Zhouqiao (Bridge) Zhao

2026 SAE World Congress

This work presents a context-aware framework that fuses driver behavior, vehicle kinematics, and the surrounding environment to classify driver-initiated transfers of control and predict their timing. A hierarchical graph neural network supports prediction, while explainable AI reveals how behavioral and contextual cues evolve before transitions between manual driving and partial automation.

• Read Publication →

Proceedings of the Human Factors and Ergonomics Society Annual Meeting

This study uses a large-language-model-agent-enabled workflow to analyze semi-structured interviews and compare driver experiences across partial automation systems. The work connects scalable qualitative analysis with human-factors questions about how drivers understand, use, and evaluate vehicle automation.

• Read Publication →

2025 IEEE Conference on Technologies for Sustainability (SusTech)

This paper presents a Connected Eco-Bus system that co-optimizes the vehicle dynamics and powertrain controls of a plug-in hybrid electric bus (PHEB) to improve its energy efficiency. The system is evaluated through both simulation and Dynamometer-in-the-Loop testing, and the results show that it can achieve energy efficiency improvements of up to 32.4%.

• Read Publication →

IEEE Transactions on Intelligent Vehicles

This paper provides a systematic review of personalization in driving behavior. It proposes a taxonomy to categorize personalized driving behaviors and surveys relevant datasets, modeling methodologies, and validation techniques. The paper emphasizes the need for intelligent vehicles to adapt to the complex and heterogeneous behaviors of human drivers to create a safe and efficient traffic environment.

• Read Publication →

Human Factors and Ergonomics Society Annual Meeting (HFES)

This study investigates driver behavior in response to forward collision warnings (FCWs) in various driving contexts. The findings suggest that the severity of FCW alerts is associated with the driving environment, and that integrating more contextual information into FCW systems could enhance their performance and reduce unnecessary alerts.

• Read Publication →

2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC)

This paper proposes a real-time framework for Personalized Adaptive Cruise Control (P-ACC) that uses Inverse Reinforcement Learning (IRL) and Gaussian Process Regression (GPR) to learn a driver's car-following preferences and adapt to their real-time feedback. The proposed framework is shown to reduce driver intervention in automatic control systems by up to 70.9% in human-in-the-loop simulation experiments.

• Read Publication →

2023 IEEE Intelligent Vehicles Symposium (IV)

This paper presents a spatio-temporal attention-based neural network that predicts driver lane-change intention by fusing video from in-cabin and forward-facing cameras. The model, which leverages a CNN-RNN architecture with a multi-head self-attention mechanism, achieves high accuracy in real-time, providing a more proactive approach to driver assistance.

• Read Publication →

IEEE Internet of Things Journal

This paper develops a driver digital twin (DDT) for online prediction of personalized lane-change behavior in a mixed traffic environment. The DDT is deployed on a vehicle-edge-cloud architecture and is shown to be able to recognize lane-change intention 6 seconds in advance and improve prediction accuracy by 27.8% compared to a general model.

• Read Publication →

Sustainability

This paper proposes a bi-level strategy for dispatching a fleet of battery-electric trucks. The strategy uses routing zone partitioning and metaheuristic-based vehicle routing to solve large-scale dispatching problems with pickup and delivery. The proposed strategy is shown to reduce travel distance and time significantly in a real-world case study.

• Read Publication →

2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC)

This paper proposes a real-time learning framework for Personalized Adaptive Cruise Control (P-ACC) that incorporates driver feedback to adapt the car-following behavior. The framework uses Inverse Reinforcement Learning (IRL) to learn a driver's preferences from historical data and then uses real-time driver feedback to update the model. The proposed method is shown to reduce driver intervention by up to 62.8% in human-in-the-loop simulation experiments.

• Read Publication →

Transportation Research Record

This paper discusses the use of robotic competitions to design and test future transport systems. It presents the case of the JRC AUTOTRAC 2020, the first European robotic traffic competition for automated miniature vehicles. The competition aimed to engage a broader community in designing and testing solutions for improving traffic flow using connected and automated vehicles (CAVs).

• Read Publication →

2022 IEEE International Conference on Robotics and Automation (ICRA)

This paper proposes a hierarchical learning-based approach for online lane-change prediction in a mixed traffic environment. The method uses a Long-Short Term Memory (LSTM) network for decision prediction and Inverse Reinforcement Learning (IRL) for trajectory prediction. The proposed system is validated on a human-in-the-loop simulation platform and is shown to be able to predict lane changes 3 seconds in advance with high accuracy.

• Read Publication →

2022 IEEE International Conference on Robotics and Automation (ICRA)

This paper proposes a Personalized Adaptive Cruise Control (P-ACC) system that learns a driver's car-following preferences from historical data using Inverse Reinforcement Learning (IRL). The P-ACC system is shown to improve the accuracy of reproducing real-world driving profiles by up to 36.5% and reduce the takeover frequency by up to 93.4% compared to the Intelligent Driver Model (IDM).

• Read Publication →

2022 IEEE 25th International Conference on Intelligent Transportation Systems (ITSC)

This paper proposes a hierarchical adaptive background subtraction (BS) method for detecting moving and stationary objects at signalized intersections using roadside fish-eye cameras. The proposed method is shown to outperform existing BS algorithms in both simulation and real-world environments.

• Read Publication →

Sustainability

This paper proposes a hierarchical, eco-friendly, cooperative ramp management system for connected and automated vehicles (CAVs). The system uses a stratified ramp metering algorithm to coordinate ramp inflow rates and a model predictive control (MPC)-based algorithm for speed control of individual CAVs. The proposed system is shown to improve mobility and energy efficiency compared to conventional ramp metering and no-control scenarios.

• Read Publication →

2021 IEEE International Intelligent Transportation Systems Conference (ITSC)

This paper proposes and investigates two real-time prediction algorithms for slow-down events on freeways, based on high-resolution information from connected vehicles. The algorithms are shown to be able to predict slow-down events on average 3.51 seconds before they occur in high-density traffic.

• Read Publication →

Sustainability

This paper proposes a demand-side cooperative shared automated mobility (DC-SAM) service framework to assess the mobility and sustainability impacts of shared automated mobility services. A case study on a New York City network shows that the proposed DC-SAM service can significantly reduce the operating costs of shared automated vehicles (SAVs) and improve customer service.

• Read Publication →

2021 IEEE Conference on Technologies for Sustainability

This paper proposes a bi-level hierarchical method to optimize the dispatch of battery-electric trucks for pickup and delivery. The method considers factors like en-route opportunity charging, time windows, and real-time traffic conditions to reduce the operational cost of the fleet.

• Read Publication →

National Center for Sustainable Transportation

This report proposes a hierarchical ramp merging system for connected and automated electric vehicles (CAEVs) that uses a centralized optimal control-based approach to smooth merging flow and improve system-wide mobility. The system is shown to improve mobility by up to 115% compared to conventional ramp metering, although this does not always translate to reduced energy consumption.

• Read Publication →

2020 IEEE Intelligent Vehicles Symposium (IV)

This paper proposes a hierarchical ramp merging system for connected and automated vehicles (CAVs) that uses a centralized optimal control-based approach to generate cooperative maneuvers and control ramp inflow rate. The system is shown to improve mobility by up to 147% and fuel savings by 47% compared to conventional ramp metering.

• Read Publication →

Journal of Traffic and Transportation Engineering

This paper reviews the research progress of cooperative eco-driving strategies for connected and automated vehicles (CAVs). It analyzes the influence of vehicle, driver, traffic network, and social factors on the energy consumption of CAVs, and discusses five representative eco-driving scenarios. The paper also identifies the limitations of existing research and points out future research directions.

• Read Publication →

SAE Technical Paper

This paper proposes a dynamometer-in-the-loop (DiL) development and testing platform for emerging mobility technologies. The platform integrates a test vehicle, chassis dynamometer, and traffic simulation to evaluate the environmental impacts of these technologies. A case study of a connected eco-operation system for a plug-in hybrid electric bus (PHEB) shows that the system can save more than 13% of fuel and reduce electricity consumption by 2%.

• Read Publication →

2020 IEEE 23rd International Conference on Intelligent Transportation Systems

This paper presents a data-driven modularized modeling approach for a plug-in hybrid electric bus (PHEB) for connected and automated vehicle (CAV) applications. The model is composed of several modules, each representing a physical component of the PHEB, and each module is modeled using a Long Short-term Memory (LSTM) network. The model is trained and validated using data from extensive dynamometer-in-the-loop (DiL) testing.

• Read Publication →

2019 IEEE Intelligent Transportation Systems Conference (ITSC)

This paper reviews the state-of-the-art of coordinated ramp control with mixed traffic conditions. It proposes a system architecture for cooperative ramp control in a mixed traffic environment and reviews the key components of the proposed system, including traffic state estimation, ramp metering, driving behavior modeling, and coordination of CAVs.

• Read Publication →