XZXiaocong Zhao

Xiaocong Zhao

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Xiaocong Zhao has a total of 33 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Escaping Stability-Plasticity Dilemma in Online Continual Learning for Motion Forecasting via Synergetic Memory Rehearsal

    by Yunlong Lin, Chao Lu, Tongshuai Wu, Xiaocong Zhao, Guodong Du, Yanwei Sun, Zirui Li, Jianwei Gong

    Deep neural networks (DNNs) have achieved remarkable success in intelligent systems such as autonomous vehicles and robots. However, most DNN-based methods suffer from catastrophic forgetting, where the DNN may fail to maintain its performance in previously learned tasks after adapting to new data. This phenomenon gives rise to a critical tradeoff between memory stability of the DNN, that is, the ability to retain learned knowledge, and learning plasticity, that is, the capacity to acquire new information effectively. To address such stability-plasticity dilemma, this study proposes a novel CL method named synergetic memory rehearsal (SyReM). SyReM maintains a compact memory buffer to represent learned knowledge. To ensure memory stability, it employs an inequality constraint that limits increments in the average loss over the memory buffer. Synergistically, a selective memory rehearsal mechanism is designed to enhance learning plasticity by selecting samples from the memory buffer that are similar to recently observed data. This selection is based on an online-measured cosine similarity of loss gradients, ensuring targeted memory rehearsal. We validate SyReM under online CL tasks for motion forecasting. Comprehensive experiments on naturalistic driving datasets demonstrate that, compared to non-CL and CL baselines, SyReM significantly mitigates catastrophic forgetting in past scenarios while improving forecasting accuracy in new ones. The code is available at https://github.com/BIT-Jack/SyReM.

  • Physics-informed interaction modeling for warning targeting in multi-risk driving scenarios

    by Jian Sun, Ruoxi Kong, Xiaocong Zhao, Haoyang Liang, Xiyan Jiang, Yiru Liu

  • Open Access

    Why Some Emergency Maneuvers Succeed: Interpretable Behavior Graphs from Expert Drivers to Automated Control

    by Yixin Zhu, Xiaocong Zhao, Yiru Liu, Haotian Shi, Zirui Li, Jian Sun

    Emergency handling is among the hardest capabilities for automated driving. Safe evasion demands staged, anticipatory, and tightly coupled longitudinal--lateral actions, yet the expert logic behind such maneuvers is scarce, hard to label by skill, and locked in continuous control signals that obscure why some maneuvers succeed and others fail. This study introduces the Driving Behavior Graph (DBG), an interpretable, probabilistic representation that turns scarce expert demonstrations into reusable knowledge linking human expert decision logic to machine-executable control. Through a two-stage procedure, we screen 25 expert drivers from 99 candidates and collect 400 driver-in-the-loop trials across four emergency scenarios (jaywalking, cut-out, cut-in, merge-in) on a virtual-reality platform. Control signals are segmented with a Hierarchical Dirichlet Process Hidden Semi-Markov Model, mapped to a scenario-specific semantic grid, and assembled into a probabilistic DBG of behavior chains and transitions. The DBG shows that expert emergency behavior is proactive and phased: evasive action is initiated far earlier (lateral-maneuver onset at about 55 vs. 115 m in the cut-out scenario), longitudinal and lateral control are coordinated, and the vehicle is deliberately recovered after the conflict; it further localizes the bifurcation points where excellent and poor episodes diverge. Notably, self-assessed driving skill does not predict objective emergency performance, indicating that competence in safety-critical handling cannot be inferred from routine self-report. To close the loop, DBG-derived primitives are synthesized into reference trajectories that guide a model predictive controller, which reproduces expert-like earlier lateral evasion (p < 0.001) and a deliberate safety-over-compliance trade-off, and runs in a vehicle-in-the-loop deployment on a Level-3 automated test vehicle. To support reproducibility, the skill-screened expert emergency-driving dataset is publicly available at https://zenodo.org/records/20710939.

  • Open Access

    From Human Semantics to Executable Control: Driving Behavior Graph for Automated Emergency Maneuvering

    by Yixin Zhu, Xiaocong Zhao, Yiru Liu, Haotian Shi, Zirui Li, Jian Sun

    Emergency handling is among the hardest capabilities for automated driving. Safe evasion demands staged, anticipatory, and tightly coupled longitudinal-lateral actions, yet the expert logic behind such maneuvers is scarce, hard to label by skill, and locked in continuous control signals that obscure why some maneuvers succeed and others fail. This study introduces the Driving Behavior Graph (DBG), an interpretable, probabilistic representation that turns scarce expert demonstrations into reusable knowledge linking human behavioral semantics to machine-executable control. Through a two-stage procedure, we screen 25 expert drivers from 99 candidates and collect 400 driver-in-the-loop trials across four emergency scenarios (jaywalking, cut-out, cut-in, merge-in) on a virtual-reality platform. Control signals are segmented with a Hierarchical Dirichlet Process Hidden Semi-Markov Model, mapped to a scenario-specific semantic grid, and assembled into a probabilistic DBG of behavior chains and transitions. The DBG shows that expert emergency behavior is proactive and phased (earlier evasive initiation, coordinated longitudinal-lateral control, and post-conflict recovery), and it localizes the bifurcation points where excellent and poor episodes diverge; notably, self-assessed skill does not predict objective emergency performance. To close the loop, DBG-derived primitives are synthesized into reference trajectories that guide a model predictive controller, which reproduces expert-like earlier lateral evasion (p < 0.001) and a deliberate safety-over-compliance trade-off, and runs in a vehicle-in-the-loop deployment on a Level-3 automated test vehicle. To support reproducibility, the skill-screened expert emergency-driving dataset is publicly available at https://zenodo.org/records/20710939.

  • Open Access

    Modeling Driver-Initiated Intervention in Automated Driving: A Dual-Stage Progressive Intention-to-Action Prediction Framework

    by Tianxiang Tang, Jie Wang, Haotian Shi, Xiaocong Zhao, Jian Sun

    In conditionally automated driving, drivers may interrupt automation without an explicit takeover request, potentially disrupting the control chain and increasing safety risks. Early anticipation of such driver-initiated intervention is therefore critical for safe human-automation interaction. However, most existing methods focus on predicting intervention likelihood from observable changes in driver state or scenario risk, while the progressive evolution of latent intention and the temporal transition toward actual intervention remain insufficiently modeled. To address this issue, this study proposes a Perception-to-Prediction Dual-stage Progressive Initiated-Intervention Prediction framework, termed P2P-DPI. The proposed framework first detects latent intervention intention and then predicts the corresponding time-to-intervention (TTI), thereby reformulating intervention prediction as a continuous progressive inference problem. Specifically, a Signal Evolution and Cognitive State EvaluatOR (SENSOR) module is developed to model pre-intervention cognitive evolution as a two-stage process of intention accumulation and intention formation based on physiological-behavioral signal evolution. Once stable intention is detected, a Time-to-intervention HieRarchical Attention-based Predictor (TRAP) is activated to hierarchically integrate drivers' physiological--behavioral states, vehicle operating states, scenario risk states, and individual traits for TTI prediction. The framework is validated through human-in-the-loop automated-driving simulation experiments with 30 participants. The results show that P2P-DPI can detect the intention for driver intervention in advance with stable hit-rate and lead-time performance, while also achieving accurate TTI prediction. Overall, by linking early cognitive-state perception with downstream temporal prediction, the proposed framework provides a progressive and interpretable approach for anticipating driver-initiated intervention and supporting safer human--automation interaction in future automated driving.

  • Open Access

    Chain-Like Cognitive Processes in Human-Machine Co-Driving: Uncovering Causal-Chain Mechanisms of Driver Takeover Failure

    by Jie Wang, Haotian Shi, Yujia Zhao, Xiaocong Zhao, Haoyang Liang, Peng Hang, Jian Sun

    Takeover failure in conditionally automated driving is rarely caused by a single erroneous action; instead, it emerges through a causal chain in which risk is transmitted and accumulated across successive takeover process. Uncovering this mechanism is therefore essential for safe human-machine interaction. However, modeling the takeover process remains challenging because it involves multimodal and time-varying factors, lacks an explicit process-level mediator, and requires causal attribution under chain-like propagation. To address these challenges, this study proposes DBSE-SOR C2Net, an interpretable and inferable causal-chain network grounded in the cognitive framework of stimulus(S)-organism(O)-response(R). Specifically, the takeover reaction process (O) is introduced as the key mediator linking the pre-takeover states (S) to the post-takeover performance outcomes (R), and is further decomposed into nodes of the postural, perceptual and cognitive stage (PPC). To capture multimodal temporal characteristics, a multi-timescale encoding scheme is developed to jointly model long-horizon baselines and short-horizon perturbations in physiological and environmental signals. On this basis, a hierarchical dynamic Bayesian structural equation solution is constructed, with a bottom measurement layer for latent construct mapping and a top mechanism layer for dependency learning. Based on a human-in-the-loop takeover experiment (n = 41), the proposed framework effectively uncovers representative high-risk failure chains, including time-pressure-driven, cognitive-bottleneck, and multi-stage prolongation patterns. Under this framework , the adaptive takeover requests enhancement reduced PPC-stage reaction times by 0.11 s, 0.07 s, and 0.45 s, respectively, and shortened the overall takeover time by 0.31 s. Moreover, downstream PPC-stage monitoring with Minimum Risk Maneuver intervention reduced the collision rate by 75%. In general, the DBSE-SOR C2Net provides a backward reasoning and forward propagation paradigm for failure-chain diagnosis and adaptive safety intervention in automated driving.