arXiv subjects

Chenguang Zhao

Publications and source records attributed to Chenguang Zhao.

16 recordsLinked to original sources

arXiv2025

Shared Control for Vehicle Lane-Changing with Uncertain Driver Behaviors

Lane changes are common yet challenging driving maneuvers that require continuous decision-making and dynamic interaction with surrounding vehicles. Relying solely on human drivers for lane-changing can lead to traffic disturbances due to the stochastic nature of human behavior and its variability under different task demands. Such uncertainties may significantly degrade traffic string stability, which is critical for suppressing disturbance propagation and ensuring smooth merging of the lane-changing vehicles. This paper presents a human-automation shared lane-changing control framework that preserves driver authority while allowing automated assistance to achieve stable maneuvers in the presence of driver's behavioral uncertainty. Human driving behavior is modeled as a Markov jump process with transitions driven by task difficulty, providing a tractable representation of stochastic state switching. Based on this model, we first design a nominal stabilizing controller that guarantees stochastic ${L}_2$ string stability under imperfect mode estimation. To further balance performance and automated effort, we then develop a Minimal Intervention Controller (MIC) that retains acceptable stability while limiting automation. Simulations using lane-changing data from the NGSIM dataset verify that the nominal controller reduces speed perturbations and shorten lane-changing time, while the MIC further reduces automated effort and enhances comfort but with moderate stability and efficiency loss. Validations on the TGSIM dataset with SAE Level 2 vehicles show that the MIC enables earlier lane changes than Level 2 control while preserving driver authority with a slight stability compromise. These findings highlight the potential of shared control strategies to balance stability, efficiency, and driver acceptance.

eess.SY↗

arXiv2025

From Micro to Macro Flow Modeling: Characterizing Heterogeneity of Mixed-Autonomy Traffic

Most autonomous-vehicles (AVs) driving strategies are designed and analyzed at the vehicle level, yet their aggregate impact on macroscopic traffic flow is still not understood, particularly the flow heterogeneity that emerges when AVs interact with human-driven vehicles (HVs). Existing validation techniques for macroscopic flow models rely on high-resolution spatiotemporal data spanning entire road segments which are rarely available for mixed-autonomy traffic. AVs record detailed Lagrangian trajectories of the ego vehicle and surrounding traffic through onboard sensors. Leveraging these Lagrangian observations to validate mixed-autonomy flow models therefore remains an open research challenge. This paper closes the gap between microscopic Lagrangian data and macroscopic Euclidean traffic models by introducing a continuous traffic-heterogeneity attribute. We represent traffic flow with two coupled conservation laws with one for vehicle number and one for the traffic attribute. Reconstruction methods are designed to derive the traffic attribute from Lagrangian vehicle trajectories. When abundant trajectory data are available, we characterize traffic heterogeneity by extracting drivers' desired speed and local behavioral uncertainty from trajectories. In data-scarce mixed traffic, we design an end-to-end mapping that infers the traffic heterogeneity solely from trajectories in the current spatiotemporal region. Experiments across multiple traffic datasets show that the proposed model effectively captures traffic heterogeneity by clustering the fundamental diagram scatter into attribute-based groups. The calibration errors of traffic flow dynamics are also reduce by 20% relative to the Aw-Rascle-Zhang model benchmark. Detailed analyses further show that the model generalizes well, maintaining nearly the same accuracy when evaluated under a variety of previously unseen traffic conditions.

eess.SY↗

arXiv2025

Physics-Informed Neural Networks for Nonlocal Flow Modeling of Connected Automated Vehicles

Connected automated vehicles (CAVs) cruising control strategies have been extensively studied at the microscopic level. CAV controllers sense and react to traffic both upstream and downstream, yet most macroscopic models still assume locality, where the desired speed only depends on local density. The nonlocal macroscopic traffic flow models that explicitly capture the ``look ahead'' and ``look behind'' nonlocal CAV dynamics remain underexplored. In this paper, we propose a Physics-informed Neural Network framework to directly learn a macroscopic non-local flow model from a generic looking-ahead looking-behind vehicle motion model, which bridges the micro-macro modeling gap. We reconstruct macroscopic traffic states from synthetic CAV trajectories generated by the proposed microscopic control designs, and then learn a non-local traffic flow model that embeds a non-local conservation law to capture the resulting look-ahead look-behind dynamics. To analyze how CAV control parameters affect nonlocal traffic flow, we conduct high-fidelity driving simulator experiments to collect human drivers' trajectory data with varying downstream and upstream visibility, which serves as a baseline for tuning CAV control gains. Our analysis validates that the learned non-local flow model predicts CAV traffic dynamics more accurately than local models, and the fundamental diagram exhibits far less scatter in the speed - density relation. We further show that the looking-ahead/looking-behind control gains mainly reshape the non-local kernels, while the macroscopic speed and non-local density relation mainly depends on the desired speed function choice of the CAV controller. Our results provide a systematic approach for learning non-local macroscopic traffic-flow models directly from generic CAV control designs.

eess.SY↗

arXiv2025

Adaptive Event-triggered Formation Control of Autonomous Vehicles

This paper presents adaptive event-triggered formation control strategies for autonomous vehicles (AVs) subject to longitudinal and lateral motion uncertainties. The proposed framework explores various vehicular formations to enable safe and efficient navigation in complex traffic scenarios, such as narrow passages, collaborative obstacle avoidance, and adaptation to cut-in maneuvers. In contrast to conventional platoon control strategies that rely on predefined communication topologies and continuous state transmission, our approach employs a sampling-based observer to reconstruct vehicle dynamics. Building upon an adaptive backstepping continuous-time controller, we design three distinct event-triggered mechanisms, each offering a different trade-off between formation tracking performance and control efficiency by reducing the frequency of control signal updates. A Lyapunov-based stability analysis is conducted to guarantee bounded tracking errors and to avoid Zeno behavior. Finally, the proposed event-triggered controllers are validated through simulations of vehicular formation in three scenarios, highlighting their impact on traffic safety and mobility.

eess.SY↗

arXiv2024

Transcranial low-level laser stimulation in near infrared-II region for brain safety and protection

Background: The use of near-infrared lasers for transcranial photobiomodulation (tPBM) offers a non-invasive method for influencing brain activity and is beneficial for various neurological conditions. Objective: To investigate the safety and neuroprotective properties of tPBM using near-infrared (NIR)-II laser stimulation. Methods: We conducted thirteen experiments involving multidimensional and quantitative methods and measured serum neurobiomarkers, performed electroencephalogram (EEG) and magnetic resonance imaging (MRI) scans, assessed executive functions, and collected a subjective questionnaire. Results: Significant reductions (n=15) in neuron specific enolase (NSE) levels were observed after treatment, indicating neuroprotective effects. No structural or functional brain abnormalities were observed, confirming the safety of tPBM. Additionally, cognitive and executive functions were not impaired, with participants' feedback indicating minimal discomfort. Conclusions: Our data indicate that NIR-II tPBM is safe with specific parameters, highlighting its potential for brain protection.

q-bio.NC↗

arXiv2024

Leveraging Cooperative Connected Automated Vehicles for Mixed Traffic Safety

The introduction of connected and automated vehicles (CAV) is believed to reduce congestion, enhance safety, and improve traffic efficiency. Numerous research studies have focused on controlling pure CAV platoons in fully connected automated traffic, as well as single or multiple CAVs in mixed traffic with human-driven vehicles (HVs). CAV cruising control designs have been proposed to stabilize the car-following traffic dynamics, but few studies has considered their safety impact, particularly the trade-offs between stability and safety. In this paper, we study how cooperative control strategies for CAVs can be designed to enhance the safety and smoothness of mixed traffic under varying penetrations of connectivity and automation. Considering mixed traffic where a pair of CAVs travels amongst HVs, we design cooperative feedback controllers for the pair CAVs to stabilize traffic via cooperation and, possibly, by also leveraging connectivity with HVs. The real-time safety impact of the CAV controllers is investigated using control barrier functions (CBF). We construct CBF safety constraints, based on which we propose safety-critical control designs to guarantee CAV safety, HV safety and platoon safety. Both theoretical and numerical analyses have been conducted to explore the effect of CAV cooperation and HV connectivity on stability and safety. Our results show that the cooperation of CAVs helps to stabilize the mixed traffic while safety can be guaranteed with the safety filters. Moreover, connectivity between CAVs and HVs offers additional benefits: if an HV connects to an upstream CAV (i.e., the CAV looks ahead), it helps the CAV to stabilize the upstream traffic, while if an HV connects to a downstream CAV (i.e., the CAV looks behind), the safety of this connected HV can be enhanced.

eess.SY↗

arXiv2023

Learning "Look-Ahead" Nonlocal Traffic Dynamics in a Ring Road

The macroscopic traffic flow model is widely used for traffic control and management. To incorporate drivers' anticipative behaviors and to remove impractical speed discontinuity inherent in the classic Lighthill-Whitham-Richards (LWR) traffic model, nonlocal partial differential equation (PDE) models with ``look-ahead" dynamics have been proposed, which assume that the speed is a function of weighted downstream traffic density. However, it lacks data validation on two important questions: whether there exist nonlocal dynamics, and how the length and weight of the ``look-ahead" window affect the spatial temporal propagation of traffic densities. In this paper, we adopt traffic trajectory data from a ring-road experiment and design a physics-informed neural network to learn the fundamental diagram and look-ahead kernel that best fit the data, and reinvent a data-enhanced nonlocal LWR model via minimizing the loss function combining the data discrepancy and the nonlocal model discrepancy. Results show that the learned nonlocal LWR yields a more accurate prediction of traffic wave propagation in three different scenarios: stop-and-go oscillations, congested, and free traffic. We first demonstrate the existence of ``look-ahead" effect with real traffic data. The optimal nonlocal kernel is found out to take a length of around 35 to 50 meters, and the kernel weight within 5 meters accounts for the majority of the nonlocal effect. Our results also underscore the importance of choosing a priori physics in machine learning models.

cs.LG↗

arXiv2023

Robust Safety for Mixed-Autonomy Traffic with Delays and Disturbances

Various control strategies and field experiments have been designed for connected and automated vehicles (CAVs) to stabilize mixed traffic that contains both CAVs and Human-driven Vehicles (HVs). The effect of these stabilizing CAV control strategies on traffic safety is still under investigation. In an effort to prioritize safety over stability, a safety-critical filter via control barrier functions (CBFs) can be designed by modifying the stabilizing nominal control input in a minimal fashion and imparting collision-free driving behaviors for CAVs and HVs. However, such formal safety guarantees can be violated if there are delays in the actuation and communication channels of the CAV. Considering both actuator and sensor delays, and disturbances, we propose robust safety-critical traffic control (RSTC) design to ensure ``robust safety'' of the mixed traffic. While predictor-based CBF has been developed to compensate for the actuator delay, uncertain speed disturbances from the head vehicle cause prediction error and require novel robust CBF design. Besides, safety-critical control with sensor delay also remains an open question. In RSTC, a state predictor with bounded error is designed, and robust CBF constraints are constructed to guarantee safety under actuator delay and disturbances. When there is a sensor delay, a state observer is designed and integrated with a predictor-based CBF to ensure robust safety. Numerical simulations demonstrate that the proposed RSTC avoids rear-end collisions for two unsafe traffic scenarios in the presence of actuator, sensor delays and disturbances.

eess.SY↗

arXiv2023

Safety-Critical Traffic Control by Connected Automated Vehicles

Connected automated vehicles (CAVs) have shown great potential in improving traffic throughput and stability. Although various longitudinal control strategies have been developed for CAVs to achieve string stability in mixed-autonomy traffic systems, the potential impact of these controllers on safety has not yet been fully addressed. This paper proposes safety-critical traffic control (STC) by CAVs -- a strategy that allows a CAV to stabilize the traffic behind it, while maintaining safety relative to both the preceding vehicle and the following connected human-driven vehicles (HDVs). Specifically, we utilize control barrier functions (CBFs) to impart collision-free behavior with formal safety guarantees to the closed-loop system. The safety of both the CAV and HDVs is incorporated into the framework through a quadratic program-based controller, that minimizes deviation from a nominal stabilizing traffic controller subject to CBF-based safety constraints. Considering that some state information of the following HDVs may be unavailable to the CAV, we employ state observer-based CBFs for STC. Finally, we conduct extensive numerical simulations -- that include vehicle trajectories from real data -- to demonstrate the efficacy of the proposed approach in achieving string stable and, at the same time, provably safe traffic.

eess.SY↗

arXiv2022

A Spatial-Temporal Attention Multi-Graph Convolution Network for Ride-Hailing Demand Prediction Based on Periodicity with Offset

Ride-hailing service is becoming a leading part in urban transportation. To improve the efficiency of ride-hailing service, accurate prediction of transportation demand is a fundamental challenge. In this paper, we tackle this problem from both aspects of network structure and data-set formulation. For network design, we propose a spatial-temporal attention multi-graph convolution network (STA-MGCN). A spatial-temporal layer in STA-MGCN is developed to capture the temporal correlations by temporal attention mechanism and temporal gate convolution, and the spatial correlations by multigraph convolution. A feature cluster layer is introduced to learn latent regional functions and to reduce the computation burden. For the data-set formulation, we develop a novel approach which considers the transportation feature of periodicity with offset. Instead of only using history data during the same time period, the history order demand in forward and backward neighboring time periods from yesterday and last week are also included. Extensive experiments on the three real-world datasets of New-York, Chicago and Chengdu show that the proposed algorithm achieves the state-of-the-art performance for ride-hailing demand prediction.

cs.LG↗

arXiv2020

PDLight: A Deep Reinforcement Learning Traffic Light Control Algorithm with Pressure and Dynamic Light Duration

Existing ineffective and inflexible traffic light control at urban intersections can often lead to congestion in traffic flows and cause numerous problems, such as long delay and waste of energy. How to find the optimal signal timing strategy is a significant challenge in urban traffic management. In this paper, we propose PDlight, a deep reinforcement learning (DRL) traffic light control algorithm with a novel reward as PRCOL (Pressure with Remaining Capacity of Outgoing Lane). Serving as an improvement over the pressure used in traffic control algorithms, PRCOL considers not only the number of vehicles on the incoming lane but also the remaining capacity of the outgoing lane. Simulation results using both synthetic and real-world data-sets show that the proposed PDlight yields lower average travel time compared with several state-of-the-art algorithms, PressLight and Colight, under both fixed and dynamic green light duration.

cs.LG↗

arXiv2020

A Traffic Light Dynamic Control Algorithm with Deep Reinforcement Learning Based on GNN Prediction

Today's intelligent traffic light control system is based on the current road traffic conditions for traffic regulation. However, these approaches cannot exploit the future traffic information in advance. In this paper, we propose GPlight, a deep reinforcement learning (DRL) algorithm integrated with graph neural network (GNN) , to relieve the traffic congestion for multi-intersection intelligent traffic control system. In GPlight, the graph neural network (GNN) is first used to predict the future short-term traffic flow at the intersections. Then, the results of traffic flow prediction are used in traffic light control, and the agent combines the predicted results with the observed current traffic conditions to dynamically control the phase and duration of the traffic lights at the intersection. Experiments on both synthetic and two real-world data-sets of Hangzhou and New-York verify the effectiveness and rationality of the GPlight algorithm.

cs.LG↗

arXiv2019

Renormalization Scheme Ambiguities in the Models with More than One Coupling

The process of renormalization to eliminate divergences arising in quantum field theory is not uniquely defined; one can always perform a finite renormalization, rendering finite perturbative results ambiguous. The consequences of making such finite renormalizations have been examined in the case of there being one or two couplings. In this paper we consider how finite renormalizations can affect more general models in which there are more than two couplings. In particular, we consider the Standard Model in which there are essentially five couplings. We show that in this model (when neglecting all mass parameters) if we use mass independent renormalization, then the renormalization group beta-functions are not unique beyond one loop order, that it is not in general possible to eliminate all terms beyond certain order for all these beta-functions, but that for a physical process all contributions beyond one loop order can be subsumed into the beta-functions.

hep-ph↗

arXiv2018

Multiple Couplings and Renormalization Scheme Ambiguities

The ambiguities inherent in renormalization are considered when using mass-independent renormalization in massless theories that involve two coupling coupling constants. We review how there is no renormalization scheme in which the beta-functions can be chosen to vanish beyond a certain order in perturbation theory, but that the beta-functions always contain ambiguities beyond first order. We examine how the coupling constants depend on the coefficients of the beta-function beyond one loop order. A way of characterizing renormalization schemes that doesn't use coefficients of the beta-function is considered for models with either one or two couplings. The renormalization scheme ambiguities of physical quantities computed to finite order in perturbation theory are also examined. We demonstrate how summation of the logarithms that have explicit dependence on the renormalization scale parameter mu in a physical quantity R leads to a cancellation with the implicit dependence of R on mu through the running couplings. It is also shown that there exists a renormalization scheme in which all radiative effects beyond lowest order are incorporated into the behaviour of the running couplings.

hep-ph↗

arXiv2017

Quantizing the Palatini action using a transverse traceless propagator

We consider the first order form of the Einstein-Hilbert action and quantize it using the path integral. Two gauge fixing conditions are imposed so that the graviton propagator is both traceless and transverse. It is shown that these two gauge conditions result in two complex Fermionic vector ghost fields and one real Bosonic vector ghost field. All Feynman diagrams to any order in perturbation theory can be constructed from two real Bosonic fields, two Fermionic ghost fields and one real Bosonic ghost field that propagate. These five fields interact through just five three point vertices and one four point vertex.

hep-th↗

arXiv2016

Light Front Quantization with the Light Cone Gauge

The Dirac procedure for dealing with constraints is applied to the quantization of gauge theories on the light front. The light cone gauge is used in conjunction with the first class constraints that arise and the resulting Dirac brackets are found. These gauge conditions are not used to eliminate degrees of freedom from the action prior to applying the Dirac constraint procedure. This approach is illustrated by considering Yang-Mills theory and the superparticle in a 2 + 1 dimensional target space.

hep-th↗