Conferences

2026

1. Real-Time Parameter Optimization for Eco-driving Control in Connected and Automated Vehicles Using Reinforcement Learning

   Yaozhong Zhang, Rami Ammourah, Jihun Han, Ayman Moawad, Daliang Shen, and Dominik Karbowski

   In WCX SAE World Congress Experience, Apr 2026

   Abs DOI

   This paper introduces a novel methodology to enhance the energy efficiency of eco-driving controllers in Connected and Automated Vehicles (CAVs) by leveraging reinforcement learning (RL) techniques for real-time parameter optimization. Traditional eco-driving strategies rely on fixed control parameters, which limit adaptability across diverse traffic and road conditions. To address this, we apply continuous action space RL algorithms, specifically Deep Deterministic Policy Gradient (DDPG) and Proximal Policy Optimization (PPO), to dynamically tune four key parameters within a model predictive control framework that is grounded in Pontryagin’s Maximum Principle (PMP). These parameters influence acceleration, braking, cruising, and intersection-approach behaviors, making them critical for achieving optimal eco-driving performance. Our study employs Argonne National Laboratory’s RoadRunner simulator, a Simulink-based environment designed for high-fidelity CAV analysis, incorporating realistic traffic signals, road gradients, and vehicle interactions. RL agents are trained to interpret vehicle states, road attributes, and traffic light information to adjust control parameters in real time. This integration enables the controller to anticipate and respond to dynamic driving scenarios, thereby improving both energy efficiency and operational robustness. Simulation experiments across multiple driving scenarios demonstrate that the RL-enhanced eco-driving controller achieves substantial energy savings without compromising travel time. On average, our approach surpasses a baseline eco-driving controller without RL by 12% and outperforms a high-fidelity human driver model by 24.2% in terms of energy consumption reduction. These results highlight the potential of continuous action space RL to advance real-time eco-driving control in CAVs. Overall, this work provides a pathway toward more intelligent, adaptive, and sustainable vehicle control systems that can accelerate the deployment of energy-efficient mobility solutions.

2025

1. Advancing Connected Vehicle Capabilities with Scalable and Flexible Anything-In-The-Loop Testing Framework

   Jihun Han, Miriam Di Russo, Debashis Das, Qian Peng, Jongryeol Jeong, Michael Pamminger, Priyash Misra, Dominik Karbowski, and 2 more authors

   In 2025 IEEE/AIAA Transportation Electrification Conference and Electric Aircraft Technologies Symposium (ITEC+EATS), Jun 2025

   Abs DOI

   Connected vehicle technologies have the potential to improve safety and energy efficiency. However, an efficient and robust testing framework that bridges simulations and realworld conditions is required to accelerate their validation and deployment. This paper demonstrates a streamlined approach to validating and testing connected vehicles through simulations, MicroAutobox-in-the-loop (MABx-IL) tests, MABx-vehicle-in-the-loop (MABx-VIL), and Human-MABx-VIL tests. The demonstration leverages our anything-in-the-loop (XIL) framework, which interfaces various components (e.g., simulation models, MABx, vehicle, human drivers) to evaluate connected vehicle performance in a simulated Chicago urban corridor. Our automatic model-building process has validated its flexibility, scalability, and consistency, enabling stage-by-stage progressive integration of physical components in XIL testing. Using target electric vehicle (Hyundai Ioniq 5), the Human-MABx-VIL study validated eco-speed advisory driving through a phone-based humanmachine interface (HMI), with drivers successfully performing eco-approaches to virtual signalized intersections along the road. This effort was preceded by extensive simulations, MABx-IL, and MABx-VIL, progressively swapping virtual components within the XIL framework. It underscores the necessity of streamlined validation and testing and highlights the potential of connected eco-speed advisory driving using real-time vehicle-to-everything (V2X) information.

2024

1. Assessing the Energy Impacts of Connected and Automated Vehicles: A Comprehensive Validation via Simulation, Dynamometer, and Track Tests

   Jongryeol Jeong, Miriam Di Russo, Ahammad Basha Dudekula, Jihun Han, Kevin Stutenberg, Jeffrey D. Naber, and Dominik Karbowski

   In 2024 IEEE International Automated Vehicle Validation Conference (IAVVC), Oct 2024

   Abs DOI

   Connected and automated vehicles (CAVs) have the potential to enhance energy efficiency by leveraging advanced environmental data, a capability beyond the reach of traditional vehicles. Nevertheless, quantifying the precise energy impacts of CAVs remains challenging due to their complex interactions with dynamic environments. This challenge is further underscored by conventional simulations and experiments that utilize predefined speed profiles to emulate vehicle behavior. To address these gaps, our study adopts a comprehensive approach involving simulations, dynamometer usage, and track tests to assess the energy effects of CAVs more accurately. Our simulation method integrates sophisticated environmental models, such as road regulations and surrounding traffic, with both human and automated driver models equipped with an eco-driving control algorithm. This algorithm adjusts vehicle operations based on real-time environmental inputs, while the vehicle’s powertrain model provides feedback, creating a closed-loop system that mirrors real-world conditions. The eco-driving algorithm is further evaluated in the dynamometer and track tests to confirm its effectiveness and to examine the vehicle’s behavior under real conditions. By integrating “anything-in-the-loop” (XIL) workflows, our approach enables real-time interaction between the vehicle and simulated environments during these tests, thus avoiding the limitations of the predefined speed profiles used in traditional experiments. This paper primarily focuses on comparing these methodologies to thoroughly investigate and highlight their distinct characteristics, providing insights into factors that influence discrepancies in energy consumption and the accuracy of energy impact assessments. Our findings validate the reliability of our method, revealing less than a 5% discrepancy between simulation predictions and actual test outcomes under various conditions. Ultimately, this comprehensive evaluation not only confirms the effectiveness of the proposed simulation model but also offers invaluable insights for developing more energy-efficient CAV technologies.

2. Energy Savings Impact of Eco-Driving Control Based on Powertrain Characteristics in Connected and Automated Vehicles: On-Track Demonstrations

   Jongryeol Jeong, Elangovan Kandaswamy, Ahammad Basha Dudekula, Jihun Han, Dominik Karbowski, and Jeffrey Naber

   In WCX SAE World Congress Experience, Apr 2024

   Abs DOI

   This research investigates the energy savings achieved through eco-driving controls in connected and automated vehicles (CAVs), with a specific focus on the influence of powertrain characteristics. Eco-driving strategies have emerged as a promising approach to enhance efficiency and reduce environmental impact in CAVs. However, uncertainty remains about how the optimal strategy developed for a specific CAV applies to CAVs with different powertrain technologies, particularly concerning energy aspects. To address this gap, on-track demonstrations were conducted using a Chrysler Pacifica CAV equipped with an internal combustion engine (ICE), advanced sensors, and vehicle-to-infrastructure (V2I) communication systems, compared with another CAV, a previously studied Chevrolet Bolt electric vehicle (EV) equipped with an electric motor and battery. The implemented control is a universal speed planner that solves the eco-driving optimal-control problem within a receding-horizon framework, utilizing V2I communications for signal phase and timing information. The controller calculates accelerator and brake pedal positions using the vehicle’s state and real-time environmental information. Both the Pacifica, target vehicle, and the Bolt, EV, are equipped with a drive-by-wire system. The experiments encompass five road scenarios repeated three times, covering a 3.7-km track with various stop signs, traffic signals, and speed limits. Three control calibrations are employed to represent human-driver-like, non-connected automated, and V2I-connected driving. First and foremost, the results demonstrate functional eco-driving controls with no extreme acceleration or traffic law violations in the Pacifica (ICE vehicle). Energy savings of up to 6% without connectivity and up to 22% with V2I connectivity are achieved in the ICE vehicle as well. Additionally, a comparison is made between an ICE vehicle and an EV to analyze the energy-saving impacts of eco-driving controls across different powertrain characteristics. In conclusion, this study emphasizes the significance of correlating powertrain design with controls and eco-driving strategies during the development of CAVs.

3. Testing Cellular Vehicle-to-Everything Communication Performance and Feasibility in Automated Vehicles\textsuperscript*

   Zhaohui Liang, Jihun Han, Xiaopeng Li, Dominik Karbowski, Chengyuan Ma, and Aymeric Rousseau

   In 2024 IEEE Intelligent Vehicles Symposium (IV), Jun 2024

   Abs DOI

   Many studies have demonstrated the eco-driving capabilities of connected and automated vehicles (CAVs) to significantly enhance mobility systems. The majority of these studies have been conducted using simulations, which fail to capture the effects of practical uncertainties encountered in vehicle-to-anything (V2X) communications. In this paper, we investigated the performance of current cellular V2X (C-V2X) communications through systematic testing and provided a quantitative analysis of key performance indices (e.g., inter-packet gap and packet error rate) across various test scenarios. As one use case to demonstrate the benefits of C-V2X communication on the road, we tested the feasibility of eco-driving for a SAE level 3 (L3) automated vehicle (AV) communicating with a connected urban corridor capable of transmitting traffic light information (i.e., signal phase and timing). To achieve this, we implemented the eco-speed planning algorithm at a high-level in the AV control software system and ensured its interactions with other existing low-level control algorithms, as well as the C-V2X onboard unit. Finally, we experimentally demonstrated eco-driving of the L3 CAV on a scaled-down corridor with two signal-controlled intersections, revealing the AV’s ability to maintain smoother trajectories and avoid unnecessary stops compared to human-driven vehicles.

4. Is It Necessary to Calibrate All Parameters for Each Driver?

   Yanbing Wang, Felipe De Souza, Jihun Han, and Dominik Karbowski

   IFAC-PapersOnLine, vol. 58, no. 28, pp. 696–701, Oct 2024

   Abs DOI

   Learning the variability of driver behavior can help with understanding the driver heterogeneity and traffic patterns, designing driver-specific vehicle automation and control, and improving the accuracy of micro-simulation tools. However, even understanding the driving behavior of a small population can be challenging, due to the large number of total parameters that need to be calibrated. This study investigates whether individually calibrating driver models provides a more accurate description of the population or if it is inherently “overparameterized”. We propose an approach to analyze calibration performance using various reduced-order versions of a car-following model based on the Optimal Velocity model (OVM). We explore which parameters can be eliminated from calibration without sacrificing overall accuracy and how reducing the number of calibrated parameters impacts the representation of driver heterogeneity. Our preliminary results indicate that while reduced-parameter models compromise accuracy, the extent varies depending on which parameters are fixed first. Furthermore, fixing one parameter alters the distribution of other parameters, suggesting possible dependencies among these model parameters.

2023

1. An X-in-the-Loop (XIL) Testing Framework for Validation of Connected and Autonomous Vehicles

   Prakhar Gupta, Rongyao Wang, Tyler Ard, Jihun Han, Dominik Karbowski, Ardalan Vahidi, and Yunyi Jia

   In 2023 IEEE International Automated Vehicle Validation Conference (IAVVC), Oct 2023

   Abs DOI

   Validation methods for growing connected and autonomous vehicle (CAV) technologies hold the key to their real-world implementation. Setting these up is challenging owing to the number of components and safety concerns involved in CAV testing. This paper presents a modular XIL framework for the validation of CAVs. The framework allows the interfacing of an in-house instrumented drive-by-wire vehicle with roads, traffic microsimulations (virtual vehicles, traffic lights, etc.), autonomy software, computing resources, and human drivers through the use of a mixed-reality headset. Three case studies in recent and ongoing work are shown, demonstrating the suitability and versatility of the XIL framework for verifying the safety and efficiency of automated vehicle control strategies via vehicle-in-the-loop, as well as ensuring realistic human driver reactions inside of emergent mixed-autonomy traffic environments via humans-in-the-loop.

2. Model Validation of Adaptive Cruise Control in Vehicles Utilizing Real-World Driving Data

   Yaozhong Zhang, Jihun Han, Namdoo Kim, and Dominik Karbowski

   In 2023 IEEE International Automated Vehicle Validation Conference (IAVVC), Oct 2023

   Abs DOI

   Adaptive Cruise Control (ACC) is one of the most significant Advanced Driver Assistance System (ADAS) features. To fully comprehend and quantitatively analyze the impact of the on-road automation on vehicles and transportation systems, it is necessary to validate the models of ACC systems in currently available vehicles. This paper presents a workflow for the modeling and validation of the stock ACC system on the 2018 Cadillac CT6 with the Super Cruise system. The approach makes use of the real-world test-driving data collected by the instrumented test vehicle, analyzes its behavior in various driving scenarios, formulates the ACC system model based on the data, and conducts validation with both simplified vehicle kinetics and the high-fidelity Cadillac CT6 powertrain model by using the energy-focused connected and automated vehicle (CAV) simulator developed by Argonne National Laboratory.

2022

1. Analytical Anticipative Optimal Drivability Car-Following Model

   Jihun Han, Dominik Karbowski, and Aymeric Rousseau

   In 2022 American Control Conference (ACC), Jun 2022

   Abs DOI

   Replicating a human car-following behavior becomes more important for developing either a human driver model or human-like adaptive cruise control. The human driver model is especially a key element in simulation tools used for the development and evaluation of connected and automated vehicle driving controls. In this paper, we propose an analytical anticipative optimal drivability car-following model that can capture a dynamic car-following behavior while maximizing driving comfort without collisions in a computational-efficient way. We formulate drivability-oriented car-following as an optimal control problem, and then reformulate it to a bi-level optimization problem in order to facilitate analytical treatment. By employing optimal control theory, we can transform the bi-level optimization problem into a nonlinear programming problem and derive its analytical solutions. To validate the proposed model, we post-processed and used Next Generation SIMulation (NGSIM) data. Results show that the proposed parametric model can generate stable car-following behaviors and its vehicle state trajectories are well matched with NGSIM data, thereby significantly improving root-mean-square error of speed and distance gap compared to the existing car-following model.

2. Potential Energy Saving of V2V-Connected Vehicles in Large-Scale Traffic

   Eunjeong Hyeon, Jihun Han, Daliang Shen, Dominik Karbowski, Namwook Kim, and Aymeric Rousseau

   IFAC-PapersOnLine, vol. 55, no. 24, pp. 78–83, Aug 2022

   Abs DOI

   Most studies evaluating the energy efficiency of connected and automated vehicles (CAVs) in car-following scenarios have considered a few preceding vehicles communicating with the controlled CAVs. However, considering rapidly evolving technologies in CAVs, extended vehicle-to-vehicle (V2V) connectivity over large-scale traffic needs to be considered in estimating CAVs’ energy benefits. This paper investigates the potential energy saving of V2V-connected vehicles in large-scale downstream traffic by adopting a human driver model generating stable car-following trajectories for many consecutive vehicles. The energy-efficient driving of a CAV is demonstrated based on an optimal controller minimizing the longitudinal acceleration by forecasting an immediately preceding vehicle’s trajectory over a fixed prediction horizon. Various traffic scenarios are considered by applying different simulation parameters, including the distribution of vehicle time gaps, the number of connected vehicles, and prediction horizon lengths. Furthermore, a comprehensive analysis is conducted to discover the relationships between the parameters of interest and system performance, including prediction and control. Our findings from the parameter study are validated by evaluating the realistic energy consumption of a vehicle in a simulation platform operating high-fidelity powertrain models.

3. Vehicle-In-The-Loop Workflow for the Evaluation of Energy-Efficient Automated Driving Controls in Real Vehicles

   Jongryeol Jeong, Dominik Karbowski, Namdoo Kim, Jihun Han, Kevin Stutenberg, Miriam Di Russo, and Julien Grave

   In WCX SAE World Congress Experience, Mar 2022

   Abs DOI

   This paper introduces a new systematic workflow for the rapid evaluation of energy-efficient automated driving controls in real vehicles in controlled laboratory conditions. This vehicle-in-the-loop (VIL) workflow, largely standardized and automated, is reusable and customizable, saves time and minimizes costly dynamometer time. In the first case study run with the VIL workflow, an automated car driven by an energy-efficient driving control previously developed at Argonne used up to 22 % less energy than a conventional control. In a VIL experiment, the real vehicle, positioned on a chassis dynamometer, has a digital twin that drives in a virtual world that replicates real-life situations, such as approaching a traffic signal or following other vehicles. The real and virtual systems interact in a close-loop fashion: the automated driving control directs accelerator and brake pedals based on measurements from the real vehicle and from the perception of the digital twin’s surrounding virtual environment; the resulting speed of the vehicle is fed back to the virtual world to compute the position of the digital twin. The VIL workflow provides a systematic linkage between the virtual environment, the hardware and software that interact with the vehicle and the dynamometer, as well as processes that facilitate scenario setup, code generation, experimentation, and data collection. Argonne’s RoadRunner, a simulation tool dedicated to the energy-focused study of connected and automated vehicles, serves as the virtual environment and is the backbone of the workflow. The real vehicle is tied to the chassis dynamometer, and a robotic driver actuates the accelerator and brake pedals based on the demands from the automated driving controls. During experimentation, the virtual environment, the data acquisition, the automated driving controls, as well as the low-level controls are run on a real-time system (dSPACE’s MicroAutoBox).

4. Data-Driven Design of Model Predictive Control for Powertrain-Aware Eco-Driving Considering Nonlinearities Using Koopman Analysis

   Daliang Shen, Jihun Han, Dominik Karbowski, and Aymeric Rousseau

   IFAC-PapersOnLine, vol. 55, no. 24, pp. 117–122, Aug 2022

   Abs DOI

   Eco-driving is a highly nonlinear control problem. The nonlinearities include the complex energy conversion/dissipation in the powertrain, environmental influences such as road grade and aerodynamic drag, constraints due to traffic signs, safety issues, and physical limits of the vehicle system. In recent years, researchers have increasingly revisited the Koopman operator to linearize nonlinear dynamics. This paper adopts such an approximation technique to construct the lifted state space in a data-driven procedure that allows us to incorporate nonlinearities and system perturbations in the cost function. In addition, the nonlinear constraints in states can also be handled linearly. The resultant formulation of a linearly constrained quadratic program can be readily applied to design a model predictive control that enjoys a low computation load as with a linear dynamic system. Simulation results demonstrate additional energy saving potential compared to a linear approach.

2021

1. A Real-Time Intelligent Speed Optimization Planner Using Reinforcement Learning

   Woong Lee, Jihun Han, Yaozhong Zhang, Dominik Karbowski, Aymeric Rousseau, and Namwook Kim

   In SAE WCX Digital Summit, Apr 2021

   Abs DOI

   As connectivity and sensing technologies become more mature, automated vehicles can predict future driving situations and utilize this information to drive more energy-efficiently than human-driven vehicles. However, future information beyond the limited connectivity and sensing range is difficult to predict and utilize, limiting the energy-saving potential of energy-efficient driving. Thus, we combine a conventional speed optimization planner, developed in our previous work, and reinforcement learning to propose a real-time intelligent speed optimization planner for connected and automated vehicles. We briefly summarize the conventional speed optimization planner with limited information, based on closed-form energy-optimal solutions, and present its multiple parameters that determine reference speed trajectories. Then, we use a deep reinforcement learning (DRL) algorithm, such as a deep Q-learning algorithm, to find the policy of how to adjust these parameters in real-time to dynamically changing situations in order to realize the full potential of energy-efficient driving. The model-free DRL algorithm, based on the experience of the system, can learn the optimal policy through iteratively interacting with different driving scenarios without increasing the limited connectivity and sensing range. The training process of the parameter adaptation policy exploits a high-fidelity simulation framework that can simulate multiple vehicles with full powertrain models and the interactions between vehicles and their environment. We consider intersection-approaching scenarios where there is one traffic light with different signal phase and timing setup. Results show that the learned optimal policy enables the proposed intelligent speed optimization planner to properly adjust the parameters in a piecewise constant manner, leading to additional energy savings without increasing total travel time compared to the conventional speed optimization planner.

2. Assessing the Implications of Automated Merging Control in a Mixed and Heterogeneous Traffic Environment

   Jackeline Rios-Torres, Zulqarnain Khattak, Jihun Han, Chieh Wang, and Hyeonsup Lim

   In 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), Sep 2021

   Abs DOI

   Previous efforts to explore the implications of partial market penetration of connected and automated vehicles (CAVs) show a consensus on the benefits of higher market penetration rates (MPR) of vehicles enabled with connectivity and/or automation. There is, however, a level of uncertainty regarding the effects of lower market penetration rates and the consideration of heterogeneous vehicle fleets. Using VISSIM to perform microscopic traffic simulation and, vehicle simulation models, we assess the impacts of different CAVs market penetration rates on fuel consumption considering a heterogeneous traffic environment. The results show that the fuel efficiency benefits of optimal coordination control are maximized in moderate congested scenarios when the CAVs MPR exceeds 40%.

2020

1. Closed-Form Solutions for a Real-Time Energy-Optimal and Collision-Free Speed Planner with Limited Information

   Jihun Han, Dominik Karbowski, and Namdoo Kim

   In 2020 American Control Conference (ACC), Jul 2020

   Abs DOI

   Under real-world driving conditions, connected and automated vehicles (CAVs) must plan and follow an energy-optimal and collision-free speed trajectory with a high updating rate, based on available information limited by its communication range. This paper presents a speed planner using analytical closed-form optimal solutions. Using the simplest vehicle model, we derive closed-form solutions as functions of boundary conditions (BCs) and summarize them without and with pure state variable inequality constraints imposed by speed limits and the preceding vehicle. Then we introduce multiple driving modes (e.g., eco-approach to a traffic signal) for responding to dynamically changing situations and show how to set BCs for each mode while retaining the nature of globally optimal solutions. Finally, we perform a large-scale simulation study to identify and quantify the energy impacts of CAVs for real-world driving routes. A simple but effective planner based on closed-form solutions shows a significant energy saving potential compared with human-driven vehicles and adaptive cruise controlled vehicles with connectivity.

2. Fine-Tuning a Real-Time Speed Planner for Eco-Driving of Connected and Automated Vehicles

   Jihun Han, Woong Lee, Dominik Karbowski, Aymeric Rousseau, and Namwook Kim

   In 2020 IEEE Vehicle Power and Propulsion Conference (VPPC), Nov 2020

   Abs DOI

   A speed planner uses available information to enable automated vehicles to “eco-drive,” which includes eco-approach and departure at signalized intersections and leads to significant energy savings. Here we summarize the design of a proposed speed planner that generates multiple modes based on its analysis of optimal solutions, and we showcase the role of the parameters in the trajectory generation of each mode in the speed planner. For greater energy savings compared to human-driven vehicles, without sacrificing travel time, we optimally tune parameters through a global optimization method in a high-fidelity simulation framework. Finally, we analyze how the fine-tuning process makes driving more efficient and show its energy impacts.

3. State-Constrained Optimal Solutions for Safe Eco-Approach and Departure at Signalized Intersections

   Jihun Han, Dominik Karbowski, and Aymeric Rousseau

   In ASME 2020 Dynamic Systems and Control Conference, Oct 2020

   Abs DOI

   This paper provides fundamentals of how to energy-efficiently pass through signalized intersections while avoiding any rear-end collisions with leading vehicles. In our previous works [1, 2], analytical solutions with and without second-order pure state constraints imposed by the preceding vehicle were presented; these showed significant energy saving potential for connected and automated vehicles (CAVs) compared to human-driven vehicles. However, these solutions were derived assuming that the desired distance headway policy does not include a speed change over a predictive horizon, and that the preceding vehicle has constant acceleration. We use the desired time headway policy that includes the speed change to define the first-order pure state constraint. We then derive analytical solutions using the direct adjoining method based on Pontryagin’s minimum principle. We also present a novel solver to compute energy-optimal and collision-free state trajectories by accounting for a piecewise constant acceleration of the preceding vehicle without using any numerical optimization methods that require initial guesses. For simple scenarios with one intersection, we analyze how the novel solver allows CAVs to smoothly pass through the signalized intersection and then reach a desired cruising speed. We also use a simulation framework based on high-fidelity powertrain models to validate its effectiveness based on energy savings when driving on real-world urban routes.

4. Receding Horizon Reference Governor for Implementable and Optimal Powertrain-Aware Eco-Driving

   Daliang Shen, Jihun Han, Jongryeol Jeong, Dominik Karbowski, and Aymeric Rousseau

   IFAC-PapersOnLine, vol. 53, no. 2, pp. 13842–13849, Jul 2020

   Abs DOI

   This paper presents an adaptive, two-level control structure that makes it possible to implement a numerical optimization algorithm for eco-driving in real time. The reference governor in the higher level is characterized by a low, flexible sampling rate and adopts a receding horizon for preview and optimization. The optimization algorithm thereby finds the energy-minimizing solution, based on Pontryagin’s minimum principle (PMP), for traveling the selected route segments without colliding with the preceding vehicle. The tracking control in the lower level compensates for errors due to modeling inaccuracies and unmodeled disturbances. The hierarchical structure can accommodate different types of numerical solvers and control schemes that apply to various vehicle powertrain configurations. A large-scale simulation study using real-world route data with high-fidelity powertrain models validates the proposed control structure and its online implementation.

2018

1. Impact of Model Simplification on Optimal Control of Combustion Engine and Electric Vehicles Considering Control Input Constraints

   Jihun Han, Jackeline Rios-Torres, Ardalan Vahidi, and Antonio Sciarretta

   In 2018 IEEE Vehicle Power and Propulsion Conference (VPPC), Aug 2018

   Abs DOI

   This paper analyzes the impact of model simplification on optimal control for electric and conventional engine-powered vehicles. An optimal control problem is formulated to minimize energy/fuel consumption subject to control input constraints and solved analytically using the Pontryagin’s Minimum Principle. We found that simplified solutions without nonlinear aerodynamic drag are sub-optimal, and their loss of optimality increases with average travel speed. The energy/fuel savings of each control approach are evaluated using the intelligent driver model to capture driving behavior of a driver. Simulation case studies are presented for illustration.

2017

1. Handling State Constraints in Fast-computing Optimal Control for Hybrid Powertrains

   Jihun Han, Antonio Sciarretta, and Nicolas Petit

   IFAC-PapersOnLine, vol. 50, no. 1, pp. 4781–4786, Jul 2017

   Abs DOI

   To optimally design hybrid powertrains, optimal energy management strategies must be automatically and rapidly generated. Pontryagin’s minimum principle-derived optimization tool called Hybrid Optimization Tool (HOT) can guarantee the fast computing of minimal fuel consumption using an array operation as well as Picard’s method. However, in presence of state constraints (e.g., the battery state of charge limitations), the near-optimality of HOT no longer holds. Herein, we use the interior- and exterior-penalty method to impose the state constraints in HOT and highlight numerical difficulties encountered in their implementation. Then, a factor that causes the numerical difficulties is optimized by quantifying trade-off between the state constraints violation and computational demanding. Finally, through a case study of a parallel hybrid electric vehicle, the results show that despite of a complex problem with rapidly changing dynamics, the penalty methods are able to generate results comparable with dynamic programming ones while guaranteeing the low computational burden.

2. A Real-Time Eco-Driving Strategy for Automated Electric Vehicles

   Luis Leon Ojeda, Jihun Han, Antonio Sciarretta, Giovanni De Nunzio, and Laurent Thibault

   In 2017 IEEE 56th Annual Conference on Decision and Control (CDC), Dec 2017

   Abs DOI

   Over the past years, connected and automated vehicles (CAV) have become highly important in the transportation research field. Several prototypes are already introduced by established companies in cooperation with research centers. However, the crucial part of reducing their energy consumption by driving in an optimal way and facing external disturbances is sometimes overlooked. In this paper, we propose a safe- and eco-driving control system that enables the CAV to accelerate or to decelerate optimally while preventing both collision with preceding vehicle (i.e. disturbance) and violation of speed limitations. Optimal control problem (OCP) minimizing energy consumption for an electric vehicle while enforcing state constraints is formulated. Numerically, the problem is solved using a Model Predictive Control-like approach. The real-time implementation is possible thanks to the analytical solution of the state-constrained OCP. The proposed system is evaluated through a simulation for various driving scenarios, and it is shown that it can significantly reduce energy consumption compared to conventional driving while also avoiding the collision, without increasing arrival time.

2013

1. Impact of Hilly Road Profile on Optimal Energy Management Strategy for FCHEV with Various Battery Sizes

   Jihun Han, Youngjin Park, Dongsuk Kum, Seongpil Ryu, and Youn-sik Park

   In SAE/KSAE 2013 International Powertrains, Fuels & Lubricants Meeting, Oct 2013

   Abs DOI

   This study investigates how hilly road profiles affect the optimal energy management strategy for fuel cell hybrid electric vehicle (FCHEV) with various battery sizes. First, a simplified FCHEV model is developed to describe power and energy flows throughout the powertrain and evaluate hydrogen consumption. Then, an optimal control problem is formulated to find the globally optimal energy management strategy of FCHEV over driving cycles with road elevation profile. In order to solve the optimal energy management problem of the FCHEV, Dynamic Programming, a dynamic optimization method, is used, and their results are analyzed to find out how hilly road conditions affect the optimal energy management strategies. The results show that the optimal energy management with a smaller battery tends to actively prepare (e.g. pre-charge/pre-discharge) for uphill/downhill roads in order not to violate the battery state of charge (SoC) bounds. On the other hand, when the battery is large enough to handle a deep SoC swing due to hilly road profile, the optimal energy management strategy is not significantly affected by various battery sizes. In conclusion, when an energy management strategy is designed for FCHEV, the designer needs to utilize the road altitude information in order to achieve near-optimal fuel economy with charge-sustenance.

2012

1. A Novel Updating Method of Equivalent Factor in ECMS for Prolonging the Lifetime of Battery in Fuel Cell Hybrid Electric Vehicle

   Jihun Han, Youngjin Park, and Youn-sik Park

   IFAC Proceedings Volumes, vol. 45, no. 30, pp. 227–232, Oct 2012

   Abs DOI

   A novel updating method of equivalent factor of the Equivalent Consumption Minimization Strategy (ECMS) is proposed to prolong the lifetime of battery in fuel cell hybrid electric vehicle. ECMS can generate the sub-optimal solution at any instance for current driving condition. Equivalent factor which is dimensionless conversion ratio of electrical power flow into chemical power flow plays an important role in charge-sustaining performance. Well-tuned equivalent factor from iterative search in off-line can only guarantee the charge-sustaining performance, not the lifetime of the battery. Therefore additional updating method of equivalent factor is necessary in order to prolong the lifetime of the battery which only takes the numbers of charge and discharge cycle into account. Backward simulation is used to verify the effectiveness of the updating method.

2011

1. Adaptive Regenerative Braking Control in Severe Cornering for Guaranteeing the Vehicle Stability of Fuel Cell Hybrid Electric Vehicle

   Jihun Han, Youngjin Park, and Youn-sik Park

   In 2011 IEEE Vehicle Power and Propulsion Conference, Sep 2011

   Abs DOI

   Front-wheel-drive electric vehicle has only 1 electric motor which is connected to the front drive axle. With this system constraint, regenerative braking by using an electric motor can be only applied on front wheels symmetrically. Additional mechanical friction braking can be independently applied on each of all wheels using brake by wire such as EMB (Electro-Mechanical Brake). During severe cornering with braking, excessive regenerative braking force distribution to the front axle for improving the fuel economy can cause the vehicle to approach its handling limit, i.e., if the front tires saturate first, the vehicle may plow out of the curve and is called to limit understeer. When the vehicle comes in a danger of crossing the limit, adaptive regenerative braking controller engages, assisting the driver to guarantee the vehicle stability. Controller can distribute an optimal regenerative brake torque and additional mechanical friction brake torques in order to prevent loss of control for the vehicle. Carsim™ software is used to verify the effectiveness of the proposed controller.

2. Cooperative Regenerative Braking Control Strategy Considering Nonlinear Tire Characteristic in Front-Wheel-Drive Hybrid Electric Vehicle

   Jihun Han and Youngjin Park

   In 1st International Electric Vehicle Technology Conference, May 2011

   Abs DOI

   An electric motor for regenerative braking in front-wheel-drive hybrid electric vehicle is only connected to the front axle, and mechanical friction braking can be independently applied on each of the 4 wheels. Excessive regenerative braking only at front wheels to improve fuel economy can cause under-steer and eventually vehicle instability. Nonlinear tire characteristic may cause this vehicle instability in severe cornering with hard braking. Therefore, cooperative braking control strategy has to be considered nonlinear tire characteristic for guaranteeing the vehicle stability while enhancing the braking energy recovery. This paper is to compare the performance of cooperative braking control strategy according to consider the influence of braking force on the lateral force. Carsim™ software is used to evaluate the performance of cooperative regenerative braking control regarding to the vehicle stability and regenerative braking efficiency.