identification and review of the research gaps preventing a ......2002/08/08 · identified...

1

ARTICLE INFOArticle ID: 08-08-02-0009Copyright © 2019 SAE Internationaldoi:10.4271/08-08-02-0009

HistoryReceived: 28 Jun 2019Revised: 18 Sep 2019Accepted: 29 Oct 2019e-Available: 14 Nov 2019

KeywordsOptimal energy management, Energy management strategy, Optimal control, Fuel economy, Hybrid electric vehicles

CitationAsher, Z., Patil, A., Wifvat, V., Frank, A. et al., “Identification and Review of the Research Gaps Preventing a Realization of Optimal Energy Management Strategies in Vehicles,” SAE Int. J. Alt. Power. 8(2):2019,doi:10.4271/08-08-02-0009.

ISSN: 2167-4191e-ISSN: 2167-4205

Identification and Review of the Research Gaps Preventing a Realization of Optimal Energy Management Strategies in VehiclesZachary D. Asher,1 Amol A. Patil,1 Van T. Wifvat,2 Andrew A. Frank,3 Scott Samuelsen,2 and Thomas H. Bradley4

1Western Michigan University, USA2University of California Irvine, USA3University of California Davis, USA4Colorado State University, USA

AbstractThe development of new vehicle control strategies that achieve improved fuel economy (FE) is an active subject of research due to the economic, environmental, and societal impact of transporta-tion. These control strategies can be classified as either driving behavior modifications (e.g., Eco-Driving, Eco-Routing) or powertrain operation modifications (e.g., an Optimal Energy Management Strategy, or Optimal EMS). This literature review is focused on the Optimal EMS and seeks to develop a novel understanding of the current research gaps and to provide a novel compre-hensive overview of initial studies addressing the identified research gaps. Research gaps are derived by utilizing a systems-level viewpoint of an Optimal EMS realization in vehicles and studying the subsystem integration readiness levels (IRLs). Identified research gaps include (1) incorporation of both perception and planning subsystems, (2) studying the effects of mispredictions on the planning subsystem, and (3) physical demonstrations of the planning subsystem. Studies which have begun to fill each research gap are identified, and recommendations are presented for future research to bridge each research gap. It is the authors’ contention that once the identified research gaps are closed by future studies, Optimal EMS will be achievable in modern vehicles resulting in improved transportation sustainability.

Downloaded from SAE International by Zachary Asher, Tuesday, November 26, 2019

2 Asher et al. / SAE Int. J. Alt. Power. / Volume 8, Issue 2, 2019

© 2019 SAE International. All Rights Reserved.

Introduction

Automobiles first achieved global adoption in the early twentieth century and continue to provide large economic and societal benefits. But global adoption

of automobiles also has drawbacks which include public safety risks from collisions, large contributions to energy consump-tion, contributions to air pollution, and contributions to climate change. To combat these drawbacks, technological advancements have been continually applied to the light-duty transportation sector. These technologies include safety-focused technologies such as the frontal collision warning, lane departure warning, and limited driver-assistance (collec-tively known as Advanced Driving-Assistance Systems, i.e. ADAS) as well as environment-focused technologies such as electrification and operation for improved efficiency [1].

These advancements evidence an ongoing vehicular evolution, in which human driving responsibilities are increasingly deferred to robust computerized systems of sensor technologies. With each stage of this evolution come improvements in vehicular safety as well as opportunities for minimizing fuel consumption.

The Importance of Minimizing Nonrenewable Fuel ConsumptionIn terms of global energy consumption, the transportation sector is the second largest consumer behind only the indus-trial sector. Transportation accounts for 30% of the world’s energy consumption, and the transportation energy demand is projected to increase 30% from current levels by 2040 [2]. Associated utilization of energy conversion devices, such as the internal combustion engine, results in issues spanning climate stability, domestic energy security, and human health risks from local air quality impacts.

The global transportation sector accounted for 64.5% of worldwide petroleum consumption in 2014 [2]. On a per-country basis, petroleum consumption is often unbalanced from domestic production, where disparity between the two creates the issue of energy security and vulnerability to geopo-litical stability [3]. For example, a 2016 estimation shows that the United States alone paid $150 billion to the Organization of the Petroleum Exporting Countries [4, 5] and that esti-mated annual costs to ensure continued petroleum importa-tion are around $200 billion [6].

Greenhouse gas emissions result in increased climate change, and the transportation sector was also responsible for 23% of global greenhouse gas emissions in 2014 [7, 8]. To combat these climate impacts, the Paris climate agreement has been adopted by most countries to limit greenhouse gas emissions and thus global warming to 2°C [9]. Limiting green-house gas emissions from transportation is proposed to be accomplished primarily through the increasing of FE with technologies such as improved vehicle operation efficiency and electrification [10]. It is widely accepted that without

successfully stabilizing global climate, major degradation of the global ecosystem and environment will be imminent [8].

The transportation sector is also a major contributor to air pollution. Of the six primary air pollutants, transportation significantly contributes to worldwide nitrogen oxide/nitrogen dioxide (NOx), carbon monoxide (CO), volatile organic compounds (Voc), particulate matter (PM), and sulfur dioxide (SO2) [11]. As a result, 6.5 million premature deaths were attributed to air pollution in 2012, making it the world’s fourth-largest threat to human health [12]. Additionally, it has been shown that premature deaths from air pollution is expected to rise to 7.4 million by 2040 with the caveat that it is strongly region dependent [11].

Overall, increasing vehicle FE (reducing petroleum consumption) results in lower global energy consumption, lower greenhouse gas emissions, and lower air pollution emissions. Automotive FE standards, such as those adopted by the United States, Japan, Canada, Australia, China, Taiwan, South Korea, and others, have proven to be one of the most effective tools in controlling petroleum demand and greenhouse gas emissions in many regions and coun-tries around the world [13].

The Evolution of Modern VehiclesA major trend in the automotive field is that modern vehicles are gradually incorporating electrification to evolve from conventional vehicles (CVs) to hybrid electric vehicles (HEVs), plug-in HEVs (PHEVs), and fully electric vehicles (EVs) with increasing electrification projected to continue to beyond 2040 [14]. But, of those four vehicle types, HEVs and PHEVs are unique in that they have two sources of vehicle propulsion energy available. These vehicles can be powered from either battery power, engine power, or a combination of both. This additional operational degree of freedom unlocks potential for improved overall powertrain efficiency with intelligent control strategies. Examples of how this increased vehicle powertrain operational freedom can be used to reduce fuel consumption include regenerating energy during braking, storing excess energy from the engine during coasting, and modifying the power-split powertrain component operation for maximum efficiency [15,16].

A second trend, which is in stark contrast to the first 100 years of automobile development, is that the past 10 years have brought forth the most safety and convenience upgrades due to rapid advancements in computational technology [1]. The modern vehicle is one distinguished by an ability to perceive its environment by the use of sensor technologies and computer systems. This vehicle intelligence can manifest as driver assis-tance, provided by ADAS [17], or it can manifest as a vehicle which is afforded responsibility to make driving decisions independent of a human driver through autonomous technolo-gies [18]. Near-future technologies include vehicle-to-vehicle communication (V2V), vehicle-to-infrastructure communica-tion (V2I), and vehicle-to-everything communication (V2X). These technologies enable, for the first time, vehicle and/or


Asher et al. / SAE Int. J. Alt. Power. / Volume 8, Issue 2, 2019 3


powertrain control that can be predictive and forward-looking rather than simply backward-looking or reactive.

This second trend has resulted in researchers defining an intelligent vehicle. An intelligent vehicle is defined as a system that can sense the driving environment and provide informa-tion or vehicle control to assist the driver in improved vehicle operation [19]. Intelligent vehicle aspects include the ability to (1) sense the vehicle’s own status and its environment [20], (2) communicate with the environment [20], and (3) plan and execute appropriate maneuvers [19]. Note that the “environ-ment” is defined as the vehicle surroundings which can include other vehicles, traffic lights, pedestrians, and so on. This increased environmental knowledge enables improved vehicle control strategies that can operate in tandem with inputs from a human driver.

The thesis of many researchers in this field is that the combination of these two trends (intelligent vehicle control and electrification) is a synergistic means to realize large fuel consumption improvements [21]. This article reviews numerous studies which show that the FE improvements with current intelligent HEVs are largely untapped. Existing intel-ligent electrified vehicles currently for sale include the 2020 Toyota Prius, 2020 Ford Fusion, 2020 Chevrolet Malibu, and others which offer HEV and PHEV architecture, pre-collision and pedestrian detection systems, lane departure alert with steering assist, adaptive cruise control with stop-and-go, blind-spot information with cross-traffic alert, and vehicle tracking through the global positioning system (GPS) [22, 23, 24, 25, 26].

Using Modern Vehicles to Minimize Fuel ConsumptionThere are three types of vehicle control that reduce fuel consumption for a drive cycle with a fixed starting point and a fixed ending point: (1) Eco-Driving, (2) Eco-Routing, and (3) an improved EMS. Eco-Driving and Eco-Routing decrease fuel consumption by decreasing the energy output of the vehicle through modification of the drive cycle. An improved EMS decreases fuel consumption by increasing the efficiency of the vehicle powertrain operation without modification of the drive cycle.

Eco-Driving reduces fuel consumption for all types of vehicles by implementing fuel-efficient driving behaviors along a fixed route which may alter the travel time. Due to this increase in travel time, it is challenging to convince drivers to adopt Eco-Driving habits [27]. If the driver input is removed or ignored, Eco-Driving can be formulated as an optimal control problem if the driving conditions along the route can be predicted. Current practical use of Eco-Driving is realized through a heuristic set of goals such as removing stops, traveling at a fuel-efficient speed (in general, this could be a higher or lower overall speed), and limiting acceleration and deceleration magnitudes, which together can achieve FE improvements of approximately 10% for modern vehicles and 30% for fully autonomous vehicles [28]. Research has typically focused on the FE impact of one intelligent vehicle technology,

such as camera systems, radar systems, LiDAR, V2V, V2I, or V2X. As an example of a typical Eco-Driving study, one group of researchers used predictions of traffic light signal phase and timing (a V2I technology) to change driving behavior and demonstrated a FE improvement of 12-14% [29]. In practice, Eco-Driving is challenging to implement because most drivers do not like to give up control.

Eco-Routing reduces fuel consumption for all types of vehicles by exploring alternate vehicle routes between a fixed starting and ending location. Routes that minimize travel time enable social efficiency while routes that minimize fuel consumption enable vehicle energy output efficiency. Modern commercially available routing techniques are currently designed for minimum travel time (increase social efficiency) only. Eco-Routing is an active research subject that seeks to balance travel time and fuel consumption to maintain social efficiency and improve energy output efficiency. Research on a large geographic scale indicates that a 3.9% FE improvement for a 4.5% travel time increase is possible in Cleveland, Ohio and a 6.6% FE improvement for a 1.0% travel time increase is possible in Columbus, Ohio [30]. As vehicles become more intelligent, Eco-Routing can assist vehicles in real time, but several research gaps preventing real-world implementation of Eco-Routing have been identified [31].

An improved EMS seeks to reduce the energy consump-tion over a fixed drive cycle through improved powertrain operation efficiency. Typically, an optimal control problem is formulated and an Optimal EMS is derived. An Optimal EMS realizes FE improvements by explicitly or implicitly modeling vehicle operation and controlling the vehicle powertrain components to minimize fuel consumption. An Optimal EMS does not require a change in driver behavior; thus this FE improvement technique has a consumer acceptance advantage over Eco-Driving and Eco-Routing. An Optimal EMS can realize FE improvements for CVs and EVs, but the greatest FE improvements are realized from vehicles with more powertrain operation degrees of freedom such as HEVs and PHEVs. The exact FE improvement from an Optimal EMS is strongly dependent on the chosen drive cycle and vehicle architecture. As an example, one of the earliest Optimal EMS studies demonstrated a 28% FE improvement in a hybrid electric truck through optimal control of the gear shifting and battery charging and discharging [32]. FE improvements realized through Optimal EMS are the focus of this literature review.

To summarize, if FE is improved by modifying the drive cycle or route, then it is considered Eco-Driving and Eco-Routing, respectively, and it is not within the scope of this research. FE improvements over a fixed drive cycle are realized through an improved EMS or Optimal EMS, which is an active area of research in which 300+ papers have been published in just the last decade [33].

Optimal EMS Background Developing and imple-menting an Optimal EMS has most commonly been posed as an application of optimal control. A mathematical optimiza-tion problem is formulated by defining a dynamic equation which describes the current state of the vehicle, a cost function that penalizes using fuel, and constraints that ensure a desired




final value of the battery state of charge is met, powertrain component limitations are not violated, and that the drive cycle is fixed. This optimization problem is also described in Equations 1-3.

Dynamic Equation:

New Battery State of Charge

Battery State of Charge, Engin= f ee Power, velocity( ) Eq. (1)

Cost Function:

Cost Sum mass of fuel used= ( ) Eq. (2)

Constraints:

ii

Desired battery state of charge at end of drive cycle

Powertrrain component physical limitations

Fixed drive cyclei Eq. (3)

This framework can be utilized as either a second-by-second instantaneous optimization or as a global optimization which includes future vehicle operation prediction. The solution from either of these schemes is the minimum fuel consumption strategy (or Optimal EMS) which can be then be applied to operate the vehicle powertrain.

Instantaneous Optimal EMS An instantaneous Optimal EMS involves finding the optimal control strategy that minimizes fuel consumption at the instant in time for which sampled data is available. In practical implementation, the instantaneous Optimal EMS depends on the vehicle type.

Many aspects of modern CV engine and transmission control techniques are classifiable as an instantaneous Optimal EMS including deceleration fuel cutoff [34], fuel enrichment [35], variable valve timing [36], cylinder deactiva-tion [37], and more. A basic example of the use of an instan-taneous Optimal EMS in a CV is choosing the gear in which the vehicle will operate. When the vehicle is operating at fixed speed and fixed driver torque request, the vehicle requires a fixed power demand and the vehicle must choose one trans-mission gear in which to operate. A control strategy that selects the gear to minimize fuel consumption is an example of an instantaneous Optimal EMS.

For HEVs, large FE improvements can be achieved by restricting engine power operation to the minimum fuel consumption solution (also known as the ideal operating line) [38], which is the primary fuel-saving technique employed by HEVs today [39].

In PHEVs such as the Toyota Prius Prime and the Chevrolet Volt, studies using an instantaneous Optimal EMS have led to the “charge-depleting, charge-sustaining” EMS, where all excess battery power is used first, and then the battery charge is sustained afterwards [40]. This EMS is the standard operating mode for PHEVs to minimize nonrenew-able liquid fuel consumption through maximization of renew-ably generated electrical energy consumption when prediction of future driving conditions is not available [41].

EVs also make use of instantaneous Optimal EMS improvements to minimize energy consumption and improve range through powertrain configuration optimization. Tesla Motors, an electric vehicle company, has multiple patents related to the use of an instantaneous Optimal EMS [42,43]. These patents improve the EV powertrain efficiency based on instantaneous driving conditions.

Instantaneous Optimal EMS remains an active research topic. Current efforts are focused on an implementation and quantification of the effects of an instantaneous equivalence calculation between electrical energy and fuel energy [44, 45, 46].

Predictive Optimal EMS A predictive Optimal EMS involves finding the optimal control strategy that minimizes fuel consumption for the window of time in which prediction data is available. Hundreds of papers have been written on the development and application of a predictive Optimal EMS in the past decade alone, but consistent real-world operation remains elusive [47, 48]. Predictive Optimal EMS development can fit into one of three categories: (1) a globally Optimal EMS with deterministic prediction, (2) an Optimal EMS with stochastic prediction, and (3) a computationally limited Optimal EMS to enable practical implementation [49].

A globally Optimal EMS with deterministic prediction is derived using either dynamic programming (DP) [32] or Pontryagin’s minimization principle (PMP) which is based on calculus of variations [37]. When deriving a globally Optimal EMS using deterministic prediction, DP has been the overwhelming favorite of researchers due to its ease of use and robustness, and that no derivatives or analytic expressions are required [47, 50]. A globally Optimal EMS with determin-istic prediction is difficult to implement in practice because of the high computational cost, but it is still beneficial in simu-lation to define the upper practical limit on FE benefits for a given vehicle and drive cycle.

An Optimal EMS with stochastic prediction is used in applications where researchers are willing to forgo a guarantee of global optimal FE in favor of a robustness to stochastic prediction errors. In other words, stochastic derivation strate-gies are appropriate for applications where a small increase in FE over a wide range of drive cycles is desired. Stochastic derivation strategies include stochastic dynamic program-ming (SDP) [51] and adaptive equivalent consumption mini-mization strategy (a-ECMS) [52].

Lastly, a computationally limited practical implementa-tion Optimal EMS also forgoes the guarantee of global optimal FE in favor of computationally efficient algorithms that can be used in current and near-future vehicles. Practical imple-mentation derivation strategies in current vehicles include optimized rules-based control [53], ECMS [54], and model predictive control (MPC) using fast optimizers [55].

The significant research activity surrounding predictive Optimal EMS shows no signs of slowing down as evidenced by numerous Optimal EMS studies that continue to be published monthly. But unless a shift in research scope is applied to predictive Optimal EMS development, there will




continue to be many published papers with no measurable steps toward real-world realization. This literature review is intended to provoke a discussion about new and novel research that would help realize predictive Optimal EMS in current vehicles.

Novel Contributions of This Research Gap Literature ReviewResearch that identifies and reviews existing research gaps is an essential part of scientific development [56], and it is common in multiple disciplines [57, 58, 59]. This article, which focuses on Optimal EMS in vehicles, builds from the concepts and literature discussed so far to deliver four novel contribu-tions to the field:

1. A holistic and systems-level understanding of the subsystems and integrations needed to implement a predictive Optimal EMS in vehicles allowing for comparison between all studies in the field.

2. Application of technology, integration, and system readiness analysis to predictive Optimal EMS realization in vehicles.

3. A definition of the research gaps existing between the current state of the art and realization of predictive Optimal EMS usage in vehicles.

4. A review of initial studies that have begun targeting the identified research gaps.

Research Gap DerivationMany studies of predictive Optimal EMS (now referred to as just “Optimal EMS”) conclude that vehicle perception, opera-tion prediction, optimal control calculation, and optimal control actuation must all be achieved for successful imple-mentation. However, a systems level understanding of how each of these pieces fit together is not as well defined.

To improve communication between academic researchers, the automotive industry, government stake-holders, and other entities, a systems-level viewpoint of Optimal EMS implementation, shown in Figure 1, is proposed. This systems-level viewpoint is intended to maintain close alignment with the well-accepted systems-level viewpoint of autonomous vehicle operation [18,60]. The systems-level view-point is composed of three subsystems: a vehicle perception subsystem, a vehicle planning subsystem, and a vehicle plant subsystem, which include a vehicle running controller.

The input to the Optimal EMS system is a suite of sensors which detect environmental information, thus defining vehic-ular surroundings (commonly referred to as the worldview in autonomous vehicle literature). This worldview can be used to generate a prediction of future vehicle states through arti-ficial intelligence, stochastic modeling, regression analysis, and more. The vehicle state prediction can then be utilized in

a mathematical optimization problem to determine powertrain operation that maximized FE. The maximum FE powertrain operation is then issued as a request to the vehicle running controller, which enforces component constraints and may be subject to various disturbances such as future vehicle state prediction error. The powertrain operation from the running controller is actuated in the vehicle plant, and the FE or energy consumption can be measured.

In order to assess the research gaps for an Optimal EMS implementation, a technology maturity framework is adopted. The Technology Readiness Level (TRL), originally proposed by the U.S. National Aeronautics and Space Administration, is one of the most prominent methods to assess technological maturity and help improve research and development outcomes [61]. The TRL scale has since been formally adopted worldwide despite known difficulties in application to large systems [62]. In response to these challenges, a popular and more comprehensive analysis is available which requires defining not only individual subsystem TRLs but IRLs, and an overall System Readiness Level (SRL)[63]. For the purpose of identifying the technological maturity of an Optimal EMS implementation, we adopt the SRL framework which requires that the TRL of each subsystem be defined and that the IRL of each subsystem integration be defined.

TRLsTable 1 describes the authors’ assessment of the TRL for each of the subsystems defined in Figure 1. The technologies from the first column of Table 1 are (1) vehicle perception for world-view building and prediction, (2) vehicle planning for deter-mining fuel consumption reductions through optimal powertrain operation, and (3) usage of a physical vehicle plant to measure actual fuel consumption reductions. The percep-tion subsystem receives sensor and signal inputs, defines vehicle surroundings, and thus computes future vehicle opera-tion as an output. The planning subsystem receives the vehicle operation prediction as an input and computes the optimal control as an output. Note that the planning subsystem is only required to compute the optimal control and issue a control request, this subsystem is not tasked with achieving the

FIGURE 1 A proposed systems-level viewpoint of Optimal EMS implementation.

© 2019 SAE International. All Rights Reserved




optimal control in the vehicle; achieving the optimal control is accomplished with the vehicle running controller. The final subsystem is comprised of the vehicle running controller and the vehicle plant which receives the optimal control request and the current vehicle state as inputs; determines physically feasible vehicle operation that does not violate torque, battery state of charge, speed, acceleration, etc.; and limits and actuates the vehicle plant, thus producing a measurable FE or energy consumption as an output.

IRLsTable 2 describes the authors’ assessment of the IRL for the three integration points possible in Figure 1. Descriptions of each integration scope are shown in column 1 of Table 2. While the TRL is used to evaluate individual subsystems, the IRL is used to evaluate the readiness for each subsystem to integrate with one another [54]. An accurate evaluation of each subsystem integration requires a larger scope than that of the individual subsystem, which typically consists of a simple input/output framework. Three conceptually unique integration points exist if the vehicle running controller and the vehicle plant are treated as one subsystem of high IRL: (1) perception and planning integration, (2) planning and control

disturbances integration, and (3) planning and running controller/vehicle plant integration. Each of these integration points was found to have a low technological maturity, due to the starkly limited amount of research that uses these integration scopes.

SRLWhere the TRL analysis has been applied to individual subsys-tems, and the IRL has been applied to the integration between subsystems, the SRL analysis is the appropriate means of evaluation for the entire system of Optimal EMS implementa-tion. Despite the relatively high TRLs of each subsystem, the low IRLs result in a low overall SRL as shown in Table 3. SRL analysis indicates that if the IRLs are improved, then the overall SRL will be improved and Optimal EMS can experi-ence commercial production in vehicles.

Research Gap Analysis SummarySRL analysis has clearly identified three research gaps preventing Optimal EMS realization, all of which come from

TABLE 2 The IRL analysis demonstrates that the technology integrations involved in predictive Optimal EMS implementation require significant research.

Integration and IRL Integration description IRL definition IRL justificationPerception and planning integration: IRL 1

Receives sensor/signal data and computes the maximum FE control

“An interface between technologies has been identified with enough detail to allow characterization of the relationship”

Data transfer types are not standardized and worldviews are not standardized

Planning and control disturbances: IRL 2

Receives incorrect vehicle operation prediction but still requests FE-improved control

“There is some level of specificity to characterize the interaction between technologies through their interface”

There is limited research demonstrating improved FE with incorrect predictions

Planning and use of a vehicle plant: IRL 3

Receives vehicle operation prediction and achieves maximum FE control in the vehicle

“There is compatibility between technologies to orderly and efficiently integrate and interact”

There are a few demonstrations of this integration in the literature

© 2

019

SA

E In

tern

atio

nal.

All

Rig

hts

Res

erve

d

TABLE 1 The TRL analysis demonstrates that the individual technologies involved in predictive Optimal EMS implementation are very mature.

Technology and TRL Technology description TRL definition TRL justificationPerception subsystem: TRL 7 Receives sensor/signal data and

derives a vehicle operation prediction

“System prototype demonstration in an operational environment”

The success of autonomous vehicle grand challenges and adoption of ADAS technology in modern vehicles

Planning subsystem: TRL 8 Receives vehicle operation prediction and computes the maximum FE control

“Actual system completed and qualified through test and demonstration”

Hundreds of research paper demonstrations for various EMS types, vehicles, and drive cycles

Running controller and vehicle plant: TRL 9

Receives driver requests and component statuses and actuates vehicle operation

“Actual system proven through successful mission operations”

Modern vehicles currently use running controllers to enforce component constraints and meet power demands ©

20

19 S

AE

Inte

rnat

iona

l. A

ll R

ight

s R

eser

ved




subsystem integration. The integrations that need to be addressed by research include

1. Performance of integrated perception and planning models

2. Performance of a planning model subject to disturbances

3. Performance of a planning subsystem integrated with a physical vehicle plant

This article will now thoroughly review high-quality studies that include one of these integrations only. As previ-ously discussed, there are hundreds of papers incorporating Optimal EMS, but most of them do not include one of these integrations.

The following sections are dedicated to each integration research need. A definition of the research scope is defined, a relevant research question is proposed, and literature matching the scope and research question is reviewed. Studies that do not include sufficient integration to meet the relevant research scope are not included.

Research Gap 1: Performance of Optimal EMS with Actual Velocity PredictionsProjects addressing research gap 1 would include all of the following: (1) a perception model with sensor/signal inputs and future vehicle operation prediction as an output, (2) a planning model that uses future vehicle operation prediction to derive an Optimal EMS, and (3) FE results. Only literature that meets each of these three requirements will be reviewed in this section. Note that some researchers have included this scope across multiple publications (see Michigan Technological University [64,65] and North Carolina State University [66,67] studies), but since there isn’t one complete study that includes these three components, they were omitted from full discus-sion here. This first research gap was identified by applying an integration scope that includes the perception and planning subsystems as shown in Figure 2.

The FE results from an Optimal EMS that uses actual predictions from a perception model are not well understood. Are there certain types of perception algorithms that work

better than others (e.g., Markov chain, autoregression, machine learning, etc.)? Are there certain sensor inputs that enable high-quality predictions (e.g., GPS, ADAS detections, V2V, weather information, etc.)? Although numerous vehicle sensor and signal information could be used, an Optimal EMS may only require a limited worldview with a low computa-tional cost perception algorithm to achieve optimal FE control. An as-yet unanswered research question could be posed as

• What worldviews can enable an Optimal EMS?

Starting in 2008, the earliest and presently most cited research pertaining to perception, Optimal EMS planning, and FE results comes from researchers at the University of Florida and the University of Wisconsin-Milwaukee. Their study used V2I and GPS signals as inputs into a perception model. The perception model was then able to output vehicle operation predictions using an analytical traffic model [68], and the approach was later revised to employ an artificial neural network (NN) [69]. The use of a NN to characterize a driving pattern was found to be effective, and FE improve-ments using an Optimal EMS in a PHEV were realized.

In 2013, researchers at the University of Stuttgart, Germany, have integrated a perception model and a planning model. Their perception model is based on repeated drive cycles of a hydraulic hybrid garbage truck. They use GPS and a current vehicle state to characterize and store a drive cycle. The perception model creates predictions by using the stored drive cycle data for current vehicle operation. The planning Optimal EMS only determines the optimal time to implement the hydraulic power. This team demonstrated significant FE

TABLE 3 The SRL argues that significant research is required before predictive Optimal EMS implementation can be achieved in practice.

System and SRL System description SRL definition SRL justificationOptimal EMS implementation: SRL 2

Perception and planning subjected to disturbances and implemented in a vehicle plant

“Reduce technology risks and determine appropriate set of technologies to integrate into a full system”

IRLs are too low and there is almost no research that includes a near fully realized system

© 2019 SAE International. All Rights Reserved

FIGURE 2 The integration scope defined in research gap 1: perception, an Optimal EMS, and FE results.

© 2

019

SA

E In

tern

atio

nal.

All

Rig

hts

Res

erve

d




improvements when driving behavior was predicted but noted that inclusion of traffic information is required for real-world implementation [70].

In 2014, researchers at the University of Minnesota applied a traffic model to predict future vehicle velocity with V2V and V2I as inputs. They employed PMP to derive their Optimal EMS and realized a modest FE improvement. FE gains were limited by the difficulty of incorporating constraints [71].

Then, in 2015, researchers from the University of California at Berkeley recognized the important relationship between perception and planning and investigated three perception models for use with a MPC Optimal EMS. They use previous driving data and the current vehicle state as inputs to test an exponentially varying perception model, a stochastic Markov chain perception model, and an NN perception model [72]. Their results show that the NN percep-tion model with MPC as the Optimal EMS produced the best FE results, though this result was not a full realization of globally optimal FE. Additional studies by this group have shown FE results closer to optimal when traffic information is included [73].

Starting in 2017, researchers at Colorado State University began publishing research that included a perception model, Optimal EMS, and FE results. The first study used current and previous vehicle velocity and GPS data input to a shallow NN perception model with an Optimal EMS computed using DP on a validated model of a 2010 Toyota Prius. The maximum FE improvement was achieved using 30 seconds of prediction [74]. The next study used V2V communication as an additional sensor/signal input to study the trade-offs between V2V and non-V2V signal inputs with an Optimal EMS. The results of this second study show that predictions

without V2V are more accurate at shorter prediction horizons while predictions with V2V are more accurate at longer prediction horizons [75]. Another follow-on study used custom camera detections relevant for velocity prediction [76], travel time data, as well as current and previous vehicle velocity and GPS data as the sensor/signal inputs, again with a shallow NN and an Optimal EMS. These currently available sensor inputs enabled an overall FE improvement, and it was also shown that travel time data negatively impacts predic-tions for city-focused driving, but positively impacts highway-focused driving [77].

Starting in 2017, researchers from the Beijing Institute of Technology, China, demonstrated greater and more consistent FE improvements with the aid of historical velocity data and a NN perception model to inform an ECMS-derived Optimal EMS [78]. Note that ECMS only requires the computation of an equivalence factor to deter-mine engine/battery operation, but for real-world driving, the optimal equivalence factor must be continually updated; thus there is an improvement from prediction as discussed by other researchers [79]. Additionally, other researchers from the Beijing Institute of Technology used previous drive cycle data as an input into a Markov chain and NN percep-tion subsystem with an Optimal EMS based on a genetic algorithm. These researchers were able to realize significant FE improvements in the highway fuel economy test (HWFET), LA92, and Japanese urban driving cycles even when running in real time [80].

Two additional studies that include a similar scope were published in 2017 from researchers at Jilin University, China, and in 2018 from researchers at Tongji University, China. The study from Jilin University incorporated self-reported driving behavior and predicted traffic information to develop what

TABLE 4 Summary of existing research that includes the integration scope of perception, Optimal EMS planning, and FE results, thus addressing research gap 1.

Research group Sensors/signals Perception model Planning techniques Vehicle modelUniv. of Wisconsin Milwaukee 2009 [68,69]

GPS, V2I Traffic model, NN DP Generic hybrid SUV

Univ. of Stuttgart 2013 [70] Vehicle state, route, GPS Database look-up Custom hydraulic power optimization

Parallel hydraulic hybrid truck

Univ. of Minnesota 2014 [71] V2V, V2I Traffic model PMP Generic power-split HEV

Univ. of California Berkeley 2015 [72,73]

Vehicle state, drive data, V2I

Exp. varying Markov chain & NN

MPC Generic power-split HEV

Colorado State University 2017, 2018 [74, 75, 76, 77]

Vehicle state, drive data, GPS, V2V, travel time, ADAS

NN DP Validated 2010 Toyota Prius

Beijing Institute of Technology 2017, 2018 [78,80]

Vehicle state, drive data Markov chain and NN a-ECMS and genetic algorithm

Generic power-split HEV

Jilin University 2017 [81] Travel info., driving behavior, GPS

NN Improved a-ECMS PHEV

Tongji University 2018 [82] Vehicle state, drive data Markov chain DP Hybrid electric bus

Univ. of California Riverside 2019 [83]

Traffic data Deep NN Reinforcement learning

Generic power-split PHEV

Univ. of Michigan 2019 [84,85] Vehicle state, GPS, V2V, V2I Deep NN, Markov chain, conditional Gaussian, and more

DP Validated 2017 Toyota Prius Prime

© 2

019

SA

E In

tern

atio

nal.

All

Rig

hts

Res

erve

d




they call “improved a-ECMS” which realized a larger FE improvement than a-ECMS [81]. The study from Tongji University compares ECMS, a-ECMS, and DP with a velocity prediction perception model [82]. These researchers conclude that DP with a real-world velocity prediction may provide the best FE improvement.

Recent trends have begun to implement advancements in machine learning and have replaced shallow NNs with deep NNs. This was first demonstrated by researchers at the University of California Riverside where a deep reinforce-ment network was implemented as a perception and planning subsystem in that traffic data was input and optimal control of a generic PHEV was the output. This technique resulted in a 16.3% FE improvement in simulation [83]. Additionally, research that was conducted at the University of Michigan and then transitioned to Western Michigan University explored a variety of perception models including autore-gressive moving average, shallow NN, long short-term memory (LSTM) deep NN, Markov chain, and conditional linear Gaussian models. It was determined that the LSTM deep NN provided the best prediction fidelity (measured in mean absolute error) [84] which realized a FE improvement of 3% in a dynamometer validated model of a 2017 Toyota Prius Prime [85].

Each of these studies has been summarized in Table 4 using columns to show the important differences between the studies (i.e., which sensors/signals were used, what perception model was used, etc.). Taken together, these papers show that the best results were obtained when a NN is used as a percep-tion model. But there is still no comprehensive study investi-gating the effect of different groups of signal inputs or percep-tion models. Future research in this area will likely leverage the recent breakthroughs in deep learning for an improved perception model.

Research Gap 2: Performance of Optimal EMS when Subjected to DisturbancesProjects addressing research gap 2 would include all of the following: (1) a planning model that uses future vehicle opera-tion prediction to derive an Optimal EMS, (2) disturbances in actual control from disturbances such as mispredictions, and (3) FE results. Literature that meets each of these three requirements will be reviewed in this section. This second research gap was identified by applying an integration scope that includes the planning subsystem and disturbances as shown in Figure 3. Note that other researchers have also iden-tified this as an important research gap [42].

The FE results from an Optimal EMS subjected to disturbances are not well understood. Are there certain types of disturbances that have large FE impacts (e.g.,

speed-up mispredictions, unrecognized stop signs, current vehicle weight mispredictions, outside wind speed/direc-tion changes, etc.)? Are there certain Optimal EMS types that are more resistant to certain types of disturbances (e.g., stochastic DP, MPC, ECMS, etc.)? Although numerous disturbances are possible, perhaps some disturbances are worse than others or perhaps most disturbances have no impact. An as-yet unanswered research question could be posed as

• How sensitive are optimal FE strategies to worldview/prediction scope, fidelity, and uncertainty?

In 2006, researchers at the National Renewable Energy Laboratory utilized DP for half battery size and full battery size PHEVs to compare optimal solutions for drive cycles of various lengths. They found that for drive cycle distances exceeding the PHEV all electric range (charge depleting strategy), the optimal solution seeks a “blended” strategy (a mix of charge depleting and charge sustaining strategies) such that the minimum battery state of charge is achieved at the end of the drive cycle [86]. But in considering the “cost of being wrong,” they demonstrated that if the vehicle distance is less than predicted for optimal control, there is a significant FE loss. In other words, if the Optimal EMS assumes the trip will be longer than it actually is, the vehicle will not use all the stored electrical energy and thus requiring more liquid fuel consumption.

In 2011, researchers at the Ohio State University conducted a study using alternate drive cycles as a prediction input error for two real-time implementable Optimal EMS: a modified MPC strategy and a-ECMS [87]. They found that the FE cost of predicting the wrong drive cycle was small for both of the Optimal EMS, but no baseline FE results were shown.

In 2012, researchers at Clemson University investigated the effects of stochastic misprediction errors on a-ECMS [88]. This group used the U.S. EPA drive cycles of Urban Dynamometer Driving Schedule (UDDS), US06, and HWFET for 25, 50, 75, 100, and 150 miles. Their results show a degra-dation in FE at all values of mean absolute percent error

FIGURE 3 The integration scope defined in research gap 2: Optimal EMS planning subject to disturbances.

© 2

019

SA

E In

tern

atio

nal.

All

Rig

hts

Res

erve

d




greater than zero. But no correlation between error magni-tudes and FE was found.

Then in 2014, researchers at the University of Michigan built upon their initial SDP-derived Optimal EMS from 2004 [51] and studied the impact of driving alternate drive cycles [89]. These researchers used a large real-world drive cycle data set to derive a new stochastic Optimal EMS. The stochastic Optimal EMS was demonstrated to improve FE over a baseline strategy when tested on alternate drive cycles. Additionally, a researcher from this project has continued to study the effect of uncertain events on energy management for HEV, PHEV, EV, and fuel cell vehicles. A generalized stochastic model was proposed that incorporates uncertain route predictions, desti-nations, charging locations, traffic predictions, historical speed, grade data, and driver models. SDP is used with a generic PHEV sedan on a drive cycle similar to the FTP72, and the results show increased battery usage when approaching the destination [90].

Starting in 2015, researchers at Colorado State University expanded on initial research [91,92] by rigorously investi-gating the effect of real-world velocity prediction errors and vehicle parameter prediction errors on FE realized through Optimal EMS. A set of driving-derived disturbances such as mispredicted stops, traffic, and route changes were compared using a validated model of a 2010 Toyota Prius using an Optimal EMS derived using DP. It was found that a wide range of FE improvements can still be achieved despite what would seem to be significant velocity mispredictions [50]. The work was also expanded into a new application that uses a precom-puted and robust Optimal EMS applied to acceleration event predictions [93].

In 2016, researchers at the University of Minnesota inves-tigated an Optimal EMS which was focused on fast computa-tion through separable programming while also being subjected to disturbances [94]. Using a traffic simulator to develop drive cycles and 30 different vehicle platforms, they compared their Optimal EMS to DP and showed favorable results. When the drive cycle predictions were subjected to a normally distributed random prediction error, FE improve-ments were maintained at a 3% mean absolute deviation

percentage, but FE improvements were lost at a 6% mean absolute deviation percentage.

Starting in 2017, a group of researchers from the Beijing Institute of Technology studied the influence of an imprecise driving cycle prediction on an Optimal EMS implemented in a PHEV. First, they derived an Optimal EMS using particle swarm optimization assuming a precise prediction of the future drive cycle is available. Next, they analyzed the influence of imprecise prediction and proposed an online correction algorithm based on a fuzzy logic control strategy. This overall system is demonstrated on the US06, EUDC, and REP05 drive cycles and they show that their method significantly improves the results [95]. An additional study from the Beijing Institute of Technology analyzed the effect of driving style on a-ECMS-derived FE improvements. They validated their model by conducting experiments with five different drivers (three moderate and two aggressive) in a real-world urban cycle. They also developed tools to precisely recognize the instantaneous driving style, and they compare their a-ECMS to traditional ECMS. The FE improvement was significant but depended on correct prediction of the driving style [96].

Each study addressing this research gap has been summa-rized in Table 5, using columns to show the important differ-ences between the studies. Taken together, these papers show that most Optimal EMS derivations are robust to prediction errors. But there is still no comprehensive study investigating the effect of different types of prediction errors on different Optimal EMS derivations.

Research Gap 3: Performance of Optimal EMS in Real VehiclesProjects addressing research gap 3 would include all of the following: (1) a planning model that uses future vehicle opera-tion prediction to derive an Optimal EMS; (2) actual vehicle

TABLE 5 Summary of existing research that includes the integration scope of Optimal EMS planning, disturbances, and FE results and thus addressing research gap 2.

Research group Drive cycle Planning technique Error type Vehicle modelNational Renewable Energy Lab 2006 [86]

LA92, UDDS DP Drive cycle distance Half-size PHEV, full-size PHEV

The Ohio State Univ. 2011 [87] Real world MPC, a-ECMS Alternate drive cycle Generic power-split PHEV

Clemson University 2012 [88] UDDS, US06, HWFET a-ECMS Stochastic Generic power-split PHEV

Univ. of Michigan 2014, 2016 [89,90]

Real world, FTP72 Stochastic DP Alternate real-world, uncertain events

Prototype Volvo S-80, generic PHEV

Colorado State University 2015-2018 [91, 92, 93]

Real world DP Real world, vehicle power

Validated 2010 Toyota Prius

Univ. of Minnesota 2016 [94] Traffic simulation Separable programming

Stochastic Generic power-split PHEV

Beijing Institute of Technology 2017, 2018 [95,96]

Real world, US06, EUDC, REP05

DP, a-ECMS Slightly off target Commuter PHEV

© 2

019

SA

E In

tern

atio

nal.

All

Rig

hts

Res

erve

d




hardware, preferable a full vehicle; and (3) physical FE results. Literature that meets each of these three requirements will be reviewed in this section. This third research gap was identi-fied by applying an integration scope that includes the percep-tion and planning technologies as shown in Figure 4.

The FE results from an Optimal EMS that uses actual predictions from a perception model are not well understood. Are there certain types of perception algorithms that work better than others (e.g., Markov chain, autoregression, machine learning, etc.)? Are there certain sensor inputs that enable high-quality predictions (e.g., GPS, ADAS detections, V2V, weather information, etc.)? Although numerous vehicle sensor and signal information could be used, an Optimal EMS may only require a limited worldview with a low computa-tional cost perception algorithm to achieve optimal FE control. An as-yet unanswered research question could be posed as

• What operational challenges have not been validated and quantified in the conceptual work?

There is limited research that includes Optimal EMS planning, vehicle plant considerations, and FE results because software-in-the-loop, hardware-in-the-loop, and vehicle-in-the-loop research is expensive and the technologies for specific vehicles are proprietary. Because of this, initial research in this area has been focused on a rules-based Optimal EMS due to the ease of implementation [97, 98, 99, 100]. But there are research programs designed to overcome cost and proprietary information challenges such as FutureCar, EcoCAR, and the most recent NEXTCAR program, which is an initiative to reduce vehicle energy consumption by 20% [101]. These programs have enjoyed significant success and, in some cases, have been the catalyst for research addressing research gap 3 at universities.

In 2001, a computationally limited Optimal EMS imple-mentation was achieved by researchers at the Ohio State University who implemented ECMS in an SUV as part of the FutureTruck program [102]. This initial demonstration produced promising FE results and inspired rigorous research of ECMS at the Ohio State University that has been ongoing for nearly two decades [103]. This research is also being explored as part of the NEXTCAR program [104].

Starting in 2008, researchers at the University of Michigan took their initial SDP-derived Optimal EMS [51] and first included engine-in-the-loop simulations to closer approxi-mate a real vehicle [105] and then applied it in a prototype vehicle [106]. Although a hardware problem prevented them from attaining a full FE assessment from their SDP controller, they concluded that an Optimal EMS can be implemented in a real vehicle and requires minimal tuning.

Starting in 2011, researchers at IFP Energies Nouvelles, or French Institute of Petroleum, have successfully imple-mented an Optimal EMS in a hardware-in-the-loop simula-tion for a diesel hybrid vehicle. Diesel-powered vehicles add the extra challenge of incorporating emissions minimization into the optimization scheme. Using ECMS, they were able to significantly improve FE, while reducing oxides of Nitrogen (NOx, an emissions constituent) through parallel HEV archi-tecture [107, 108, 109, 110].

In 2012, researchers at the Eindhoven University of Technology, the Netherlands, have successfully imple-mented an Optimal EMS using PMP to achieve a moderate FE improvement over a baseline, rules-based, control strategy on a hybrid electric truck. The FE improvement was demonstrated on a variety of drive cycles, including road grade variations. They achieved success in imple-menting the Optimal EMS on the vehicle’s engine control unit (ECU) with the assistance of stored solutions from off line computations [111].

Also, in 2012, researchers at Shanghai Jiao Tong University, China, demonstrated FE improvements over a baseline rules-based control strategy using a hardware-in-the-loop real-time simulation. These researchers derived an Optimal EMS using a form of DP and then used the solution to train a NN. The trained NN then accepted drive cycle inputs and output an improved control strategy [112].

In 2013, researchers at the University of Stuttgart, Germany, incorporated a perception model and a planning model into an actual vehicle. They were able to test their Optimal EMS using a rapid prototyping unit on their hybrid hydraulic truck. When testing on a closed track, this team found FE improvements as well as instances of reduced FE when compared to the baseline control strategy, though these results were briefly mentioned and no detailed analysis was shown [70].

Also, in 2013, researchers at the University of Salento, Italy, investigated the implementation of a PMP-derived Optimal EMS in a hardware-in-the-loop simulation of a custom-built series PHEV. These researchers investigated multiple local drive cycle section prediction cases. The researchers were able to demonstrate an overall FE improve-ment despite needing to reduce the complexity of their Optimal EMS planning formulation to satisfy real-time running constraints [113].

In 2016, researchers from Tsinghua University, China, proposed an Optimal EMS that is a combination of offline SDP and ECMS for a plug-in hybrid electric bus. The Optimal EMS was tested in simulation and hardware-in-loop using real-world urban bus routes. The real-world bus route data is

FIGURE 4 The integration scope defined in research gap 3: Optimal EMS planning with actual vehicle plant considerations.

© 2

019

SA

E In

tern

atio

nal.

All

Rig

hts

Res

erve

d




divided into segments according to the position of bus stops which are used as inputs into the SDP Optimal EMS formula-tion. The results are then converted into a three-dimensional lookup table of parameters for the real-time operation ECMS. They observed a significant increase in FE but noted that the hardware-in-the-loop case consumes slightly more fuel than the simulated one [114].

In 2017, researchers from Arizona State University and Clemson University proposed an integrated vehicle hardware-in-the-loop approach to address the limitation associated with a traditional hardware-in-the-loop approach. Both rules based and ECMS were tested using the integrated vehicle hardware-in-the-loop approach implemented on a parallel through-the-road HEV. The drive cycles analyzed include the FHDS, FUDS, and US06. The general results show strong correlation between the simulated vehicle and the integrated vehicle hardware-in-the-loop approach. The results from physical vehicle fuel consumption measurements indicate that the ECMS achieved a higher FE [115].

Also in 2017, researchers from the Beijing Institute of Technology proposed an EMS using Markov Chain Monte Carlo velocity prediction with average filtering and quadratic fitting, a stochastic MPC as the Optimal EMS, and hardware-in-the-loop experiments. When tested on the Chinese Typical Urban Driving Cycle (CTUDC), the observed a substantial FE improvement over a rules-based strategy, but less FE than a globally Optimal EMS [116].

In 2018, researchers from Kunming University of Science and Technology, China, investigated a PMP Optimal EMS that controls engine on/off state using simulations and hardware-in-the-loop experiments. The proposed strategy was demon-strated on a variety of drive cycles such as the UDDS, HWFET, and New European Driving Cycle (NEDC), and real-world cycles. They selected a charge-depleting/charge-sustaining strategy as their lower bound and a globally Optimal EMS through DP as the highest FE bound. The results show a signifi-cant FE improvement at a low computational cost that can be practically implemented [117].

As part of the NEXTCAR program, researchers at Michigan Technological University have explored an opti-mized blended mode strategy in the Chevy Volt PHEV. DP is implemented and FE improvements between 2 and 12% are measured in a fleet of four instrumented Chevy Volts [118].

Each of the initial studies addressing this research gap has been summarized in Table 6, using columns to show the impor-tant differences between the studies. Taken together, these studies show that through gradual incorporation of vehicle plant components through software-in-the-loop, hardware-in-the-loop, and vehicle-in-the-loop experiments, Optimal EMS implementation that improves FE can be realized. Numerous research opportunities still exist that could fill this research gap by focusing on the demonstration of an Optimal EMS in a variety of vehicle operational environments with a variety of vehicle types, platforms, and powertrains.

TABLE 6 Summary of existing research that includes the integration scope of Optimal EMS planning, a vehicle plant, and FE results, thus addressing research gap 3.

Research group Drive cycle Planning technique Vehicle plant model Vehicle realizationThe Ohio State Univ. 2001, 2019 [102, I04]

Real world ECMS 2000 Chevrolet Suburban Actual vehicle

Univ. of Michigan 2008, 2013 [105,106]

Real World Stochastic DP Prototype Volvo S-80 Actual vehicle

French Institute of Petroleum 2011 [107, 108, 109, 110]

NEDC, FTP ECMS Micro- hybrid, parallel hybrid

Hardware-in-the-loop

Eindhoven University of Tech. 2012 [111]

Real world PMP Parallel hybrid electric truck


Shanghai Jiao Tong University 2012 [112]

CTBDC Iterative DP Power-split hybrid electric bus


Univ. of Stuttgart 2013 [70] OC Bus Cycle, Real world

Custom hydraulic power opt.

Parallel hydraulic hybrid truck

Actual vehicle

Univ. of Salento 2013 [113] Real world PMP Prototype series PHEV Hardware-in-the-loop

Tsinghua University 2016 [114] Real world SDP, ECMS Parallel PHEB Hardware-in-the-loop

Arizona State University 2017 [115]

FHDS, US06, FUDS Rules based, ECMS Through-the-road HEV Integrated vehicle hardware-in-the-loop

Beijing Institute of Technology 2017 [116]

CTUDC Stochastic MPC Plug-in hybrid electric bus


Kunming Univ. of Science and Technology 2018 [117]

UDDS, HWFET, NEDC, real World

PMP Power-split PHEV Hardware-in-the-loop

Michigan Technological University 2019 [118]

Real World Limited preview DP Chevy Volt PHEV Actual vehicle

© 2

019

SA

E In

tern

atio

nal.

All

Rig

hts

Res

erve

d




ConclusionsIn this literature review, vehicle control for optimal FE was introduced by discussing methods that increase FE by modi-fying vehicle velocity (Eco-Driving, Eco-Routing) as well as a method that increases FE without modifying vehicle velocity (Optimal EMS). Upon further exploration of Optimal EMS literature, it was discovered that despite the existence of several hundred studies addressing the concept of Optimal EMS, integration and system-level research gaps still exist. When systems-level analysis is applied, three research gaps are identified: Optimal EMS planning with perception, Optimal EMS planning including disturbances, and Optimal EMS execution with a vehicle plant. In other words, gaps in academic knowledge exist regarding

1. An understanding of essential sensors and signals for perception and prediction enabling optimal FE vehicle control

2. A deep understanding of disturbance types that can affect optimal FE vehicle control

3. The operational and real-world vehicle implementation challenges of optimal FE vehicle control.

A review of insightful research efforts in each of these areas was presented, and recommendations for future efforts that can advance the body of knowledge of Optimal EMS were discussed. Overall, four novel contributions were made. They include a systems-level viewpoint of the subsystems and inte-grations needed for Optimal EMS allowing comparisons between all studies in the field, system readiness analysis applied to predictive Optimal EMS realization in vehicles, formal definitions and example research questions for existing research gaps, and a targeted review of initial studies that address the research gaps.

Engineering research is responsible for providing action-able ideas for industry incorporation. Because there are hundreds of combinations of vehicle architectures, drive cycles, and optimization routines, there have been hundreds of research papers showing the same result: an Optimal EMS can increase FE [42]. We present this comprehensive literature review on the Optimal EMS concept with the objective of focusing the efforts of the research community on increasing the TRL associated with the key interfaces in the vehicle. Once the IRL of the various subsystems is increased, Optimal EMS can be realized commercially. Once an Optimal EMS has achieved widespread usage in vehicles, the burdens of low FE such as oil importation, climate change, and air pollution can be reduced.

Contact InformationZachary D. AsherWestern Michigan [email protected]

AbbreviationsADAS - Advanced Driver-Assistance SystemCTUDC - Chinese Typical Urban Driving CycleCV - Conventional VehicleDP - Dynamic ProgrammingECMS - Equivalent Consumption Minimization StrategyECU - Engine Control UnitEMS - Energy Management StrategyEPA - Environmental Protection AgencyEV - Electric VehicleFE - Fuel EconomyGPS - Global Positioning SystemHEV - Hybrid Electric VehicleHWFET - Highway Fuel Economy TestIRL - Integration Readiness LevelMPC - Model Predictive ControlNEDC - New European Driving CycleNOx - Oxides of NitrogenOptimal EMS - Optimal Energy Management StrategyPHEV - Plug-in Hybrid Electric VehicleSDP - Stochastic Dynamic ProgrammingSRL - System Readiness LevelSUV - Sport Utility VehicleTRL - Technology Readiness LevelUDDS - Urban Dynamometer Driving ScheduleV2I - Vehicle-to-Infrastructure CommunicationV2V - Vehicle-to-Vehicle Communication

References 1. Fagnant, D.J. and Kockelman, K., “Preparing a Nation for

Autonomous Vehicles: Opportunities, Barriers and Policy Recommendations,” Transportation Research Part A: Policy and Practice 77:167-181, 2015.

2. International Energy Agency, Key World Energy Statistics 2016 (Paris, France: International Energy Agency, 2015).

3. van Moerkerk, M. and Crijns-Graus, W., “A Comparison of Oil Supply Risks in EU, US, Japan, China and India under Different Climate Scenarios,” Energy Policy 88:148-158, 2016/1.

4. U.S. Energy Information Administration, “Petroleum and Other Liquids: Europe Brent Spot Price FOB,” 2017, https://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=RBRTE&f=M, accessed May 8, 2017.

5. U.S. Energy Information Administration, “Petroleum and Other Liquids: U.S. Imports of Crude Oil and Petroleum Products,” https://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=MTTIMUS1&f=M, accessed May 8, 2016


[email protected]

https://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=RBRTE&f=M



https://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=MTTIMUS1&f=M

https://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=MTTIMUS1&f=M



6. Greene, D.L. and Ahmad, S., Costs of US Oil Dependence: 2005 Update (United States: Department of Energy, 2005).

7. International Energy Agency, “CO2 Emissions from Fuel Combustion,” 2016.

8. National Research Council, “Division on Earth and Life Studies, Board on Atmospheric Sciences and Climate, America’s Climate Choices: Panel on Advancing the Science of Climate Change,” In: Advancing the Science of Climate Change. (Washington, DC: National Academies Press, 2011).

9. Rogelj, J., den Elzen, M., Höhne, N., Fransen, T. et al., “Paris Agreement Climate Proposals Need a Boost to Keep Warming Well below 2°C,” Nature 534:631-639, 2016.

10. International Energy Agency, “Energy and Climate Change - World Energy Outlook Special Report,” 2015.

11. International Energy Agency, “World Energy Outlook Special Report: Energy and Air Pollution,” 2016.

12. World Health Organization, World Health Statistics 2016: Monitoring Health for the Sustainable Development Goals (SDGs) (Geneva, Switzerland: World Health Organization, 2016).

13. Atabani, A.E., Badruddin, I.A., Mekhilef, S., and Silitonga, A.S., “A Review on Global Fuel Economy Standards, Labels and Technologies in the Transportation Sector,” Renewable Sustainable Energy Review 15:4586-4610, 2011.

14. Energy Information Administration, “Annual Energy Outlook 2017,” 2017.

15. Lukic, S.M. and Emadi, A., “Effects of Drivetrain Hybridization on Fuel Economy and Dynamic Performance of Parallel Hybrid Electric Vehicles,” IEEE Transactions on Vehicular Technology 53:385-389, 2004.

16. Rajamani, R., Vehicle Dynamics and Control (New York, NY: Springer Science & Business Media, 2011).

17. Bengler, K., Dietmayer, K., Farber, B., Maurer, M. et al., “Three Decades of Driver Assistance Systems: Review and Future Perspectives,” IEEE Intelligent Transportation System Magazine 6:6-22, 2014.

18. Thrun, S., “Toward Robotic Cars,” Commun ACM 53:99-106, 2010.

19. Bishop, R., “Intelligent Vehicle Technology and Trends,” trid.trb.org 2005.

20. Gusikhin, O., Filev, D., and Rychtyckyj, N., “Intelligent Vehicle Systems: Applications and New Trends,” In: Informatics in Control Automation and Robotics. (Heidelberg, Springer Berlin, 2008), 3-14.

21. Fulton, L., Mason, J., and Meroux, D., Three Revolutions in Urban Transportation (USA: University of California Davis, 2017).

22. Toyota, “2020 Toyota Prius Features,” https://www.toyota.com/prius/prius-features/, accessed Sep. 18, 2019

23. Chevrolet, “2020 Chevy Malibu|Stylish & Powerful Midsize Car,” https://www.chevrolet.com/cars/malibu, accessed Sep. 18, 2019.

24. Ford Motor Company, “2020 Ford® Fusion Sedan|Fuel-Efficient Midsize Sedan & Hybrid|Ford.com,” https://www.ford.com/cars/fusion/, accessed Sep. 18, 2019.

25. Gray, D., “10 Cars with Adaptive Cruise Control,” Autobytel, http://www.autobytel.com/car-buying-guides/features/10-cars-with-adaptive-cruise-control-131148/, accessed May 8, 2017.

26. Enge, P. and Misra, P., “Special Issue on Global Positioning System,” Proceedings of the IEEE 87:3-15, 1999.

27. Gonder, J., Earleywine, M., and Sparks, W., “Analyzing Vehicle Fuel Saving Opportunities through Intelligent Driver Feedback,” SAE Int. J. Passeng. Cars - Electron. Electr. Syst. 5:450-461, 2012, doi:10.4271/2012-01-0494.

28. Michel, P., Karbowski, D., and Rousseau, A., “Impact of Connectivity and Automation on Vehicle Energy Use,” SAE Technical Paper 2016-01-0152, 2016, 2016, doi:10.4271/2016-01-0152.

29. Mandava, S., Boriboonsomsin, K., and Barth, M., “Arterial Velocity Planning Based on Traffic Signal Information under Light Traffic Conditions,” in 2009 12th International IEEE Conference on Intelligent Transportation Systems, St. Louis, MO, 2009, 1-6, ieeexplore.ieee.org.

30. Ahn, K. and Rakha, H.A., “Network-Wide Impacts of Eco-Routing Strategies: A Large-Scale Case Study,” Transportation Research Part D: Transport and Environment 25:119-130, 2013.

31. Pillac, V., Gendreau, M., Guéret, C., and Medaglia, A.L., “A Review of Dynamic Vehicle routing Problems,” European Journal of Operational Research 225:1-11, 2013.

32. Lin, C.C., Kang, J.M., Grizzle, J.W., and Peng, H., “Energy Management Strategy for a Parallel Hybrid Electric Truck,” in Proceedings of the 2001 American Control Conference (Cat. No.01CH37148), Arlington, VA, 2001, 2878-2883, vol. 4, ieeexplore.ieee.org.

33. Zhang, P., Yan, F., and Du, C., “A Comprehensive Analysis of Energy Management Strategies for Hybrid Electric Vehicles Based on Bibliometrics,” Renewable Sustainable Energy Review 48:88-104, 2015/8.

34. Evans, D.G., Polom, M.E., Poulos, S.G., and Van Maanen, K.D., “Powertrain Architecture and Controls Integration for GM’s Hybrid Full-Size Pickup Truck,” SAE Technical Paper 2003-01-0085, 2003, doi:10.4271/2003-01-0085.

35. Grandin, B. and Ångström, H.-E., “Replacing Fuel Enrichment in a Turbo Charged SI Engine: Lean Burn or Cooled EGR,” SAE Technical Paper 1999-01-3505, 1999, doi:10.4271/1999-01-3505.

36. Ma, T.H., “Effect of Variable Engine Valve Timing on Fuel Economy,” SAE Technical Paper 880390, 1988, doi:10.4271/880390.

37. Leone, T.G. and Pozar, M., “Fuel Economy Benefit of Cylinder Deactivation-Sensitivity to Vehicle Application and Operating Constraints,” SAE Technical Paper 2001-01-3591, 2001, doi:10.4271/2001-01-3591.

38. Frank, A.A., “Control Method and Apparatus for Internal Combustion Engine Electric Hybrid Vehicles,” US Patent, 2000.

39. Muta, K., Yamazaki, M., and Tokieda, J., “Development of New-Generation Hybrid System THS II-Drastic Improvement of Power Performance and Fuel Economy,”


https://www.toyota.com/prius/prius-features/

https://www.toyota.com/prius/prius-features/

https://www.chevrolet.com/cars/malibu

https://www.ford.com/cars/fusion/

https://www.ford.com/cars/fusion/

http://www.autobytel.com/car-buying-guides/features/10-cars-with-adaptive-cruise-control-131148/

http://www.autobytel.com/car-buying-guides/features/10-cars-with-adaptive-cruise-control-131148/

http://dx.doi.org/10.4271/2012-01-0494

https://www.sae.org/publications/technical-papers/content/2016-01-0152

http://dx.doi.org/10.4271/2016-01-0152

http://dx.doi.org/10.4271/2016-01-0152

ieeexplore.ieee.org

ieeexplore.ieee.org


http://dx.doi.org/10.4271/2003-01-0085


http://dx.doi.org/10.4271/1999-01-3505

https://www.sae.org/publications/technical-papers/content/880390

http://dx.doi.org/10.4271/880390


http://dx.doi.org/10.4271/2001-01-3591



SAE Technical Paper 2004-01-0064, 2004, doi:10.4271/2004-01-0064.

40. Wang, Q. and Frank, A.A., “Plug-in HEV with CVT: Configuration, Control, and Its Concurrent Multi-Objective Optimization by Evolutionary Algorithm,” International Journal of Automotive Technology 15:103-115, 2014.

41. Bradley, T.H. and Quinn, C.W., “Analysis of Plug-in Hybrid Electric Vehicle Utility Factors,” Journal of Power Sources 195:5399-5408, 2010.

42. Tang, Y., “Control System for an All-Wheel Drive Electric Vehicle,” US Patent, 2010.

43. Baglino, A.D., Hofmann, H.F., and Solberg, G.G., “Varying Flux versus Torque for Maximum Efficiency,” US Patent, 2010.

44. Sciarretta, A., Back, M., and Guzzella, L., “Optimal Control of Parallel Hybrid Electric Vehicles,” IEEE Transactions on Control Systems Technology 12:352-363, 2004.

45. Changjian, H., Gao, Y., and Alex, Q.H., “Power Management Strategy of Hybrid Electric Vehicles Based on Particle Swarm Optimization,” in 2015 IEEE Transportation Electrification Conference and Expo (ITEC), Metro Detroit, MI, 2015, 1-6.

46. Zhao, J. and Sciarretta, A., “Design and Control Co-Optimization for Hybrid Powertrains: Development of Dedicated Optimal Energy Management Strategy,” IFAC-PapersOnLine 49:277-284, 2016.

47. Asher, Z.D., Wifvat, V., Navarro, A., Samuelsen, S. et al., “The Importance of HEV Fuel Economy and Two Research Gaps Preventing Real World Implementation of Optimal Energy Management,” SAE Technical Paper 2017-26-0106, 2017, doi:10.4271/2017-26-0106.

48. Wu, G., Qi, X., Barth, M., and Boriboonsomsin, K., “Advanced Energy Management Strategy Development for Plug-in Hybrid Electric Vehicles,” Technical Report, 2016.

49. Kim, N., Cha, S., and Peng, H., “Optimal Control of Hybrid Electric Vehicles Based on Pontryagin’s Minimum Principle,” IEEE Transactions on Control Systems Technology 19:1279-1287, 2011.

50. Asher, Z.D., Baker, D.A., and Bradley, T.H., “Prediction Error Applied to Hybrid Electric Vehicle Optimal Fuel Economy,” IEEE Transactions on Control Systems Technology, 26(6):2121-2134, 2017.

51. Lin, C.C., Peng, H., and Grizzle, J.W., “A Stochastic Control Strategy for Hybrid Electric Vehicles,” in Proceedings of the 2004 American Control Conference, Boston, MA, 2004, 4710-4715, vol. 5, ieeexplore.ieee.org.

52. Onori, S., Serrao, L., and Rizzoni, G., “Adaptive Equivalent Consumption Minimization Strategy for Hybrid Electric Vehicles,” in ASME 2010 Dynamic Systems and Control Conference, Cambridge, MA, American Society of Mechanical Engineers, 2010, 499-505.

53. Bianchi, D., Rolando, L., Serrao, L., Onori, S. et al., “A Rule-Based Strategy for a Series/Parallel Hybrid Electric Vehicle: An Approach Based on Dynamic Programming,” in ASME 2010 Dynamic Systems and Control Conference, Cambridge, MA, American Society of Mechanical Engineers, 2010, 507-514.

54. Paganelli, G., Delprat, S., Guerra, T.M., Rimaux, J. et al., “Equivalent Consumption Minimization Strategy for Parallel Hybrid Powertrains,” in Vehicular Technology Conference IEEE 55th Vehicular Technology Conference, VTC Spring 2002 (Cat. No.02CH37367), Birmingham, AL, 2002, 2076-2081, vol. 4, ieeexplore.ieee.org.

55. Borhan, H.A., Vahidi, A., Phillips, A.M., Kuang, M.L. et al., “Predictive Energy Management of a Power-Split Hybrid Electric Vehicle,” in 2009 American Control Conference, St. Louis, MO, 2009, 3970-3976, ieeexplore.ieee.org.

56. Robinson, K.A., Saldanha, I.J., and McKoy, N.A., “Development of a Framework to Identify Research Gaps from Systematic Reviews,” Journal of Clinical Epidemiology 64:1325-1330, 2011.

57. Peden, A.E., Franklin, R.C., and Leggat, P.A., “Fatal River Drowning: The Identification of Research Gaps through a Systematic Literature Review,” Injury Prevention 22:202-209, 2016.

58. Eerkes-Medrano, D., Thompson, R.C., and Aldridge, D.C., “Microplastics in Freshwater Systems: A Review of the Emerging Threats, Identification of Knowledge Gaps and Prioritisation of Research Needs,” Water Research 75:63-82, 2015.

59. Westgate, M.J., Barton, P.S., Pierson, J.C., and Lindenmayer, D.B., “Text Analysis Tools for Identification of Emerging Topics and Research Gaps in Conservation Science,” Conservation Biology 29:1606-1614, 2015.

60. Campbell, M., Egerstedt, M., How, J.P., and Murray, R.M., “Autonomous Driving in Urban Environments: Approaches, Lessons and Challenges,” Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Science 368:4649-4672, 2010.

61. Mankins, J.C., “Technology Readiness Levels,” White Paper 6, 1995.

62. Mankins, J.C., “Technology Readiness Assessments: A Retrospective,” Acta Astronautica 65:1216-1223, 2009.

63. Sauser, B. and Verma, D., “From TRL to SRL: The Concept of Systems Readiness Levels,” UCLA Law Review, 2006, Paper #126, 1-10.

64. Rezaei, A. and Burl, J.B., “Effects of Time Horizon on Model Predictive Control for Hybrid Electric Vehicles,” IFAC-PapersOnLine 48:252-256, 2015.

65. Rezaei, A. and Burl, J.B., “Prediction of Vehicle Velocity for Model Predictive Control,” IFAC-PapersOnLine 48:257-262, 2015.

66. Changjian, H., Gao, Y., and Alex, Q.H., “Power Management Strategy of Hybrid Electric Vehicles Based on Particle Swarm Optimization,” in 2015 IEEE Transportation Electrification Conference and Expo (ITEC), Metro Detroit, MI, 2015, 1-6.

67. Hu, C., Huang, A.Q., and Gao, Y., “Comprehensive Lost Minimization Strategy for Parallel Plug-in Hybrid Electric Vehicles,” in 2013 IEEE Transportation Electrification Conference and Expo (ITEC), Dearborn/Detroit, MI, 2013, 1-7.

68. Gong, Q., Li, Y., and Peng, Z.R., “Trip-Based Optimal Power Management of Plug-in Hybrid Electric Vehicles,” IEEE Transactions on Vehicle Technology 57:3393-3401, 2008.



http://dx.doi.org/10.4271/2004-01-0064

http://dx.doi.org/10.4271/2004-01-0064


http://dx.doi.org/10.4271/2017-26-0106

ieeexplore.ieee.org

ieeexplore.ieee.org

ieeexplore.ieee.org



69. Gong, Q., Li, Y., and Peng, Z., “Power Management of Plug-in Hybrid Electric Vehicles Using Neural Network Based Trip Modeling,” in 2009 American Control Conference, St. Louis, MO, 2009, 4601-4606, ieeexplore.ieee.org.

70. Bender, F.A., Kaszynski, M., and Sawodny, O., “Drive Cycle Prediction and Energy Management Optimization for Hybrid Hydraulic Vehicles,” IEEE Transactions on Vehicle Technology 62:3581-3592, 2013.

71. Mohd Zulkefli, M.A., Zheng, J., Sun, Z., and Liu, H.X., “Hybrid Powertrain Optimization with Trajectory Prediction Based on Inter-Vehicle-Communication and Vehicle-Infrastructure-Integration,” Transportation Research Part C: Emerging Technologies 45:41-63, 2014/8

72. Sun, C., Hu, X., Moura, S.J., and Sun, F., “Velocity Predictors for Predictive Energy Management in Hybrid Electric Vehicles,” IEEE Transactions on Control Systems Technology 23:1197-1204, 2015.

73. Sun, C., Moura, S.J., Hu, X., Hedrick, J.K. et al., “Dynamic Traffic Feedback Data Enabled Energy Management in Plug-in Hybrid Electric Vehicles,” IEEE Transactions on Control Systems Technology 23:1075-1086, 2015.

74. Baker, D., Asher, Z., and Bradley, T., “Investigation of Vehicle Speed Prediction from Neural Network Fit of Real World Driving Data for Improved Engine On/Off Control of the EcoCAR3 Hybrid Camaro,” SAE Technical Paper 2017-01-1262, 2017, doi:10.4271/2017-01-1262.

75. Baker, D., Asher, Z.D., and Bradley, T., “V2V Communication Based Real-World Velocity Predictions for Improved HEV Fuel Economy,” SAE Technical Paper 2018-01-1000, 2018, doi:10.4271/2018-01-1000.

76. Tunnell, J.A., Asher, Z.D., Pasricha, S., and Bradley, T.H., “Towards Improving Vehicle Fuel Economy with ADAS,” SAE Technical Paper 2018-01-0593, 2018, doi:10.4271/2018-01-0593.

77. Asher, Z.D., Tunnell, J.A., Baker, D.A., Fitzgerald, R.J. et al., “Enabling Prediction for Optimal Fuel Economy Vehicle Control,” SAE Technical Paper 2018-01-1015, 2018, doi:10.4271/2018-01-1015.

78. Sun, C., Sun, F., and He, H., “Investigating Adaptive-ECMS with Velocity Forecast Ability for Hybrid Electric Vehicles,” Applied Energy 185(Part 2):1644-1653, 2017.

79. Onori, S., Serrao, L., and Rizzoni, G., “Adaptive Equivalent Consumption Minimization Strategy for Hybrid Electric Vehicles,” in ASME 2010 Dynamic Systems and Control Conference, Cambridge, MA, American Society of Mechanical Engineers, 2010, 499-505.

80. Liu, H., Li, X., Wang, W., Han, L. et al., “Markov Velocity Predictor and Radial Basis Function Neural Network-Based Real-Time Energy Management Strategy for Plug-in Hybrid Electric Vehicles,” Energy 152:427-444, 2018.

81. Zhang, Y., Chu, L., Fu, Z., Xu, N. et al., “Optimal Energy Management Strategy for Parallel Plug-in Hybrid Electric Vehicle Based on Driving Behavior Analysis and Real Time Traffic Information Prediction,” Mechatronics 46:177-192, 2017.

82. Shen, P., Zhao, Z., Zhan, X., Li, J. et al., “Optimal Energy Management Strategy for a Plug-in Hybrid Electric Commercial Vehicle Based on Velocity Prediction,” Energy 155:838-852, 2018.

83. Qi, X., Luo, Y., Wu, G., Boriboonsomsin, K. et al., “Deep Reinforcement Learning Enabled Self-Learning Control for Energy Efficient Driving,” Transportation Research Part C: Emerging Technologies 99:67-81, 2019.

84. Liu, K., Asher, Z., Gong, X., Huang, M. et al., “Vehicle Velocity Prediction and Energy Management Strategy Part 1: Deterministic and Stochastic Vehicle Velocity Prediction Using Machine Learning,” SAE Technical Paper 2019-01-1051, 2019, doi:10.4271/2019-01-1051.

85. Gaikwad, T.D., Asher, Z.D., Liu, K., Huang, M. et al., “Vehicle Velocity Prediction and Energy Management Strategy Part 2: Integration of Machine Learning Vehicle Velocity Prediction with Optimal Energy Management to Improve Fuel Economy,” SAE Technical Paper 2019-01-1212, 2019, doi:10.4271/2019-01-1212.

86. O’Keefe, M.P. and Markel, T., Dynamic Programming Applied to Investigate Energy Management Strategies for a Plug-in HEV (Colorado, USA: National Renewable Energy Laboratory, 2006).

87. Fu, L., Zgüner, Ü.Ö., Tulpule, P., and Marano, V., “Real-Time Energy Management and Sensitivity Study for Hybrid Electric Vehicles,” in Proceedings of the 2011 American Control Conference, San Francisco, CA, 2011, 2113-2118, ieeexplore.ieee.org

88. He, Y., Chowdhury, M., Pisu, P., and Ma, Y., “An Energy Optimization Strategy for Power-Split Drivetrain Plug-in Hybrid Electric Vehicles,” Transportation Research Part C: Emerging Technologies 22:29-41, 2012/6.

89. Opila, D.F., Wang, X., McGee, R., Brent Gillespie, R. et al., “Real-World Robustness for Hybrid Vehicle Optimal Energy Management Strategies Incorporating Drivability Metrics,” Journal of Dynamic Systems, Measurement and Control 136:061011, 2014.

90. Opila, D.F., “Uncertain Route, Destination, and Traffic Predictions in Energy Management for Hybrid, Plug-in, and Fuel-Cell Vehicles,” in 2016 American Control Conference (ACC), Seattle, WA, 2016, 1685-1692, ieeexplore.ieee.org.

91. Cummings, T., Bradley, T.H., and Asher, Z.D., “The Effect of Trip Preview Prediction Signal Quality on Hybrid Vehicle Fuel Economy,” IFAC-PapersOnLine 48:271-276, 2015.

92. Asher, Z.D., Cummings, T., and Bradley, T.H., “The Effect of Hill Planning and Route Type Identification Prediction Signal Quality on Hybrid Vehicle Fuel Economy,” SAE Technical Paper 2016-01-1240, 2016, doi:10.4271/2016-01-1240.

93. Trinko, D.A., Asher, Z.D., and Bradley, T.H., “Application of Pre-Computed Acceleration Event Control to Improve Fuel Economy in Hybrid Electric Vehicles,” SAE Technical Paper 2018-01-0997, 2018, doi:10.4271/2018-01-0997.

94. Zulkefli, M.A.M. and Sun, Z., “Real-Time Powertrain Optimization Strategy for Connected Hybrid Electrical Vehicle,” in ASME 2016 Dynamic Systems and Control Conference, Pittsburgh, PA, American Society of Mechanical Engineers, 2016, V002T20A006-V002T20A006.

95. Chen, Z., Xiong, R., Wang, C., and Cao, J., “An On-Line Predictive Energy Management Strategy for Plug-in Hybrid Electric Vehicles to Counter the Uncertain Prediction of the Driving Cycle,” Applied Energy 185:1663-1672, 2017.


ieeexplore.ieee.org



http://dx.doi.org/10.4271/2017-01-1262



http://dx.doi.org/10.4271/2018-01-1000


http://dx.doi.org/10.4271/2018-01-0593


http://dx.doi.org/10.4271/2018-01-1015



http://dx.doi.org/10.4271/2019-01-1051


http://dx.doi.org/10.4271/2019-01-1212

ieeexplore.ieee.org

ieeexplore.ieee.org


http://dx.doi.org/10.4271/2016-01-1240


http://dx.doi.org/10.4271/2018-01-0997

© 2019 SAE International. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE International.

Positions and opinions advanced in this work are those of the author(s) and not necessarily those of SAE International. Responsibility for the content of the work lies solely with the author(s).


96. Yang, S., Wang, W., Zhang, F., Hu, Y. et al., “Driving-Style-Oriented Adaptive Equivalent Consumption Minimization Strategies for HEVs,” IEEE Transactions on Vehicle Technology 67:9249-9261, 2018.

97. Alzorgan, M., Carroll, J., Al-Masalmeh, E., and Mayyas, A.R.T., “Look-Ahead Information Based Optimization Strategy for Hybrid Electric Vehicles,” SAE Technical Paper 2016-01-2226, 2016, doi:10.4271/2016-01-2226.

98. Lin, C.-C., Peng, H., Grizzle, J.W., Liu, J. et al., “Control System Development for an Advanced-Technology Medium-Duty Hybrid Electric Truck,” SAE Technical Paper 2003-01-3369, 2003, doi:10.4321/2003-01-3369.

99. MohdZulkefli, M.A., Mukherjee, P., Sun, Z., Zheng, J. et al., “Hardware-in-the-Loop Testbed for Evaluating Connected Vehicle Applications,” Transportation Research Part C: Emerging Technologies 78:50-62, 2017/5.

100. Peng, J., He, H., and Xiong, R., “Rule Based Energy Management Strategy for a Series-Parallel Plug-in Hybrid Electric Bus Optimized by Dynamic Programming,” Applied Energy 185(Part 2):1633-1643, 2017.

101. arpa-e, “NEXTCAR Program Overview,” 2019, https://arpa-e.energy.gov/sites/default/files/documents/files/NEXTCAR_ProgramOverview.pdf, accessed June 1, 2019.

102. Paganelli, G., Tateno, M., Brahma, A., Rizzoni, G. et al., “Control Development for a Hybrid-Electric Sport-Utility Vehicle: Strategy, Implementation and Field Test Results,” in Proceedings of the 2001 American Control Conference (Cat. No.01CH37148), Arlington, VA, 2001, 5064-5069, vol. 6, ieeexplore.ieee.org

103. Rizzoni, G. and Onori, S., “Energy Management of Hybrid Electric Vehicles: 15 years of development at the Ohio State University,” Oil & Gas Science and Technology - Revue d’IFP Energies nouvelles 70:41-54, 2015.

104. Olin, P., Aggoune, K., Tang, L., Confer, K. et al., “Reducing Fuel Consumption by Using Information from Connected and Automated Vehicle Modules to Optimize Propulsion System Control,” SAE Technical Paper 2019-01-1213, 2019, doi:10.4271/2019-01-1213.

105. Liu, J., Hagena, J., Peng, H., and Filipi, Z.S., “Engine-in-the-Loop Study of the Stochastic Dynamic Programming Optimal Control Design for a Hybrid Electric HMMWV,” International Journal of Heavy Vehicle Systems 15:309-326, 2008.

106. Opila, D.F., Wang, X., McGee, R., and Grizzle, J.W., “Real-Time Implementation and Hardware Testing of a Hybrid Vehicle Energy Management Controller Based on Stochastic Dynamic Programming,” Journal of Dynamic Systems and Measurement Control 135:021002, 2013.

107. Chasse, A., Pognant-Gros, P., and Sciarretta, A., “Online Implementation of an Optimal Supervisory Control for a

Parallel Hybrid Powertrain,” SAE Int. J. Engines 2:1630-1638, 2009, doi:10.4271/2009-01-1868.

108. Chasse, A. and Sciarretta, A., “Supervisory Control of Hybrid Powertrains: An Experimental Benchmark of Offline Optimization and Online Energy Management,” Control Engineering Practice 19:1253-1265, 2011.

109. Grondin, O., Thibault, L., Moulin, P., Chasse, A. et al., “Energy Management Strategy for Diesel Hybrid Electric Vehicle,” in 2011 IEEE Vehicle Power and Propulsion Conference, Chicago, IL, 2011, 1-8, ieeexplore.ieee.org.

110. Grondin, O., Thibault, L., and Quérel, C., “Energy Management Strategies for Diesel Hybrid Electric Vehicle,” Oil & Gas Science and Technology - Revue d’IFP Energies nouvelles 70:125-141, 2015.

111. van Keulen, T., van Mullem, D., de Jager, B., Kessels, J.T.B.A. et al., “Design, Implementation, and Experimental Validation of Optimal Power Split Control for Hybrid Electric Trucks,” Control and Engineering Practise 20:547-558,2012/5.

112. Wang, L., Zhang, Y., Yin, C., Zhang, H. et al., “Hardware-in-the-Loop Simulation for the Design and Verification of the Control System of a Series-Parallel Hybrid Electric City-Bus,” Simulation Modelling Practice and Theory 25:148-162, 2012/6.

113. Pagliara, E., Parlangeli, G., Donateo, T., and Adamo, F., “Real Time Implementation of an Optimal Power Management Strategy for a Plug-in Hybrid Electric Vehicle,” in 52nd IEEE Conference on Decision and Control, Florence, Italy, 2013, 2214-2219, ieeexplore.ieee.org.

114. Du, Y., Zhao, Y., Wang, Q., Zhang, Y. et al., “Trip-Oriented Stochastic Optimal Energy Management Strategy for Plug-in Hybrid Electric Bus,” Energy 115:1259-1271, 2016.

115. Mayyas, A.R., Kumar, S., Pisu, P., Rios, J. et al., “Model-Based Design Validation for Advanced Energy Management Strategies for Electrified Hybrid Power Trains Using Innovative Vehicle Hardware in the Loop (VHIL) Approach,” Applied Energy 204:287-302, 2017.

116. Xie, S., He, H., and Peng, J., “An Energy Management Strategy Based on Stochastic Model Predictive Control for Plug-in Hybrid Electric Buses,” Applied Energy 196:279-288, 2017.

117. Guo, N., Shen, J., Xiao, R., Yan, W. et al., “Energy Management for Plug-in Hybrid Electric Vehicles Considering Optimal Engine ON/OFF Control and Fast State-of-Charge Trajectory Planning,” Energy 163:457-474, 2018.

118. Rama, N., Wang, H., Orlando, J., Robinette, D. et al., “Route-Optimized Energy Management of Connected and Automated Multi-Mode Plug-In Hybrid Electric Vehicle Using Dynamic Programming,” SAE Technical Paper 2019-01-1209, 2019, doi:10.4271/2019-01-1209.



http://dx.doi.org/10.4271/2016-01-2226



http://dx.doi.org/10.4321/2003-01-3369

https://arpa-e.energy.gov/sites/default/files/documents/files/NEXTCAR_ProgramOverview.pdf



ieeexplore.ieee.org


http://dx.doi.org/10.4271/2019-01-1213

http://dx.doi.org/10.4271/2009-01-1868

http://ieeexplore.ieee.org

http://ieeexplore.ieee.org



http://dx.doi.org/10.4271/2019-01-1209

identification and review of the research gaps preventing a ......2002/08/08 · identified...

Documents