Jianming Hu

Haoyi Niu, Shubham Sharma, Yiwen Qiu et al. · tsinghua

Learning effective reinforcement learning (RL) policies to solve real-world complex tasks can be quite challenging without a high-fidelity simulation environment. In most cases, we are only given imperfect simulators with simplified dynamics, which inevitably lead to severe sim-to-real gaps in RL policy learning. The recently emerged field of offline RL provides another possibility to learn policies directly from pre-collected historical data. However, to achieve reasonable performance, existing offline RL algorithms need impractically large offline data with sufficient state-action space coverage for training. This brings up a new question: is it possible to combine learning from limited real data in offline RL and unrestricted exploration through imperfect simulators in online RL to address the drawbacks of both approaches? In this study, we propose the Dynamics-Aware Hybrid Offline-and-Online Reinforcement Learning (H2O) framework to provide an affirmative answer to this question. H2O introduces a dynamics-aware policy evaluation scheme, which adaptively penalizes the Q function learning on simulated state-action pairs with large dynamics gaps, while also simultaneously allowing learning from a fixed real-world dataset. Through extensive simulation and real-world tasks, as well as theoretical analysis, we demonstrate the superior performance of H2O against other cross-domain online and offline RL algorithms. H2O provides a brand new hybrid offline-and-online RL paradigm, which can potentially shed light on future RL algorithm design for solving practical real-world tasks.

Haoyi Niu, Qimao Chen, Yingyue Li et al. · tsinghua

The deployment of autonomous vehicles (AVs) has faced hurdles due to the dominance of rare but critical corner cases within the long-tail distribution of driving scenarios, which negatively affects their overall performance. To address this challenge, adversarial generation methods have emerged as a class of efficient approaches to synthesize safety-critical scenarios for AV testing. However, these generated scenarios are often underutilized for AV training, resulting in the potential for continual AV policy improvement remaining untapped, along with a deficiency in the closed-loop design needed to achieve it. Therefore, we tailor the Stackelberg Driver Model (SDM) to accurately characterize the hierarchical nature of vehicle interaction dynamics, facilitating iterative improvement by engaging background vehicles (BVs) and AV in a sequential game-like interaction paradigm. With AV acting as the leader and BVs as followers, this leader-follower modeling ensures that AV would consistently refine its policy, always taking into account the additional information that BVs play the best response to challenge AV. Extensive experiments have shown that our algorithm exhibits superior performance compared to several baselines especially in higher dimensional scenarios, leading to substantial advancements in AV capabilities while continually generating progressively challenging scenarios. Code is available at https://github.com/BlueCat-de/SDM.

Silvio Giancola, Anthony Cioppa, Adrien Deliège et al.

The SoccerNet 2022 challenges were the second annual video understanding challenges organized by the SoccerNet team. In 2022, the challenges were composed of 6 vision-based tasks: (1) action spotting, focusing on retrieving action timestamps in long untrimmed videos, (2) replay grounding, focusing on retrieving the live moment of an action shown in a replay, (3) pitch localization, focusing on detecting line and goal part elements, (4) camera calibration, dedicated to retrieving the intrinsic and extrinsic camera parameters, (5) player re-identification, focusing on retrieving the same players across multiple views, and (6) multiple object tracking, focusing on tracking players and the ball through unedited video streams. Compared to last year's challenges, tasks (1-2) had their evaluation metrics redefined to consider tighter temporal accuracies, and tasks (3-6) were novel, including their underlying data and annotations. More information on the tasks, challenges and leaderboards are available on https://www.soccer-net.org. Baselines and development kits are available on https://github.com/SoccerNet.

Haoyi Niu, Kun Ren, Yizhou Xu et al. · tsinghua

Autonomous driving and its widespread adoption have long held tremendous promise. Nevertheless, without a trustworthy and thorough testing procedure, not only does the industry struggle to mass-produce autonomous vehicles (AV), but neither the general public nor policymakers are convinced to accept the innovations. Generating safety-critical scenarios that present significant challenges to AV is an essential first step in testing. Real-world datasets include naturalistic but overly safe driving behaviors, whereas simulation would allow for unrestricted exploration of diverse and aggressive traffic scenarios. Conversely, higher-dimensional searching space in simulation disables efficient scenario generation without real-world data distribution as implicit constraints. In order to marry the benefits of both, it seems appealing to learn to generate scenarios from both offline real-world and online simulation data simultaneously. Therefore, we tailor a Reversely Regularized Hybrid Offline-and-Online ((Re)$^2$H2O) Reinforcement Learning recipe to additionally penalize Q-values on real-world data and reward Q-values on simulated data, which ensures the generated scenarios are both varied and adversarial. Through extensive experiments, our solution proves to produce more risky scenarios than competitive baselines and it can generalize to work with various autonomous driving models. In addition, these generated scenarios are also corroborated to be capable of fine-tuning AV performance.

Zhihang Song, Zimin He, Xingyu Li et al.

Autonomous driving techniques have been flourishing in recent years while thirsting for huge amounts of high-quality data. However, it is difficult for real-world datasets to keep up with the pace of changing requirements due to their expensive and time-consuming experimental and labeling costs. Therefore, more and more researchers are turning to synthetic datasets to easily generate rich and changeable data as an effective complement to the real world and to improve the performance of algorithms. In this paper, we summarize the evolution of synthetic dataset generation methods and review the work to date in synthetic datasets related to single and multi-task categories for to autonomous driving study. We also discuss the role that synthetic dataset plays the evaluation, gap test, and positive effect in autonomous driving related algorithm testing, especially on trustworthiness and safety aspects. Finally, we discuss general trends and possible development directions. To the best of our knowledge, this is the first survey focusing on the application of synthetic datasets in autonomous driving. This survey also raises awareness of the problems of real-world deployment of autonomous driving technology and provides researchers with a possible solution.

Haoyi Niu, Tianying Ji, Bingqi Liu et al. · tsinghua

Solving real-world complex tasks using reinforcement learning (RL) without high-fidelity simulation environments or large amounts of offline data can be quite challenging. Online RL agents trained in imperfect simulation environments can suffer from severe sim-to-real issues. Offline RL approaches although bypass the need for simulators, often pose demanding requirements on the size and quality of the offline datasets. The recently emerged hybrid offline-and-online RL provides an attractive framework that enables joint use of limited offline data and imperfect simulator for transferable policy learning. In this paper, we develop a new algorithm, called H2O+, which offers great flexibility to bridge various choices of offline and online learning methods, while also accounting for dynamics gaps between the real and simulation environment. Through extensive simulation and real-world robotics experiments, we demonstrate superior performance and flexibility over advanced cross-domain online and offline RL algorithms.

Haoyi Niu, Qimao Chen, Tenglong Liu et al. · tsinghua

Reusing pre-collected data from different domains is an appealing solution for decision-making tasks, especially when data in the target domain are limited. Existing cross-domain policy transfer methods mostly aim at learning domain correspondences or corrections to facilitate policy learning, such as learning task/domain-specific discriminators, representations, or policies. This design philosophy often results in heavy model architectures or task/domain-specific modeling, lacking flexibility. This reality makes us wonder: can we directly bridge the domain gaps universally at the data level, instead of relying on complex downstream cross-domain policy transfer procedures? In this study, we propose the Cross-Domain Trajectory EDiting (xTED) framework that employs a specially designed diffusion model for cross-domain trajectory adaptation. Our proposed model architecture effectively captures the intricate dependencies among states, actions, and rewards, as well as the dynamics patterns within target data. Edited by adding noises and denoising with the pre-trained diffusion model, source domain trajectories can be transformed to align with target domain properties while preserving original semantic information. This process effectively corrects underlying domain gaps, enhancing state realism and dynamics reliability in source data, and allowing flexible integration with various single-domain and cross-domain downstream policy learning methods. Despite its simplicity, xTED demonstrates superior performance in extensive simulation and real-robot experiments.

Haoyi Niu, Yizhou Xu, Xingjian Jiang et al. · tsinghua

The safety of autonomous vehicles (AV) has been a long-standing top concern, stemming from the absence of rare and safety-critical scenarios in the long-tail naturalistic driving distribution. To tackle this challenge, a surge of research in scenario-based autonomous driving has emerged, with a focus on generating high-risk driving scenarios and applying them to conduct safety-critical testing of AV models. However, limited work has been explored on the reuse of these extensive scenarios to iteratively improve AV models. Moreover, it remains intractable and challenging to filter through gigantic scenario libraries collected from other AV models with distinct behaviors, attempting to extract transferable information for current AV improvement. Therefore, we develop a continual driving policy optimization framework featuring Closed-Loop Individualized Curricula (CLIC), which we factorize into a set of standardized sub-modules for flexible implementation choices: AV Evaluation, Scenario Selection, and AV Training. CLIC frames AV Evaluation as a collision prediction task, where it estimates the chance of AV failures in these scenarios at each iteration. Subsequently, by re-sampling from historical scenarios based on these failure probabilities, CLIC tailors individualized curricula for downstream training, aligning them with the evaluated capability of AV. Accordingly, CLIC not only maximizes the utilization of the vast pre-collected scenario library for closed-loop driving policy optimization but also facilitates AV improvement by individualizing its training with more challenging cases out of those poorly organized scenarios. Experimental results clearly indicate that CLIC surpasses other curriculum-based training strategies, showing substantial improvement in managing risky scenarios, while still maintaining proficiency in handling simpler cases.

Haoyi Niu, Jianming Hu, Guyue Zhou et al. · tsinghua

The burgeoning fields of robot learning and embodied AI have triggered an increasing demand for large quantities of data. However, collecting sufficient unbiased data from the target domain remains a challenge due to costly data collection processes and stringent safety requirements. Consequently, researchers often resort to data from easily accessible source domains, such as simulation and laboratory environments, for cost-effective data acquisition and rapid model iteration. Nevertheless, the environments and embodiments of these source domains can be quite different from their target domain counterparts, underscoring the need for effective cross-domain policy transfer approaches. In this paper, we conduct a systematic review of existing cross-domain policy transfer methods. Through a nuanced categorization of domain gaps, we encapsulate the overarching insights and design considerations of each problem setting. We also provide a high-level discussion about the key methodologies used in cross-domain policy transfer problems. Lastly, we summarize the open challenges that lie beyond the capabilities of current paradigms and discuss potential future directions in this field.

Jingxuan Yang, Ruoxuan Bai, Haoyuan Ji et al. · tsinghua

The assessment of safety performance plays a pivotal role in the development and deployment of connected and automated vehicles (CAVs). A common approach involves designing testing scenarios based on prior knowledge of CAVs (e.g., surrogate models), conducting tests in these scenarios, and subsequently evaluating CAVs' safety performances. However, substantial differences between CAVs and the prior knowledge can significantly diminish the evaluation efficiency. In response to this issue, existing studies predominantly concentrate on the adaptive design of testing scenarios during the CAV testing process. Yet, these methods have limitations in their applicability to high-dimensional scenarios. To overcome this challenge, we develop an adaptive testing environment that bolsters evaluation robustness by incorporating multiple surrogate models and optimizing the combination coefficients of these surrogate models to enhance evaluation efficiency. We formulate the optimization problem as a regression task utilizing quadratic programming. To efficiently obtain the regression target via reinforcement learning, we propose the dense reinforcement learning method and devise a new adaptive policy with high sample efficiency. Essentially, our approach centers on learning the values of critical scenes displaying substantial surrogate-to-real gaps. The effectiveness of our method is validated in high-dimensional overtaking scenarios, demonstrating that our approach achieves notable evaluation efficiency.

Shu Li, Jingxuan Yang, Honglin He et al. · tsinghua

Testing and evaluating the safety performance of autonomous vehicles (AVs) is essential before the large-scale deployment. Practically, the number of testing scenarios permissible for a specific AV is severely limited by tight constraints on testing budgets and time. With the restrictions imposed by strictly restricted numbers of tests, existing testing methods often lead to significant uncertainty or difficulty to quantifying evaluation results. In this paper, we formulate this problem for the first time the "few-shot testing" (FST) problem and propose a systematic framework to address this challenge. To alleviate the considerable uncertainty inherent in a small testing scenario set, we frame the FST problem as an optimization problem and search for the testing scenario set based on neighborhood coverage and similarity. Specifically, under the guidance of better generalization ability of the testing scenario set on AVs, we dynamically adjust this set and the contribution of each testing scenario to the evaluation result based on coverage, leveraging the prior information of surrogate models (SMs). With certain hypotheses on SMs, a theoretical upper bound of evaluation error is established to verify the sufficiency of evaluation accuracy within the given limited number of tests. The experiment results on cut-in scenarios demonstrate a notable reduction in evaluation error and variance of our method compared to conventional testing methods, especially for situations with a strict limit on the number of scenarios.

Dongxiu Liu, Haoyi Niu, Zhihao Wang et al. · tsinghua

Current robotic planning methods often rely on predicting multi-frame images with full pixel details. While this fine-grained approach can serve as a generic world model, it introduces two significant challenges for downstream policy learning: substantial computational costs that hinder real-time deployment, and accumulated inaccuracies that can mislead action extraction. Planning with coarse-grained subgoals partially alleviates efficiency issues. However, their forward planning schemes can still result in off-task predictions due to accumulation errors, leading to misalignment with long-term goals. This raises a critical question: Can robotic planning be both efficient and accurate enough for real-time control in long-horizon, multi-stage tasks? To address this, we propose a Latent Space Backward Planning scheme (LBP), which begins by grounding the task into final latent goals, followed by recursively predicting intermediate subgoals closer to the current state. The grounded final goal enables backward subgoal planning to always remain aware of task completion, facilitating on-task prediction along the entire planning horizon. The subgoal-conditioned policy incorporates a learnable token to summarize the subgoal sequences and determines how each subgoal guides action extraction. Through extensive simulation and real-robot long-horizon experiments, we show that LBP outperforms existing fine-grained and forward planning methods, achieving SOTA performance. Project Page: https://lbp-authors.github.io

Ruibo Ming, Jingwei Wu, Zhewei Huang et al.

Recent advancements in auto-regressive large language models (LLMs) have led to their application in video generation. This paper explores the use of Large Vision Models (LVMs) for video continuation, a task essential for building world models and predicting future frames. We introduce ARCON, a scheme that alternates between generating semantic and RGB tokens, allowing the LVM to explicitly learn high-level structural video information. We find high consistency in the RGB images and semantic maps generated without special design. Moreover, we employ an optical flow-based texture stitching method to enhance visual quality. Experiments in autonomous driving scenarios show that our model can consistently generate long videos.

Dingyi Yao, Xinyao Han, Ruibo Ming et al.

Ensuring the reliability of autonomous driving perception systems requires extensive environment-based testing, yet real-world execution is often impractical. Synthetic datasets have therefore emerged as a promising alternative, offering advantages such as cost-effectiveness, bias free labeling, and controllable scenarios. However, the domain gap between synthetic and real-world datasets remains a major obstacle to model generalization. To address this challenge from a data-centric perspective, this paper introduces a profile extraction and discovery framework for characterizing the style profiles underlying both synthetic and real image datasets. We propose Style Embedding Distribution Discrepancy (SEDD) as a novel evaluation metric. Our framework combines Gram matrix-based style extraction with metric learning optimized for intra-class compactness and inter-class separation to extract style embeddings. Furthermore, we establish a benchmark using publicly available datasets. Experiments are conducted on a variety of datasets and sim-to-real methods, and the results show that our method is capable of quantifying the synthetic-to-real gap. This work provides a standardized profiling-based quality control paradigm that enables systematic diagnosis and targeted enhancement of synthetic datasets, advancing future development of data-driven autonomous driving systems.

Guan Wang, Haoyi Niu, Jianxiong Li et al. · tsinghua

Among various branches of offline reinforcement learning (RL) methods, goal-conditioned supervised learning (GCSL) has gained increasing popularity as it formulates the offline RL problem as a sequential modeling task, therefore bypassing the notoriously difficult credit assignment challenge of value learning in conventional RL paradigm. Sequential modeling, however, requires capturing accurate dynamics across long horizons in trajectory data to ensure reasonable policy performance. To meet this requirement, leveraging large, expressive models has become a popular choice in recent literature, which, however, comes at the cost of significantly increased computation and inference latency. Contradictory yet promising, we reveal that lightweight models as simple as shallow 2-layer MLPs, can also enjoy accurate dynamics consistency and significantly reduced sequential modeling errors against large expressive models by adopting a simple recursive planning scheme: recursively planning coarse-grained future sub-goals based on current and target information, and then executes the action with a goal-conditioned policy learned from data rela-beled with these sub-goal ground truths. We term our method Recursive Skip-Step Planning (RSP). Simple yet effective, RSP enjoys great efficiency improvements thanks to its lightweight structure, and substantially outperforms existing methods, reaching new SOTA performances on the D4RL benchmark, especially in multi-stage long-horizon tasks.

Zhihang Song, Lihui Peng, Jianming Hu et al.

The multi-modal perception methods are thriving in the autonomous driving field due to their better usage of complementary data from different sensors. Such methods depend on calibration and synchronization between sensors to get accurate environmental information. There have already been studies about space-alignment robustness in autonomous driving object detection process, however, the research for time-alignment is relatively few. As in reality experiments, LiDAR point clouds are more challenging for real-time data transfer, our study used historical frames of LiDAR to better align features when the LiDAR data lags exist. We designed a Timealign module to predict and combine LiDAR features with observation to tackle such time misalignment based on SOTA GraphBEV framework.

Ruibo Ming, Zhewei Huang, Jingwei Wu et al.

Future Frame Synthesis (FFS), the task of generating subsequent video frames from context, represents a core challenge in machine intelligence and a cornerstone for developing predictive world models. This survey provides a comprehensive analysis of the FFS landscape, charting its critical evolution from deterministic algorithms focused on pixel-level accuracy to modern generative paradigms that prioritize semantic coherence and dynamic plausibility. We introduce a novel taxonomy organized by algorithmic stochasticity, which not only categorizes existing methods but also reveals the fundamental drivers--advances in architectures, datasets, and computational scale--behind this paradigm shift. Critically, our analysis identifies a bifurcation in the field's trajectory: one path toward efficient, real-time prediction, and another toward large-scale, generative world simulation. By pinpointing key challenges and proposing concrete research questions for both frontiers, this survey serves as an essential guide for researchers aiming to advance the frontiers of visual dynamic modeling.

Guan Wang, Haoyi Niu, Desheng Zhu et al.

Heated debates continue over the best autonomous driving framework. The classic modular pipeline is widely adopted in the industry owing to its great interpretability and stability, whereas the fully end-to-end paradigm has demonstrated considerable simplicity and learnability along with the rise of deep learning. As a way of marrying the advantages of both approaches, learning a semantically meaningful representation and then use in the downstream driving policy learning tasks provides a viable and attractive solution. However, several key challenges remain to be addressed, including identifying the most effective representation, alleviating the sim-to-real generalization issue as well as balancing model training cost. In this study, we propose a versatile and efficient reinforcement learning framework and build a fully functional autonomous vehicle for real-world validation. Our framework shows great generalizability to various complicated real-world scenarios and superior training efficiency against the competing baselines.

Haoyi Niu, Jianming Hu, Zheyu Cui et al.

How to explore corner cases as efficiently and thoroughly as possible has long been one of the top concerns in the context of deep reinforcement learning (DeepRL) autonomous driving. Training with simulated data is less costly and dangerous than utilizing real-world data, but the inconsistency of parameter distribution and the incorrect system modeling in simulators always lead to an inevitable Sim2real gap, which probably accounts for the underperformance in novel, anomalous and risky cases that simulators can hardly generate. Domain Randomization(DR) is a methodology that can bridge this gap with little or no real-world data. Consequently, in this research, an adversarial model is put forward to robustify DeepRL-based autonomous vehicles trained in simulation to gradually surfacing harder events, so that the models could readily transfer to the real world.

Haoyi Niu, Jianming Hu, Zheyu Cui et al.

Increasing the response time of emergency vehicles(EVs) could lead to an immeasurable loss of property and life. On this account, tactical decision making for EVs' microscopic control remains an indispensable issue to be improved. In this paper, a rule-based avoiding strategy(AS) is devised, that CVs in the prioritized zone ahead of EV should accelerate or change their lane to avoid it. Besides, a novel DQN method with speed-adaptive compact state space (SC-DQN) is put forward to fit in EVs' high-speed feature and generalize in various road topologies. Afterward, the execution of AS feedback to the input of SC-DQN so that they joint organically as a combinational method. The following approach reveals that DRL could complement rule-based avoiding strategy in generalization, and on the contrary, the rule-based avoiding strategy could complement DRL in stability, and their combination could lead to less response time, lower collision rate and smoother trajectory.

Yusen Huo, Qinghua Tao, Jianming Hu

Traffic signal control has long been considered as a critical topic in intelligent transportation systems. Most existing learning methods mainly focus on isolated intersections and suffer from inefficient training. This paper aims at the cooperative control for large scale multi-intersection traffic signal, in which a novel end-to-end learning based model is established and the efficient training method is proposed correspondingly. In the proposed model, the input traffic status in multi-intersections is represented by a tensor, which not only significantly reduces dimensionality than using a single matrix but also avoids information loss. For the output, a multidimensional boolean vector is employed for the control policy to indicate whether the signal state changes or not, which simplifies the representation and abides the practical phase changing rules. In the proposed model, a multi-task learning structure is used to get the cooperative policy by learning. Instead of only using the reinforcement learning to train the model, we employ imitation learning to integrate a rule based model with neural networks to do the pre-training, which provides a reliable and satisfactory stage solution and greatly accelerates the convergence. Afterwards, the reinforcement learning method is adopted to continue the fine training, where proximal policy optimization algorithm is incorporated to solve the policy collapse problem in multi-dimensional output situation. In numerical experiments, the advantages of the proposed model are demonstrated with comparison to the related state-of-the-art methods.

Guan Wang, Jianming Hu, Zhiheng Li et al.

In this paper, we study how to learn an appropriate lane changing strategy for autonomous vehicles by using deep reinforcement learning. We show that the reward of the system should consider the overall traffic efficiency instead of the travel efficiency of an individual vehicle. In summary, cooperation leads to a more harmonic and efficient traffic system rather than competition

Xiaolong Ma, Xiatian Zhu, Shaogang Gong et al.

Most existing person re-identification (ReID) methods rely only on the spatial appearance information from either one or multiple person images, whilst ignore the space-time cues readily available in video or image-sequence data. Moreover, they often assume the availability of exhaustively labelled cross-view pairwise data for every camera pair, making them non-scalable to ReID applications in real-world large scale camera networks. In this work, we introduce a novel video based person ReID method capable of accurately matching people across views from arbitrary unaligned image-sequences without any labelled pairwise data. Specifically, we introduce a new space-time person representation by encoding multiple granularities of spatio-temporal dynamics in form of time series. Moreover, a Time Shift Dynamic Time Warping (TS-DTW) model is derived for performing automatically alignment whilst achieving data selection and matching between inherently inaccurate and incomplete sequences in a unified way. We further extend the TS-DTW model for accommodating multiple feature-sequences of an image-sequence in order to fuse information from different descriptions. Crucially, this model does not require pairwise labelled training data (i.e. unsupervised) therefore readily scalable to large scale camera networks of arbitrary camera pairs without the need for exhaustive data annotation for every camera pair. We show the effectiveness and advantages of the proposed method by extensive comparisons with related state-of-the-art approaches using two benchmarking ReID datasets, PRID2011 and iLIDS-VID.