On August 1–2, the 7th China Robotics Academic Annual Conference (CCRS 2026) was held at the National Exhibition and Convention Center (Shanghai). As the highest-profile and most comprehensive academic event in China's robotics field, the conference was themed "Intelligence Converges at Huangpu, Innovation Empowers the New Era." It brought together leading experts, research teams, and industry pioneers from top universities and institutions worldwide, featuring in-depth discussions on foundational breakthroughs, key technological challenges, and deep industry-academia-research collaboration – presenting a full spectrum of China's cutting-edge robotics research and industrial development trends.

At the conference, CHINGMU, a global leader in motion capture, co-hosted an exhibition booth with Unitree Robotics, a benchmark player in humanoid robotics. By integrating optical mocap technology with robot motion control, the joint exhibition became a standout highlight in the industry exhibition zone, drawing significant attention from attendees.

Motion Capture Dual Empowerment: Live Validation of Robot Interaction & Data Infrastructure

At the core demo zone, the robot teleoperation showcase drew a large crowd of attendees. A performer wearing a motion capture suit was tracked in real time by CHINGMU's high-precision optical mocap system – capturing full-body motion with millisecond-level latency and synchronously mapping it to the Unitree robot. As the music played, the robot precisely replicated the performer's dance movements, from body sways to footwork, achieving remarkable fluidity and sync accuracy.

Beyond its visual appeal, this demonstration validated the high reliability of optical mocap in robot teleoperation scenarios – offering a practical pathway for applications such as special-environment operations, human-robot collaborative interaction, and robot kinematics algorithm validation.

In robotic training and R&D workflows, precise and multidimensional data is the foundation for iteration. The desktop-scale data acquisition solution on display, powered by CHINGMU's Project Decode multimodal high-quality Embodied AI data acquisition system, integrates optical, inertial, and tactile sensing technologies. It synchronously captures fine-grained manipulation actions – grasping, twisting, inserting, and removing – with full joint angle and pose data in real operational scenarios, while simultaneously collecting video, first-person view, tactile, object, and environmental data. This transforms what conventionally requires hours of repeated multimodal acquisition into a single, streamlined capture, with data precision and consistency far surpassing traditional manual annotation. The immersive on-site demonstration gave attendees a tangible sense of how Embodied AI data infrastructure translates into real-world value.

Joint Exhibition: A Complete Data-to-Embodied AI Pipeline

At CCRS 2026, the CHINGMU and Unitree joint exhibition was more than two standalone demos – it represented a complete closed loop spanning data acquisition, algorithm training, teleoperation control, and robot execution. This is precisely the direction CHINGMU continues to deepen.

The two companies have already collaborated across multiple scenarios, accumulating extensive experience in adapting motion data to robot platforms – covering robotics R&D training, event technical support, and teleoperation validation. Customized mocap solutions have been delivered for humanoid and quadruped robot product lines. In the CMG World Robot Fighting Championship & UFC Elite Training Center project, CHINGMU provided pre-competition training data acquisition for the Unitree G1 robot – capturing complex combat movements from professional athletes, with synchronized recording of joint trajectories, pose angles, punch velocity, and motion coherence. This multimodal data helps robots continuously learn and evolve, advancing them toward real-world competition arenas.

Breakthroughs in Embodied AI rely on high-precision spatial perception and motion data. Looking ahead, CHINGMU will continue to partner with Unitree and other industry partners to empower full-scenario robotics innovation with spatial intelligence – providing sustained technical momentum for China's robotics research and industrial advancement, and driving the field toward higher levels of independent innovation.