Peng Zhou

Peng Zhou

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Peng Zhou has a total of 127 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    World Models for Robotic Manipulation: A Survey

    by Fangyuan Wang, Ziyuan Wang, Guorui Pei, Mengshi Zhang, Canxi Liang, Jun Hu, Zhongxuan Li, Jinsong Wu, Ning Han, Zeqing Zhang, et al.

    Robotic manipulation depends on the ability to anticipate how actions reshape objects, contacts, and scene geometry before execution. Learned world models provide this capability by predicting task‐relevant future evolution under robot intervention, yet the term now spans latent dynamics models, action‐conditioned video generators, three‐ and four‐dimensional scene predictors, physics‐informed simulators, and predictive modules inside vision‐language‐action systems. This breadth has fragmented the literature and obscured the design choices that matter for manipulation. We survey world models for robotic manipulation through three questions: what future representation is predicted, how prediction is connected to action, and when prediction is used in the robot‐learning pipeline. We operationally define a world model as an action‐conditioned predictive system and distinguish it from perception modules, inverse models, policies, rewards, and value functions. We then organize existing work into five representation families, develop a functional taxonomy that separates integrated prediction–action models from explicit predictive planners, and characterize infrastructure roles including synthetic experience generation, candidate filtering, search‐based evaluation, learned environments, and outcome verification. We further map these roles across pretraining, post‐training, and inference adaptation, review 38 manipulation datasets, and synthesize evaluation protocols for predictive fidelity, task performance, and simulator reliability. This survey shows that world models are evolving from task‐specific dynamics predictors into predictive infrastructure for robot learning, while exposing open challenges in contact modeling, hallucination control, action alignment, and benchmarking under closed‐loop use.


  • Learning to Count Without Labels: Self-Supervised Skill Abstraction and Bayesian Counting for Garment Production Monitoring

    by Zhongxuan Li, Yilin Wen, Lei Yang, Jia Pan, Peng Zhou

    Repetition counting in long-horizon manufacturing processes is critical for quality assurance and productivity monitoring. In garment production, reliable sewn piece counting is particularly challenging due to deformable materials, unstructured workflows, and high inter-worker execution variability. Existing methods depend on supervised action annotations or handcrafted rules, limiting their scalability and robustness in industrial settings. We propose a self-supervised framework for video-based sewn piece counting that learns task representations and dynamics directly from unsegmented demonstrations. The framework operates in three stages: task-state manifold learning via self-supervised visual pretraining; discrete skill discovery through information-bottleneck-guided vector quantization; and hierarchical probabilistic inference for robust progress estimation and verified counting. This joint modeling of representation, abstraction, and task dynamics enables the system to handle execution variability and partial observations without manual annotation. We evaluate our method on a large-scale real-world sewing dataset from factory environments. Our approach achieves accurate repetition counting while substantially reducing annotation requirements, closely matching or surpassing supervised baselines under challenging industrial conditions.

  • Open Access

    Failure-Aware Bimanual Teleoperation via Conservative Value Guided Assistance

    by Peng Zhou, Zhongxuan Li, Jinsong Wu, Jiaming Qi, Jun Hu, David Navarro-Alarcon, Jia Pan, Lihua Xie, Shiyao Zhang, Zeqing Zhang

    Teleoperation of high-precision manipulation is con-strained by tight success tolerances and complex contact dy-namics, which make impending failures difficult for human operators to anticipate under partial observability. This paper proposes a value-guided, failure-aware framework for bimanual teleoperation that provides compliant haptic assistance while pre-serving continuous human authority. The framework is trained entirely from heterogeneous offline teleoperation data containing both successful and failed executions. Task feasibility is mod-eled as a conservative success score learned via Conservative Value Learning, yielding a risk-sensitive estimate that remains reliable under distribution shift. During online operation, the learned success score regulates the level of assistance, while a learned actor provides a corrective motion direction. Both are integrated through a joint-space impedance interface on the master side, yielding continuous guidance that steers the operator away from failure-prone actions without overriding intent. Experimental results on contact-rich manipulation tasks demonstrate improved task success rates and reduced operator workload compared to conventional teleoperation and shared-autonomy baselines, indicating that conservative value learning provides an effective mechanism for embedding failure awareness into bilateral teleoperation. Experimental videos are available at https://www.youtube.com/watch?v=XDTsvzEkDRE

  • A Hierarchical Laser Control Strategy for Spatially Uniform Thermal Dose Delivery in Laser-Induced Thermal Therapy

    by Jinsong Wu, Muhammad Muddassir, Qihang Li, David Navarro-Alarcon, Peng Zhou

    Laser-induced hyperthermia and ablation treatments are promising modalities in oncology, ophthalmology, cardiology, urology, and dermatology, offering minimally invasive alternatives to conventional treatment approaches. Precise spatiotemporal control of laser-induced thermal energy delivery to the target, while minimizing collateral damage to surrounding healthy tissues, is the key to achieving the desired treatment efficacy. However, nonlinear bioheat transfer processes, patient-specific tissue characteristics, diverse treatment objectives, and disturbances present significant challenges in spatiotemporal laser control for achieving precise thermal dose delivery. This article introduces a novel hierarchical laser control architecture comprising two main layers: a multiobjective genetic algorithm to determine the optimal global laser irradiation strategy and a model predictive control to optimize the local thermal dose distribution. The proposed architecture was validated on the phantom and ex vivo porcine tissue. Results demonstrate the uniform and accurate spatial thermal dose delivery across various scenarios, including arbitrary target thermal dose, target treatment temperatures, target area geometries, and heterogeneous and varying biothermal properties. The significance lies in enhancing the precision and efficacy of personalized treatment in real clinical scenarios for robotic-assisted laser-induced thermal therapies, while offering a generalizable control paradigm for diverse thermal therapy modalities employing alternative heat sources.

  • BagIt! An Adaptive Dual-Arm Manipulation of Fabric Bags for Object Bagging

    by Peng Zhou, Jiaming Qi, Hongmin Wu, Chen Wang, Yizhou Chen, Zeqing Zhang

    Bagging tasks, commonly found in industrial scenarios, are challenging considering deformable bags' complicated and unpredictable nature. This paper presents an automated bagging system from the proposed adaptive Structure-of-Interest (SOI) manipulation strategy for dual robot arms. The system dynamically adjusts its actions based on real-time visual feedback, removing the need for pre-existing knowledge of bag properties. Our framework incorporates Gaussian Mixture Models (GMM) for estimating SOI states, optimization techniques for SOI generation, motion planning via Constrained Bidirectional Rapidly exploring Random Tree (CBiRRT), and dual-arm coordination using Model Predictive Control (MPC). Extensive experiments validate the capability of our system to perform precise and robust bagging across various objects, showcasing its adaptability. This work offers a new solution for robotic deformable object manipulation (DOM), particularly in automated bagging tasks


  • AuSoScan: Automatic Scoliosis Assessment by Ultrasound Scanning With Soft Contact Control

    by Anqing Duan, Wanli Liuchen, Peng Zhou, Dezhen Song, Chenguang Yang, Yongping Zheng, David Navarro-Alarcon

    In this article, we present AuSoScan, a novel robotic platform that enables automatic scoliosis assessment through ultrasound scanning. Ultrasound imaging has been widely adopted in physical examinations due to its many benefits, such as being radiation-free, cost-effective, and highly portable. However, the scanning procedure is often tedious and labor-intensive, requiring sonographers to perform repetitive manual scanning tasks. With the fast development of robotic technologies in medical and healthcare applications, robotic ultrasound imaging presents a promising solution by combining the strengths of both robotic systems and ultrasonic devices. This has driven the development of AuSoScan, a platform specifically aimed at diagnosing scoliosis, a condition characterized by an abnormal lateral curvature of the spine. The hardware of the research platform consists of an ultrasound probe, a robotic arm, a depth camera, a force/torque sensor, and a workstation. The software architecture of the system comprises an ultrasonic perception model, a control system, and a 3-D spinal image reconstruction program. The control system of AuSoScan is implemented using model predictive control, explicitly accounting for the soft contact between the ultrasound probe and the patient’s back to enable precise force control. The effectiveness of AuSoScan is evaluated by real-world experiments of assessing the scoliosis on a phantom.


  • A Coarse-to-Fine Robotic Fabric Alignment System Integrating Visual Servoing and Admittance Control

    by Jiaming Qi, Liang Lu, Lei Yang, Yan Ding, Pai Zheng, David Navarro-Alarcon, Jia Pan, Peng Zhou

    Fabric alignment is essential to key production processes such as cutting, sewing, and fusing in garment manufacturing. Traditionally, this task has relied heavily on the dexterity and expertise of skilled human workers. Although automated systems have been introduced, they often lack the flexibility required for complex alignment tasks. In this paper, we present a novel robotic fabric alignment framework that fully automates the process with high precision and adaptability. First, we propose a coarse-to-fine alignment strategy, where an initial imprecise target position is roughly computed based on a basic perception module and eye-to-hand calibration. This is followed by a sliding mode control (SMC)-based visual servoing approach (in an eye-in-hand configuration) to ensure a close-up view of feedback features for the fine alignment process. Additionally, we consider system disturbances estimated by a fuzzy logic system (FLS) and combine it with the controller to further enhance the system’s robustness. Finally, we developed an advanced end-effector equipped with force/torque (F/T) sensors and air-powered needle grippers for gentle fabric manipulation using admittance control. We validate our framework through a series of experiments that demonstrate its effectiveness in fabric alignment tasks.