Bridge Zhao
Assistant Professor of Electrical and Computer Engineering Loyola Marymount University
- Los Angeles CA
Develops human-centered AI for safer, more sustainable mobility—from human behavior to intelligent vehicles and transportation networks.
Biography
His work combines machine learning, computer vision, robotics, control, optimization, digital twins, and multimodal data analytics. Using naturalistic driving data—including vehicle telemetry, video, and GPS—alongside human-in-the-loop simulation and field experiments, he studies driver behavior and human–automation interaction. Applications include personalized advanced driver-assistance systems (ADAS), driver intention and trajectory prediction, automation disengagement analysis, cooperative automated driving, and network-level traffic management. His research spans scales from individual behavior and vehicle trajectories to multi-vehicle systems, roadway networks, and system-level operations.
Before joining LMU, Dr. Zhao was a Postdoctoral Associate at the MIT AgeLab and Center for Transportation & Logistics and worked with the Advanced Vehicle Technology consortium, where he studied real-world use of ADAS and automated-driving technologies. He has also conducted research at Honda Research Institute and Toyota Motor North America InfoTech Labs and contributed to government- and industry-funded projects on connected and automated vehicles, intelligent transportation systems, and safety-critical AI.
Dr. Zhao received his Ph.D. in Electrical and Computer Engineering from the University of California, Riverside, in 2023, his M.S. from The Ohio State University, and his B.S. from the University of Electronic Science and Technology of China. He serves as an Associate Editor of the IEEE Open Journal of Intelligent Transportation Systems.
Education
University of California, Riverside
Ph.D.
Electrical and Computer Engineering
2023
The Ohio State University
Master of Science
Electrical and Computer Engineering
2017
Social
Areas of Expertise
Industry Expertise
Languages
- English
- Mandarin
- Japanese
Patents
Personalized vehicle operation for autonomous driving with inverse reinforcement learning
17572486
Systems and methods are provided for implementing personalized adaptive cruise control techniques in connection with, but not necessarily, autonomous and semi-autonomous vehicles. In accordance with one embodiment, a method comprises receiving first vehicle operating data and associated first environmental data of a plurality of vehicles; classifying the first vehicle operating data and the first environmental data into a plurality of driver type classifications; training a control policy model for each driver type classification based on the first vehicle operating data and the first environmental data; receiving a real-time classification of a target vehicle based on second vehicle operating data and associated second environmental data of the target vehicle; and output a trained control policy model the to target vehicle based on the real-time classification of the vehicle, wherein the target vehicle is controlled according to the trained control policy model.
Personalized adaptive cruise control based on steady-state operation
17578330
A personalized adaptive cruise control (P-ACC) system and associated algorithm are disclosed for determining a driver's preferred following gap in relation to vehicle speed based on periods of steady-state operation of a vehicle. While the P-ACC system is activated, vehicle transition states initiated by driver manual interventions such as takeover or overwrite events are used to identify subsequent periods of vehicle steady-state operation. Vehicle dynamics data captured during periods of steady-state operation is stored as steady-state data, which is then used to train a machine learning model to learn the driver's preferred following gap. This learned relationship is fed into second-order vehicle dynamics to determine a target acceleration for achieving the desired following gap while the P-ACC system is activated. Upon achieving the desired following gap, the vehicle speed may be held constant to maintain the following gap unless a change in lead vehicle speed necessitates updating the following gap.
Systems and methods for predicting driver visual impairment with artificial intelligence
12071141
According to various embodiments of the disclosed technology provide for systems and methods for predictive assessment of driver perception abilities based on driving behavior personalized to the driver.
In accordance with some embodiments, a method is provided that comprises receiving first vehicle operating data and associated first gaze data of a driver operating a vehicle; training a model for the driver based on the first vehicle operating data and the first gaze data, the model indicating driving behavior of the driver; receiving second vehicle operating data and associated second gaze data of the driver; and determining that an ability of the driver to perceive hazards is impaired based on applying the model to the second vehicle operating data and associated second gaze data.
In another aspect, a system is provided that comprises a memory and one or more processors that are configured to execute machine readable instructions stored in the memory for performing a method. The method comprises receiving historical vehicle operating data and associated historical gaze data of a driver operating a vehicle; learning a reward function based on application of inverse reinforcement learning (IRL) to the historical vehicle operating data and the historical environmental; calculating a cumulative reward from the reward function based on real-time vehicle operating data and associated real-time gaze data of the driver; and determining that an ability of the driver to perceive hazards is impaired based on the cumulative reward.
Other features and aspects of the disclosed technology will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, which illustrate, by way of example, the features in accordance with embodiments of the disclosed technology. The summary is not intended to limit the scope of any inventions described herein, which are defined solely by the claims attached hereto.
Method and system for personalized car following with transformers
12083883
In an embodiment, a method may include determining a vectorized representation of a position or road agents and road geometry based on sensor data from a vehicle, inputting the vectorized representation of the positions of the road agents and the road geometry into a trained transformer network, predicting one or more road agent trajectories at one or more future time steps based on an output of the transformer network, predicting an acceleration of the vehicle at the one or more future time steps based on the predicted one or more road agent trajectories at the one or more future time steps, and causing the vehicle to perform the predicted acceleration at the one or more future time steps.
In another embodiment, a remote computing device may include a controller. The controller may determine a vectorized representation of positions of road agents and road geometry based on sensor data from a vehicle. The controller may input the vectorized representation of the positions of the road agents and the road geometry into a trained transformer network. The controller may predict one or more road agent trajectories at one or more future time steps based on an output of the transformer network. The controller may predict an acceleration of the vehicle at the one or more future time steps based on the predicted one or more road agent trajectories at the one or more future time steps. The controller may cause the vehicle to perform the predicted acceleration at the one or more future time steps.
A system may include a vehicle including one or more vehicle sensors and a remote computing device including a controller. The vehicle sensors may collect sensor data including positions of road agents at a plurality of time steps and road geometry. The controller of the remote computing device may determine a vectorized representation of the positions of the road agents and the road geometry based on the sensor data from the vehicle. The controller may input the vectorized representation of the positions of the road agents and the road geometry into a trained transformer network. The controller may predict one or more road agent trajectories at one or more future time steps based on an output of the transformer network. The controller may predict an acceleration of the vehicle at the one or more future time steps based on the predicted one or more road agent trajectories at the one or more future time steps...
Courses
ENGR 100 Introduction to Engineering Analysis, Problem Solving and Design
This course is designed to introduce basic concepts relevant to engineering and to promote interest in the profession. The course seeks to establish a solid foundation of technical, creative, teamwork, and communication skills for engineers through effective problem solving, analysis, and design techniques.
Articles
Large-Language-Model-Agent-Enabled Analysis of Semi-Structured Interviews: Comparing Driver Experience Across Partial Automation Systems
Proceedings of the Human Factors and Ergonomics Society Annual MeetingZhouqiao Zhao, Pnina Gershon
2026-08-26
SAE Level 2 (L2) partial automation changes the driving task, shifting the driver’s role from active operator to supervisor, raising questions about trust, satisfaction, perceived limitations, and system understanding. This study presents a locally hosted, auditable large language model (LLM) agent pipeline for analyzing 161 semi-structured post-study interviews from a 30-day naturalistic driving study of Tesla Autopilot, Cadillac Super Cruise, and Volvo Pilot Assist. The pipeline segmented transcripts into semantic units and coded each unit by question theme, answer theme, subtheme, and sentiment, enabling structured comparison across systems while preserving traceability to source evidence. Results showed significant vehicle-level differences in sentiment, with satisfaction-related responses strongly positive but performance-limitation narratives strongly negative. Validation against manually coded interview data showed high accuracy for answer subthemes and sentiment. Findings suggest that driver satisfaction with partial automation can coexist with uncertainty about system limits and automation boundaries.
Precision Camera Calibration for AI-Powered Traffic Surveillance at Urban Intersections
2026 7th International Conference on Artificial Intelligence, Robotics, and Control (AIRC)Chuheng Wei, Zhouqiao Zhao, Minghao Han, Kevin Boriboonsomsin, Guoyuan Wu
2026-04-08
In the era of AI-driven Intelligent Transportation Systems (ITS), accurate camera calibration serves as the fundamental prerequisite for enabling reliable computer vision algorithms at signalized intersections. Precise geometric calibration is essential for downstream AI tasks such as object detection, multi-object tracking, trajectory prediction, and traffic behavior analysis. However, roadside cameras at intersections face unique challenges including diverse vehicle trajectories, varying pedestrian paths, dynamic lighting conditions, and multiple viewing angles from different camera placements. This paper proposes a refined calibration technique specifically designed for AI-ready intersection monitoring systems. Leveraging Zhang’s Calibration Method for intrinsic parameter estimation and the Perspectiven-Point (PnP) approach for extrinsic calibration, we optimize the entire process using the Levenberg-Marquardt technique to address the unique challenges of intersection environments. We validate our method at two distinct real-world intersections, achieving localization errors of 0.3496 m and 0.1850 m, respectively. These results demonstrate the method’s precision and effectiveness in providing geometric accuracy that meets the stringent requirements of AI-powered traffic surveillance systems, paving the way for enhanced intelligent transportation applications.
From Prediction to Design: Using Context-Aware Graph Neural Networks and Explainable AI to Anticipate Transfer-of-Control
WCX SAE World CongressZhouqiao Zhao, Pnina Gershon
2026-04-07
Drivers often interact with partial automation (SAE Level 2) systems, initiating transfer of control (TOC) either by handing control over to the automation or by taking it back. Accurately predicting these interactions may inform the design of future automation systems that adapt proactively to the operating context, enhance comfort, and ultimately may improve safety. We present a context-aware framework that generates a unified driver–vehicle–environment representation by fusing data from in-cabin video of the driver and of the forward roadway with vehicle kinematics, driver glance, and hands-on-wheel behaviors. This representation was encoded in a hierarchical Graph Neural Network that classified driver-initiated TOCs to: (i) Manual-to-automation and (ii) Automation-to-manual transitions and predicted time-to-TOC. Shapley-based explainable AI was used to quantify how the importance of behavioral, contextual, and kinematic cues evolved in the seconds preceding a TOC. Analysis of a naturalistic dataset of 1,565 driver-initiated TOCs from 16 experienced drivers revealed distinct patterns. Manual-to-automation transitions were preceded by lane count increases, acceleration, and spikes in glances to the instrument-cluster. In contrast, Automation-to-manual transitions were associated with lane count reductions, higher surrounding-vehicle density, deceleration, reduction in secondary-task engagement, and higher steering wheel control. Together, these patterns highlight key cues for predicting the TOC type and time-to-TOC. Using environment-only features, the classifier achieved 78% accuracy; adding vehicle kinematics increased accuracy to 84%, and incorporating driver behavior features further improved prediction to 90%. Across prediction horizons, the Manual-to-automation TOC was consistently predicted more accurately than the automation-to-manual TOC. Shapley analyses underscore that driver behavior provided the strongest cues for predicting TOCs, highlighting the value of fusing driving context with information obtained from monitoring the driver behavior to anticipate the type of driver-automation interaction and its timing.
Design, Implementation, and Evaluation of an Innovative Vehicle-Powertrain Eco-Operation System for Plug-In Hybrid Electric Buses
2025 IEEE Conference on Technologies for Sustainability (SusTech)Zhouqiao Zhao, Guoyuan Wu, Peng Hao, Fei Ye, Zhiming Gao, Tim J LaClair, Dylan Brown, Danial Esaid, Kanok Boriboonsomsin, Matthew J Barth
2025-04-20
Transit buses serve a vital role in sustainable transportation systems, providing mobility to millions of passengers daily. These buses primarily operate on fixed routes in urban areas, leading to frequent stops at bus stops and traffic signals, which contribute to them having low fuel economy. In recent years, hybrid electric buses and plug-in hybrid electric buses (PHEBs) have gained significant interest in transit applications. However, the energy efficiency of early HEBs and PHEBs is limited as they rely primarily on simple charge sustaining strategies. This paper presents the design, implementation, and validation of a Connected Eco-Bus system that utilizes connected and automated vehicle technology to improve the energy efficiency of a power-split PHEB. The system co-optimizes the PHEB’s vehicle dynamics and powertrain controls by leveraging connectivity and partial automation (Level 1) capability. A case study is conducted with the Connected Eco-Bus system operating on a typical urban route where its performance is evaluated through both microscopic simulation and Dynamometer-in-the-Loop testing. The results demonstrate that the Connected Eco-Bus system can achieve energy efficiency improvements of up to 32.4%, which would thereby contribute to a more sustainable transportation system.
Driver behavior in response to forward collision warnings considering driving context
Proceedings of the Human Factors and Ergonomics Society Annual MeetingZhouqiao Zhao, Linda Pipkorn, Bruce Mehler, Bryan Reimer, Pnina Gershon
Forward collision warnings (FCW) are designed to warn drivers of potential collisions, but their effectiveness may vary based on driving conditions and driver responses. This study investigated driver behavior in response to real-world FCW alerts across various driving contexts with the aims of evaluating the prevalence and characteristics of FCW events, i.e., examining how often these events occur, their severity, and the contexts in which they happen. The current analysis considered driver responses to FCW events, including changes in vehicle kinematics, visual attention, and hand-on-wheel behavior. Findings suggest that lower-severity FCWs are more common on local roads and associated with adjacent vehicles, while higher-severity FCWs were observed to be more frequent on highways and in the presence of lead vehicles. The study highlights the potential value of integrating more advanced contextual …
A Review of Personalization in Driving Behavior: Dataset, Modeling, and Validation
IEEE Transactions on Intelligent VehiclesXishun Liao, Zhouqiao Zhao, Matthew J Barth, Amr Abdelraouf, Rohit Gupta, Kyungtae Han, Jiaqi Ma, Guoyuan Wu
2024-07-08
Personalization in driving behavior research is crucial for developing intelligent vehicles that can safely coexist with human-driven vehicles in mixed-traffic environments. By accounting for the diversity of human driving behaviors, personalized modeling can improve predictive capabilities of intelligent vehicles and foster a more balanced traffic ecosystem. This paper presents a systematic review on personalization in driving behavior, evaluating their potential to enhance road safety, transportation efficiency, and human-centric mobility. It proposes a taxonomy to categorize personalized driving behaviors and surveys relevant datasets, modeling methodologies, and techniques for validating personalized driver models. Focusing on personalized driving behavior, the study emphasizes the need for intelligent vehicles to adapt to the complex and heterogeneous behaviors exhibited by human drivers to enhance predictability, responsiveness, and ultimately create a safe and efficient traffic environment. Lastly, key challenges are identified, along with promising future research directions to advance personalized driving behavior research.
Inverse reinforcement learning and Gaussian process regression-based real-time framework for personalized adaptive cruise control
2023 IEEE 26TH International conference on intelligent transportation systems (ITSC)Zhouqiao Zhao, Xishun Liao, Amr Abdelraouf, Kyungtae Han, Rohit Gupta, Matthew J Barth, Guoyuan Wu
2023-09-24
Adaptive Cruise Control (ACC) has become increasingly popular in modern vehicles, providing enhanced driving safety, comfort, and fuel efficiency. However, predefined ACC settings may not always align with a driver's preferences, leading to discomfort and possible safety hazards. To address this issue, Personalized ACC (P-ACC) has been studied by scholars. However, existing research mostly relies on historical driving data to imitate driver styles, which ignores real-time feedback from the driver. To overcome this limitation, we propose a cloud-vehicle collaborative P-ACC framework, which integrates real-time driver feedback adaptation. This framework consists of offline and online modules. The offline module records the driver's naturalistic car-following trajectory and uses inverse reinforcement learning (IRL) to train the model on the cloud. The online module utilizes the driver's real-time feedback to update the driving gap preference in real-time using Gaussian process regression (GPR). By retraining the model on the cloud with the driver's takeover trajectories, our approach achieves incremental learning to better match the driver's preference. In human-in-the-loop (HuiL) simulation experiments, the proposed framework results in a significant reduction of driver intervention in automatic control systems, up to 70.9%.
Real-time learning of driving gap preference for personalized adaptive cruise control
arXiv preprint arXiv:2309.05115Zhouqiao Zhao, Xishun Liao, Amr Abdelraouf, Kyungtae Han, Rohit Gupta, Matthew J Barth, Guoyuan Wu
2023-09-10
Advanced Driver Assistance Systems (ADAS) are increasingly important in improving driving safety and comfort, with Adaptive Cruise Control (ACC) being one of the most widely used. However, pre-defined ACC settings may not always align with driver's preferences and habits, leading to discomfort and potential safety issues. Personalized ACC (P-ACC) has been proposed to address this problem, but most existing research uses historical driving data to imitate behaviors that conform to driver preferences, neglecting real-time driver feedback. To bridge this gap, we propose a cloud-vehicle collaborative P-ACC framework that incorporates driver feedback adaptation in real time. The framework is divided into offline and online parts. The offline component records the driver's naturalistic car-following trajectory and uses inverse reinforcement learning (IRL) to train the model on the cloud. In the online component, driver feedback is used to update the driving gap preference in real time. The model is then retrained on the cloud with driver's takeover trajectories, achieving incremental learning to better match driver's preference. Human-in-the-loop (HuiL) simulation experiments demonstrate that our proposed method significantly reduces driver intervention in automatic control systems by up to 62.8%. By incorporating real-time driver feedback, our approach enhances the comfort and safety of P-ACC, providing a personalized and adaptable driving experience.
End-to-end spatio-temporal attention-based lane-change intention prediction from multi-perspective cameras
2023 IEEE Intelligent Vehicles Symposium (IV)Zhouqiao Zhao, Zhensong Wei, Danyang Tian, Bryan Reimer, Pnina Gershon, Ehsan Moradi-Pari
2023-06-04
Advanced Driver Assistance Systems (ADAS) with proactive alerts have been used to increase driving safety. Such systems’ performance greatly depends on how accurately and quickly the risky situations and maneuvers are detected. Existing ADAS provide warnings based on the vehicle’s operational status, detection of environments, and the drivers’ overt actions (e.g., using turn signals or steering wheels), which may not give drivers as much as optimal time to react. In this paper, we proposed a spatio-temporal attention-based neural network to predict drivers’ lane-change intention by fusing the videos from both in-cabin and forward perspectives. The Convolutional Neural Network (CNN)-Recursive Neural Network (RNN) network architecture was leveraged to extract both the spatial and temporal information. On top of this network backbone structure, the feature maps from different time steps and perspectives were fused using multi-head self-attention at each resolution of the CNN. The proposed model was trained and evaluated using a processed subset of the MIT Advanced Vehicle Technology (MIT-AVT) dataset which contains synchronized CAN data, 11058-second videos from 3 different views, 548 lane-change events, and 274 non-lane-change events performed by 83 drivers. The results demonstrate that the model achieves 87% F1-score within the 1-second validation window and 70% F1-score within the 5-second validation window with real-time performance.
Driver digital twin for online prediction of personalized lane-change behavior
IEEE Internet of Things JournalXishun Liao, Xuanpeng Zhao, Ziran Wang, Zhouqiao Zhao, Kyungtae Han, Rohit Gupta, Matthew J Barth, Guoyuan Wu
2023-03-27
Connected and automated vehicles (CAVs) are supposed to share the road with human-driven vehicles (HDVs) in a foreseeable future. Therefore, considering the mixed traffic environment is more pragmatic, as the well-planned operation of CAVs may be interrupted by HDVs. In the circumstance that human behaviors have significant impacts, CAVs need to understand HDV behaviors to make safe actions. In this study, we develop a driver digital twin (DDT) for the online prediction of personalized lane-change behavior, allowing CAVs to predict surrounding vehicles’ behaviors with the help of the digital twin technology. DDT is deployed on a vehicle-edge–cloud architecture, where the cloud server models the driver behavior for each HDV based on the historical naturalistic driving data, while the edge server processes the real-time data from each driver with his/her digital twin on the cloud to predict the personalized lane-change maneuver. The proposed system is first evaluated on a human-in-the-loop co-simulation platform, and then in a field implementation with three passenger vehicles driving along an on/off ramp segment connecting to the edge server and cloud through the 4G/LTE cellular network. The lane-change intention can be recognized in 6 s on average before the vehicle crosses the lane separation line, and the Mean Euclidean Distance between the predicted trajectory and GPS ground truth is 1.03 m within a 4-s prediction window. Compared to the general model, using a personalized model can improve prediction accuracy by 27.8%. The demonstration video of the proposed system can be watched at https://youtu.be/5cbsabgIOdM.
Robotic competitions to design future transport systems: The case of JRC autotrac 2020
Transportation Research RecordBiagio Ciuffo, Michail Makridis, Valter Padovan, Emilio Benenati, Kanok Boriboonsomsin, Mamen Thomas Chembakasseril, Petros Daras, Viswanath Das, Anastasios Dimou, Sergio Grammatico, Ronny Hartanto, Malte Hoelscher, Yu Jiang, Suad Krilasevic, Shangrui Liu, Quang Nhat Nguyen Le, Cas Rosier, Pingbo Ruan, Zhensong Wei, Guoyuan Wu, Xuanpeng Zhao, Zhouqiao Zhao
Vehicle automation and connectivity bring new opportunities for safe and sustainable mobility in urban and highway networks. Such opportunities are however not directly associated with traffic flow improvements. Research on exploitation of connected and automated vehicles (CAVs) toward a more efficient traffic currently remains at a theoretical level, and/or based on simulation models with limited reliability. Furthermore, testing CAVs in the real world is still costly and very challenging from an implementation perspective. A possible alternative is to use automated robots. By designing and testing both the low- and the high-level controllers of CAVs, it is indeed possible to reach a better understanding of the challenges that future vehicles will need to face. Robotic applications can effectively test these challenges within a wide variety of research communities—for example, via robotic competitions. Along this …
Bi-level fleet dispatching strategy for battery-electric trucks: A real-world case study
SustainabilityDongbo Peng, Zhouqiao Zhao, Guoyuan Wu, Kanok Boriboonsomsin
2023-01-04
Driven by new regulations concerning greenhouse gas (GHG) emissions in the transportation sector, battery-electric trucks (BETs) are considered one of the sustainable freight transportation solutions. In this paper, a dispatching problem of the BET fleet is formulated as a capacitated electric vehicle routing problem (VRP) with pick-up and delivery. As the BET dispatching problem is NP-hard, the performance of existing approaches deteriorates in large instance problems, especially when the customers have different preferences and constraints. This article proposes a bi-level strategy that incorporates routing zone partitioning and metaheuristic-based vehicle routing to solve the large-scale BET dispatching problem, considering the delivery types, limited travel distances, and cargo payloads. We apply this strategy to a real-world fleet dispatching scenario with around 300 customer positions for pickups and drop-offs. The experimental results demonstrate that the proposed bi-level strategy can reduce total travel distance and travel time by 24–31%, compared to the baseline strategy implemented in the real world.
A Connected Automation Enabled Cooperative Management Framework for Mixed Traffic
University of California, RiversideZhouqiao Zhao
Safety, mobility, and environmental sustainability form the triad of challenges in modern transportation systems. To tackle these issues, there has been an increasing emphasis on intelligent transportation systems (ITS) technology, which employs interdisciplinary approaches to provide effective solutions. Transportation systems are characterized by their large scale, non-linearity, time-varying behavior, interconnectivity, heterogeneity, and distributed nature, with various participants engaging in intensive interactions.
Real-time Adaptive Background Subtraction for Traffic Scenarios at Signalized Intersections Based on Roadside Fish-eye Cameras
2022 IEEE 25th International Conference on Intelligent Transportation Systems (ITSC)Jiahe Cao, Zhouqiao Zhao, Guoyuan Wu, Matthew J Barth, Yongkang Liu, Emrah Akin Sisbot, Kentaro Oguchi
2022-10-08
Background subtraction (BS) has been a norm for moving object detection along a classical computer vision pipeline, especially when the labelled data is largely unavailable. It has been widely used for infrastructure-based sensing such as traffic surveillance with roadside cameras. Existing BS algorithms focus on detecting moving objects, while the temporal motionless objects are neglected. This leads to performance degradation in particular at signalized intersections where vehicles may stop to wait in red. In this paper, we propose a hierarchical adaptive BS method which can eliminate the cumulative errors for those temporally static objects based on images from roadside fish-eye cameras in real-time. The proposed method is validated in both the CARLA simulation and the real-world environment. The results show that our method outperforms ViBe and LOBSTER by about 45% and 39%, respectively, on recall, without compromising too much in precision.
Personalized car following for autonomous driving with inverse reinforcement learning
2022 International Conference on Robotics and Automation (ICRA)Zhouqiao Zhao, Ziran Wang, Kyungtae Han, Rohit Gupta, Prashant Tiwari, Guoyuan Wu, Matthew J Barth
2022-05-23
Driving automation is gradually replacing human driving maneuvers in different applications such as adaptive cruise control and lane keeping. However, contemporary driving automation applications based on expert systems or prede-fined control strategies are not in line with individual human driver's preference. To overcome this problem, we propose a Personalized Adaptive Cruise Control (P-ACC) system that can learn the driver's car-following preferences from historical data using model-based maximum entropy Inverse Reinforcement Learning (IRL). Once activated in real-time, the P-ACC system first classifies the driver type and the weather type (at that moment). The vehicle is then controlled using the pre-trained IRL model on the cloud of the associated class. The personalized IRL model on the cloud will be updated as more human driving data is collected from various scenarios. Numerical simulation with real-world naturalistic driving data shows that, the accuracy of reproducing the real-world driving profile improves up to 30.1% in terms of speed and 36.5% in terms of distance gap, when P-ACC is compared with the Intelligent Driver Model (IDM). Game engine-based human-in-the-loop simulation demonstrates that, the takeover frequency of the driver during the usage of P-ACC decreases up to 93.4%, compared with that during the usage of IDM-based ACC.
Online prediction of lane change with a hierarchical learning-based approach
2022 International Conference on Robotics and Automation (ICRA)Xishun Liao, Ziran Wang, Xuanpeng Zhao, Zhouqiao Zhao, Kyungtae Han, Prashant Tiwari, Matthew J Barth, Guoyuan Wu
2022-05-23
In the foreseeable future, connected and auto-mated vehicles (CAVs) and human-driven vehicles will share the road networks together. In such a mixed traffic environment, CAVs need to understand and predict maneuvers of surrounding vehicles for safer and more efficient interactions, especially when human drivers bring in a wide range of uncertainties. In this paper, we propose a learning-based lane-change prediction algorithm that considers the driving behaviors of the target human driver. To provide accurate maneuver prediction, we adopt a hierarchical structure that seamlessly seals both the lane-change decision prediction and the vehicle trajectory pre-diction together. Specifically, we propose a lane-change decision prediction method based on a Long-Short Term Memory (LSTM) network, and a trajectories prediction considering driver preference and vehicular interactions based on Inverse Reinforcement Learning (IRL). To validate the performance of the proposed methodology, a case study of an on-ramp merging scenario is conducted on a uniquely built human-in-the-loop simulation platform that can provide an immersive driving environment, collect data of lane-change behaviors, and test drivers' reactions to the prediction results in real time. It is shown in the simulation results that we can predict the lane-change decision 3 seconds before the vehicle crosses the line to another lane, and the Mean Euclidean Distance between the predicted trajectory and ground truth is 0.39 meters within a 4-second prediction window.
Connected vehicle-based advanced detection of “slow-down” events on freeways
2021 IEEE International Intelligent Transportation Systems Conference (ITSC)Zhouqiao Zhao, Guoyuan Wu, Matthew J Barth, Hossein Nourkhiz Mahjoub, Yasir Khudhair Al-Nadawi, Laith Daman, Shigenobu Saigusa
2021-09-19
From the perspective of an individual vehicle, the prediction of a “slow-down” or shockwave event on a freeway can help the driver reduce potential collision risks, enhance the driving experience, and reduce the cost of energy consumption and vehicle maintenance. From the perspective of traffic management, shockwave prediction may help regulate traffic flow effectively and allow for the response to (non-recurrent) incidents in a timely manner. In this paper, two real-time prediction algorithms are proposed and investigated, which are based on the high-resolution information provided from a set of connected vehicles within the communication range of the host vehicle. Both methods are able to predict the “slow-down” event under high traffic density at 3.51 seconds (on average) earlier than its occurrence. Both algorithm performances degrade with the decrease of the traffic density and penetration rate of the connected vehicles.
Corridor-wise eco-friendly cooperative ramp management system for connected and automated vehicles
SustainabilityZhouqiao Zhao, Guoyuan Wu, Matthew Barth
2021-07-31
Safety, mobility, and environmental sustainability are three fundamental issues that our transportation system has been confronting for decades. Intelligent transportation systems (ITS) aim to address these problems by leveraging disruptive technologies, such as connected and automated vehicles (CAVs). The cooperative potential of CAVs enable more efficient maneuvers and operation of a group of vehicles, or even the entire traffic system. In addition, CAVs may couple with other emerging technologies such as electrification to boost overall system performance and to further mitigate the aforementioned issues. In this study, we propose a hierarchical eco-friendly cooperative ramp management system, where macroscopically, a stratified ramp metering algorithm, is deployed to coordinate all of the ramp inflow rates along a corridor according to the real-time traffic condition; microscopically, a model predictive control (MPC)-based algorithm is designed for the detailed speed control of individual CAVs. Using the shared information from CAVs, the proposed ramp management system can smooth traffic flow, improve system mobility, and decrease the energy consumption of the network. Moreover, traffic simulation has been conducted using PTV VISSIM under various congestion levels for vehicles with different powertrain types, i.e., an internal combustion engine and an electric motor. Compared to conventional ramp metering, the proposed ramp management system may improve mobility by 48.6–56.7% and save energy by 24.0–35.1%. Compared to no control scenarios, savings in travel time and energy consumption are in the ranges of 79.4–89.1% and 0.8–2.5%, respectively.
Vehicle dispatching and scheduling algorithms for battery electric heavy-duty truck fleets considering en-route opportunity charging
2021 IEEE Conference on Technologies for Sustainability (SusTech)Zhouqiao Zhao, Guoyuan Wu, Kanok Boriboonsomsin, Aravind Kailas
2021-04-22
There has been growing interest in the electrification of medium- and heavy-duty vehicles (M-HDVs) in real-world, regional distribution applications. Fleet dispatch optimization of battery-electric trucks (BETs) is critical given the limited onboard energy, charging characteristics, and operational considerations. Our paper proposes a bi-level hierarchical method to optimize BET dispatch during pickup and delivery runs. With any route/scheduling change, the average speed, travel time, and energy consumption from one location to another will change accordingly because of the weight of the goods and the real-time traffic condition. So, the "electric vehicle routing problem" was extended to include pickup and delivery, time windows, and partial recharge. The proposed algorithm significantly reduces the operation cost of the BET fleet considering labor, energy consumption, and time window penalties without compromising computational efficiency.
Shared automated mobility with demand-side cooperation: A proof-of-concept microsimulation study
SustainabilityLei Zhu, Zhouqiao Zhao, Guoyuan Wu
2021-02-25
Most existing shared automated mobility (SAM) services assume the door-to-door manner, i.e., the pickup and drop-off (PUDO) locations are the places requested by the customers (or demand-side). While some mobility services offer more affordable riding costs in exchange for a little walking effort from customers, their rationales and induced impacts (in terms of mobility and sustainability) from the system perspective are not clear. This study proposes a demand-side cooperative shared automated mobility (DC-SAM) service framework, aiming to fill this knowledge gap and to assess the mobility and sustainability impacts. The optimal ride matching problem is formulated and solved in an online manner through a micro-simulation model, Simulation of Urban Mobility (SUMO). The objective is to maximize the profit (considering both the revenue and cost) of the proposed SAM service, considering the constraints in seat capacities of shared automated vehicles (SAVs) and comfortable walking distance from the perspective of customers. A case study on a portion of a New York City (NYC) network with a pre-defined fleet size demonstrated the efficacy and promise of the proposed system. The results show that the proposed DC-SAM service can not only significantly reduce the SAV’s operating costs in terms of vehicle-miles traveled (VMT), vehicle-hours traveled (VHT), and vehicle energy consumption (VEC) by up to 53, 46 and 51%, respectively, but can also considerably improve the customer service by 30 and 56%, with regard to customer waiting time (CWT) and trip detour factor (TDF), compared to a heuristic service model. In addition, the demand-side cooperation strategy can bring about additional system-wide mobility and sustainability benefits in the range of 4–10%.
Review on connected and automated vehicles based cooperative eco-driving strategies
交通运输工程学报YANG Lan, ZHAO Xiang-mo, WU Guo-yuan, XU Zhi-gang, MATTHEW Barth, HUI Fei, HAO Peng, HAN Meng-jie, ZHAO Zhou-qiao, FANG Shan, JING Shou-cai
2020-10-25
To track the research progress of connected and automated vehicles(CAV) based cooperative eco-driving strategies in recent years, the influences of four factors, including the vehicle, driver behavior, traffic network and social factor on the energy consumption of CAV were analyzed. The current ecological studies on CAV were classified with vehicle, infrastructure and traveler as objects. The status-quo of 5 representative types of cooperative eco-driving scenarios were emphatically analyzed, including the eco-approach and departure at the signalized intersection, eco-cooperative adaptive cruise control, eco-cooperative driving in the ramp merging area, eco-cooperative lane changing trajectory planning and eco-routing. Analysis result shows that compared with the human driving mode, CAVs can save up to 63% fuel consumption at any traffic flow with 100% penetration rate of CAV as well as in the light traffic condition with partial penetration rate of CAV. CAVs with partial automated and connected levels can save at least 7% fuel consumption. Few existing studies consider the trajectory tracking deviation caused by the driver's response delay and automatic controller transmission delay in the case of human-machine co-driving. The existing researches assume the vehicle-to-vehicle communication(V2V) and vehicle-to-infrastructure communication(V2I) as the ideal data interaction processes. The impacts of factors such as the communication topology, transmission delay, communication failure and packet loss on the CAV based cooperative eco-driving strategies are ignored. Few existing studies discuss the eco-driving strategies in these traffic scenarios, such as the multi-lanes, shared lanes for turning and through at intersection, U-turn, as well as the mixed traffic conditions of different automated and connected level CAVs coexisting with human-driven vehicles, pedestrians and bicycles. Limited by the immaturity and imperfection of automatic driving technology and infrastructure, the test and verification work in real traffic scenarios is not carried out. The vehicle control, V2V communication, multi-vehicles collaboration, mixed traffic flow scenario, hardware-in-the-loop simulation test and real traffic scenario test will be the further development direction of CAV based cooperative eco-driving strategies.
Optimal control-based eco-ramp merging system for connected and automated vehicles
2020 IEEE Intelligent Vehicles Symposium (IV)Zhouqiao Zhao, Guoyuan Wu, Ziran Wang, Matthew J Barth
2020-10-19
Our current transportation system suffers from a number of problems in terms of safety, mobility, and environmental sustainability. The emergence of innovative intelligent transportation systems (ITS) technologies, and in particular connected and automated vehicles (CAVs), provides many opportunities to address the aforementioned issues. In this paper, we propose a hierarchical ramp merging system that not only generates microscopic cooperative maneuvers for CAVs on the ramp to merge into the mainline traffic flow, but also provides controllability of the ramp inflow rate, thereby enabling macroscopic traffic flow control. A centralized optimal control-based approach is proposed to smooth the merging flow, improve the system-wide mobility, and decrease the overall fuel consumption of the network. Linear quadratic trackers in both finite horizon and receding horizon forms are developed to solve the optimization problem in terms of path planning and sequence determination, where a microscopic vehicle fuel consumption model is applied. Extensive traffic simulation runs have been conducted using PTV VISSIM to evaluate the impact of the proposed system on a segment of SR-91 E in Corona, California. The results confirm that under the regulated inflow rate, the proposed system can avoid potential traffic congestion and improve mobility (e.g., VMT/VHT) up to 147%, with a 47% fuel savings compared to the conventional ramp metering and the ramp without any control approach.
Developing a data-driven modularized model of a plug-in hybrid electric bus (PHEB) for connected and automated vehicle applications
2020 IEEE 23rd International Conference on Intelligent Transportation Systems (ITSC)Zhouqiao Zhao, Zhensong Wei, Guoyuan Wu, Matthew J Barth
2020-09-20
Shared Electric Connected and Automated Vehicles have the potential to improve transportation safety, mobility, and energy efficiency. A plug-in hybrid electric architecture is well suited for developing connected and automated vehicle (CAV) applications, allowing for vehicle dynamics management and powertrain control. In this paper, we developed a data-driven modularized modeling approach for a plug-in hybrid electric bus (PHEB), thereby allowing for a wide range of connected and automated vehicle applications. Instead of using an end-to-end learning approach to model the PHEB, our modularized modeling approach considers the physical connection of each component of PHEB, which provides various signals and dynamics of each subsystem for testing use or controller design. The plug-and-play (PnP) feature allows us to customize the bus model and update each individual module in a flexible manner. The modules include human driver behavior, energy management system, internal combustion engine, electric motor(s), transmission, and powertrain dynamics. For each module, a Long Short-term Memory (LSTM) network is utilized to learn each modules' behavior and dynamics using the data from extensive dynamometer-in-the-loop (DiL) testing.
Dyno-in-the-loop: an innovative hardware-in-the-loop development and testing platform for emerging mobility technologies
WCX SAE World Congress Experience 237384Guoyuan Wu, Dylan Brown, Zhouqiao Zhao, Peng Hao, Michael Todd, Kanok Boriboonsomsin, Matthew Barth, Zhiming Gao, Tim LaClair
2020-04-14
Today’s transportation is quickly transforming with the nascent advent of connectivity, automation, shared-mobility, and electrification. These technologies will not only affect our safety and mobility, but also our energy consumption, and environment. As a result, it is of unprecedented importance to understand the overall system impacts due to the introduction of these emerging technologies and concepts. Existing modeling tools are not able to effectively capture the implications of these technologies, not to mention accurately and reliably evaluating their effectiveness with a reasonable scope. To address these gaps, a dynamometer-in-the-loop (DiL) development and testing approach is proposed which integrates test vehicle(s), chassis dynamometer, and high fidelity traffic simulation tools, in order to achieve a balance between the model accuracy and scalability of environmental analysis for the next generation of transportation systems. With this DiL platform, a connected eco-operation system for the plug-in hybrid electric bus (PHEB) has been developed and tested, which can optimize the vehicle dynamics (and potentially powertrain control via smart energy management) to reduce the operational energy consumption as well as tailpipe emissions of the target PHEB. The system performance has been evaluated on the DiL platform with respect to a variety of traffic congestion levels. The results have shown that the developed system can save fuel by more than 13% while reducing the electricity consumption by 2% in the test scenarios.
Development of Eco-Friendly Ramp Control for Connected and Automated Electric Vehicles
National Center for Sustainable Transportation Research ReportGuoyuan Wu, Zhouqiao Zhao, Ziran Wang, Matthew J Barth
With on-board sensors such as camera, radar, and Lidar, connected and automated vehicles (CAVs) can sense the surrounding environment and be driven autonomously and safely by themselves without colliding into other objects on the road. CAVs are also able to communicate with each other and roadside infrastructure via vehicle-to-vehicle and vehicle-to-infrastructure communications, respectively, sharing information on the vehicles’ states, signal phase and timing (SPaT) information, enabling CAVs to make decisions in a collaborative manner. As a typical scenario, ramp control attracts wide attention due to the concerns of safety and mobility in the merging area. In particular, if the line-of-the-sight is blocked (because of grade separation), then neither mainline vehicles nor on-ramp vehicles may well adapt their own dynamics to perform smoothed merging maneuvers. This may lead to speed fluctuations or even shockwave propagating upstream traffic along the corridor, thus potentially increasing the traffic delays and excessive energy consumption. In this project, the research team proposed a hierarchical ramp merging system that not only allowed microscopic cooperative maneuvers for connected and automated electric vehicles on the ramp to merge into mainline traffic flow, but also had controllability of ramp inflow rate, which enabled macroscopic traffic flow control. A centralized optimal control-based approach was proposed to both smooth the merging flow and improve the system-wide mobility of the network. Linear quadratic trackers in both finite horizon and receding horizon forms were developed to solve the optimization problem in terms of path planning and sequence determination, and a microscopic electric vehicle (EV) energy consumption model was applied to estimate the energy consumption. The simulation results confirmed that under the regulated inflow rate, the proposed system was able to avoid potential traffic congestion and improve the mobility (in terms of average speed) as much as 115%, compared to the conventional ramp metering and the ramp without any control approach. Interestingly, for EVs (connected and automated EVs in this study), the improved mobility may not necessarily result in the reduction of energy consumption. The “sweet spot” of average speed ranges from 27–34 mph for the EV models in this study.
The state-of-the-art of coordinated ramp control with mixed traffic conditions
2019 IEEE Intelligent Transportation Systems Conference (ITSC)Zhouqiao Zhao, Ziran Wang, Guoyuan Wu, Fei Ye, Matthew J Barth
2019-10-27
Ramp metering, a traditional traffic control strategy for conventional vehicles, has been widely deployed around the world since the 1960s. On the other hand, the last decade has witnessed significant advances in connected and automated vehicle (CAV) technology and its great potential for improving safety, mobility and environmental sustainability. Therefore, a large amount of research has been conducted on cooperative ramp merging for CAVs only. However, it is expected that the phase of mixed traffic, namely the coexistence of both human-driven vehicles and CAVs, would last for a long time. Since there is little research on the system-wide ramp control with mixed traffic conditions, the paper aims to close this gap by proposing an innovative system architecture and reviewing the state-of-the-art studies on the key components of the proposed system. These components include traffic state estimation, ramp metering, driving behavior modeling, and coordination of CAVs. All reviewed literature plot an extensive landscape for the proposed system-wide coordinated ramp control with mixed traffic conditions.
Optimal control-based eco-ramp merging system for connected and automated electric vehicles
arXiv preprint arXiv:1910.07620Zhouqiao Zhao, Guoyuan Wu, Ziran Wang, Matthew J Barth
2019-10-16
Our current transportation system suffers from a number of problems in terms of safety, mobility, and environmental sustainability. The emergence of innovative intelligent transportation systems (ITS) technologies, and in particular connected and automated vehicles (CAVs), provides many opportunities to address the aforementioned issues. In this paper, we propose a hierarchical ramp merging system that not only generates microscopic cooperative maneuvers for CAVs on the ramp to merge into the mainline traffic flow, but also provides controllability of the ramp inflow rate, thereby enabling macroscopic traffic flow control. A centralized optimal control-based approach is proposed to smooth the merging flow, improve the system-wide mobility, and decrease the overall fuel consumption of the network. Linear quadratic trackers in both finite horizon and receding horizon forms are developed to solve the optimization problem in terms of path planning and sequence determination, where a microscopic vehicle fuel consumption model is applied. Extensive traffic simulation runs have been conducted using PTV VISSIM to evaluate the impact of the proposed system on a segment of SR-91 E in Corona, California. The results confirm that under the regulated inflow rate, the proposed system can avoid potential traffic congestion and improve mobility (e.g., VMT/VHT) up to 147%, with a 47% fuel savings compared to the conventional ramp metering and the ramp without any control approach.
