Submission Deadline: August 1, 2026
Embodied Physical Intelligence
具身物理智能
| Chair: |
Co-chair: |
 |
 |
| Jinguo Huang |
Jian Wang |
| Beijing University of Posts and Telecommunications, China |
Institute of Automation, Chinese Academy of Sciences, China |
| |
|
| Topics: |
|
- Embodied Cognitive Science and Robotic Physical Intelligence (具身认知科学与机器人物理智能)
- Embodied Perception and Multimodal Fusion for Physical Interaction (物理交互具身感知与多模态融合)
- Embodied Models and Learning Methods Under Physical Constraints (物理约束下具身模型与学习方法)
- Real-time Dynamic Control and Autonomous Decision-making for Robots (机器人实时动态控制与自主决策)
- Cooperative Control and Swarm Intelligence for Unmanned Systems (无人系统的协同控制与群体智能)
|
| |
|
| Summary: |
|
- Embodied Physical Intelligence (EPI), as a cutting-edge direction at the intersection of artificial intelligence and robotics, addresses the long-standing industry challenge that traditional robots can only perform pre-programmed tasks in structured environments, leading a fundamental paradigm shift in robotics from programmed execution to autonomous learning. Unlike conventional artificial intelligence that operates solely in the digital space, the core of EPI lies in the continuous dynamic interaction between intelligent agents and the real world through physical carriers, constructing a complete intelligent closed loop of perception, understanding, decision-making, action, and feedback, and realizing the fundamental paradigm shift of intelligent unmanned systems from disembodied computation to embodied closed-loop intelligence.
Currently, the field faces a series of urgent common technical bottlenecks: unstable whole-body motion control of humanoid robots, low success rate of dexterous manipulation with robotic hands, long changeover and commissioning cycles for industrial robots, poor adaptability of special-purpose robots in complex environments, and low efficiency of multi-robot coordination. This special session focuses on the core application areas of Embodied Physical Intelligence, aiming to bring together leading scholars and industry experts from home and abroad to in-depth discuss its fundamental theories, key technologies and engineering practices. We pay special attention to robots' autonomous learning and evolution capabilities in open, dynamic and unstructured environments, as well as their innovative applications in national strategic fields such as industrial manufacturing, medical rehabilitation, emergency rescue, and national defense and military industry. This special session will provide an interdisciplinary communication platform for participants to showcase the latest research achievements, share industrial practice experiences, and jointly promote the development and deployment of Embodied Physical Intelligence technologies.
|
| |
|
- 具身物理智能作为人工智能与机器人学深度融合的前沿方向，通过解决传统机器人只能在结构化环境中执行预设任务的行业顽疾，引领机器人从编程执行向自主学习的根本性范式转变。与传统仅在数字空间运行的人工智能不同，具身物理智能的核心就是强调智能体通过物理载体与真实世界的持续动态交互，构建感知、理解、决策、行动、反馈的完整智能闭环，实现智能无人系统从离身计算向具身闭环的根本性范式转变。
当前面临着一系列亟待突破的共性技术瓶颈：人形机器人全身运动控制不稳定、灵巧手精细操作成功率低、工业机器人换产调试周期长、特种机器人在复杂环境中适应性差、多机器人协同效率低下等。本专题聚焦具身物理智能的核心应用领域，旨在汇聚国内外顶尖学者和产业界专家，深入探讨其基础理论、关键技术与工程实践。我们特别关注机器人在开放、动态、非结构化环境中的自主学习与进化能力，以及在工业制造、医疗康复、应急救援、国防军工等国家战略领域的创新应用。本专题将为参会者提供一个跨学科交流平台，展示最新研究成果，分享产业实践经验，共同推动具身物理智能技术的发展与落地。
|