Yujia Zhao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yujia Zhao has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Jian Sun
Tongji University
6 shared publications Jie Wang
4 shared publicationsHaotian Shi
3 shared publications- Xiaocong Zhao
Tongji University
3 shared publications - Haoyang Liang
The Key Laboratory of Road and Traffic Engineering, Ministry of Education, Tongji University, Shanghai, 201804, China
2 shared publications - Peng Hang
Tongji University
2 shared publications
Latest Publications
- Open Access
- 14 August 2026
- Published by Elsevier in SSRN Electronic Journal
Towards Pre-Crash Safety in Closed-Loop Autonomous Driving: A Surrogate-Safety-Informed Curriculum and Corrective Replay Framework
Autonomous driving policies trained with ordinary random sampling may under-expose long-tail hazardous interactions, leaving critical failures insufficiently corrected in closed-loop training. This limitation is particularly important for pre-crash safety, because rare high-risk scenarios such as intersection conflicts, lane-change merges, mixed-complex interactions, and rule-boundary violations disproportionately shape policy safety boundaries. This study reformulates training-sample organization as a pre-crash safety intervention and proposes a surrogate-safety-informed curriculum and corrective replay framework. The framework performs scenario-level closed-loop evaluation using at-fault collision, minimum time-to-collision (TTC), comfort degradation, off-road behavior, and red-light violations. These signals are integrated into a Pre-crash Risk Severity Index to stratify training scenarios, while failure taxonomy and interaction topology are used to diagnose hazardous events. Three interventions are compared under a common learner and training budget: risk-aware curriculum learning, targeted corrective replay, and their combination. Experiments on the Waymo Open Motion Dataset using the ScenarioMax-to-V-Max pipeline show that, with a unified BC-SAC learner and 81,920 training steps, the combined intervention achieves the best overall performance on an independent test set, including the highest accuracy, comfort, minimum TTC, and V-Max Score and the lowest at-fault collision rate. The results indicate that reallocating finite training exposure according to pre-crash risk can improve aggregate closed-loop performance and key safety surrogates without changing the backbone learner.
- 14 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessTowards Pre-Crash Safety in Closed-Loop Autonomous Driving: A Surrogate-Safety-Informed Curriculum and Corrective Replay Framework
Autonomous driving policies trained with ordinary random sampling may under-expose long-tail hazardous interactions, leaving critical failures insufficiently corrected in closed-loop training. This limitation is particularly important for pre-crash safety, because rare high-risk scenarios such as intersection conflicts, lane-change merges, mixed-complex interactions, and rule-boundary violations disproportionately shape policy safety boundaries. This study reformulates training-sample organization as a pre-crash safety intervention and proposes a surrogate-safety-informed curriculum and corrective replay framework. The framework performs scenario-level closed-loop evaluation using at-fault collision, minimum time-to-collision (TTC), comfort degradation, off-road behavior, and red-light violations. These signals are integrated into a Pre-crash Risk Severity Index to stratify training scenarios, while failure taxonomy and interaction topology are used to diagnose hazardous events. Three interventions are compared under a common learner and training budget: risk-aware curriculum learning, targeted corrective replay, and their combination. Experiments on the Waymo Open Motion Dataset using the ScenarioMax-to-V-Max pipeline show that, with a unified BC-SAC learner and 81,920 training steps, the combined intervention achieves the best overall performance on an independent test set, including the highest accuracy, comfort, minimum TTC, and V-Max Score and the lowest at-fault collision rate. The results indicate that reallocating finite training exposure according to pre-crash risk can improve aggregate closed-loop performance and key safety surrogates without changing the backbone learner.
- Open Access
- 15 May 2026
- Published by Elsevier in SSRN Electronic Journal
Incorporating human factors into bicycle microscopic simulation: Methods and applications for modelling cycling capability
Incorporation of more sophisticated human factors (HFs) into microscopic traffic simulation has become a mainstream research direction in recent years. Such models are believed to enable us to further improve the performance of simulation, and to understand the mechanisms behind many puzzling traffic phenomena. In this regard, it can not only provide more accurate traffic flow estimation for management and facility design, but also support the development of autonomous driving by providing more realistic virtual testing simulation environments in traffic. Although related discussions and models for the operational behaviour of drivers and pedestrians are already extensive, studies on incorporating HFs into microscopic simulation models of cyclists are still nascent. Most existing research still focuses on how different types, genders, and cycling styles affect interaction behaviour, and captures such characteristics by distinguishing values of model parameters. Solutions for endogenously representing and depicting human cycling behaviours remain very limited. This limitation brings bicycle simulation models’ inability in accurately capturing heterogeneous human cycling behaviours and the variation of interaction states generated by HFs, resulting in unsatisfactory simulation performance. To this end, we attempt to incorporate HFs into bicycle behaviour models to address this gap. In this paper, we first select a key HF, i.e., cycling capability, as the modelling attribute of cyclist, and accordingly propose a novel cycling “Demand-Capability” framework to endogenously quantify the influence of this HF attribute on cyclist interactions. Corresponding formulations are developed, and testing is carefully designed for the model validation. Then, we further integrate the above model into our previously developed two‑dimensional bicycle simulation model. Here, two dedicated approaches are correspondingly put forward to address the effects of HFs on the operational mental (decision) and operational physical (action) activities, respectively. Finally, we perform face validation of the proposed simulation model by comparing simulation results with observations obtained from the empirical trajectory dataset. Results show the proposed model outperforms its predecessors, especially in the presence of individual behaviour decisions and actions.
- 15 May 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessIncorporating human factors into bicycle microscopic simulation: Methods and applications for modelling cycling capability
Incorporation of more sophisticated human factors (HFs) into microscopic traffic simulation has become a mainstream research direction in recent years. Such models are believed to enable us to further improve the performance of simulation, and to understand the mechanisms behind many puzzling traffic phenomena. In this regard, it can not only provide more accurate traffic flow estimation for management and facility design, but also support the development of autonomous driving by providing more realistic virtual testing simulation environments in traffic. Although related discussions and models for the operational behaviour of drivers and pedestrians are already extensive, studies on incorporating HFs into microscopic simulation models of cyclists are still nascent. Most existing research still focuses on how different types, genders, and cycling styles affect interaction behaviour, and captures such characteristics by distinguishing values of model parameters. Solutions for endogenously representing and depicting human cycling behaviours remain very limited. This limitation brings bicycle simulation models’ inability in accurately capturing heterogeneous human cycling behaviours and the variation of interaction states generated by HFs, resulting in unsatisfactory simulation performance. To this end, we attempt to incorporate HFs into bicycle behaviour models to address this gap. In this paper, we first select a key HF, i.e., cycling capability, as the modelling attribute of cyclist, and accordingly propose a novel cycling “Demand-Capability” framework to endogenously quantify the influence of this HF attribute on cyclist interactions. Corresponding formulations are developed, and testing is carefully designed for the model validation. Then, we further integrate the above model into our previously developed two‑dimensional bicycle simulation model. Here, two dedicated approaches are correspondingly put forward to address the effects of HFs on the operational mental (decision) and operational physical (action) activities, respectively. Finally, we perform face validation of the proposed simulation model by comparing simulation results with observations obtained from the empirical trajectory dataset. Results show the proposed model outperforms its predecessors, especially in the presence of individual behaviour decisions and actions.
- Open Access
- 13 May 2026
- Published by Elsevier in SSRN Electronic Journal
Chain-Like Cognitive Processes in Human-Machine Co-Driving: Uncovering Causal-Chain Mechanisms of Driver Takeover Failure
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.
- 13 May 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessChain-Like Cognitive Processes in Human-Machine Co-Driving: Uncovering Causal-Chain Mechanisms of Driver Takeover Failure
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.