Two-Legged Robot Tackles Forest Terrain
The AI-enabled robot uses reinforcement learning to react to obstacles and terrain
LimX Dynamics, a Chinese general-purpose robot company, has released a video update of its bipedal robot, showcasing its locomotion capabilities as it navigates a forest terrain.
The P1 robot uses reinforcement learning to identify and react to external stimuli, including obstacles and uneven ground in its path. It can also right itself when pushed or kicked, as demonstrated by a LimX team member.
The tests took place in Tanglang Mountain in Shenzhen, China.
“With zero-shot learning, non-protected and fully open testing conditions, P1 successfully navigated the completely strange wilderness of the forest,” the company said. “Demonstrating exceptional control and stability post reinforcement learning by dynamically locomoting over various complex terrains.”
LimX said P1 serves as an “important platform” for its development of reinforcement learning systems, with plans to deploy the tech in other robotic designs in the future.
The company plans to use these systems on its humanoid robot, CL-1, which it describes as “an ideal testing platform for embodied AI” and “one of the few humanoid robots around the world” to achieve dynamic stair climbing based on real-time terrain perception.
In a blog post, the company said CL-1 perceives its terrain in real time, using perception control to “proactively adjust” its motion and allow it to walk smoothly onto a curb, climb stairs “dynamically and fluidly” and walk down slopes.
About the Author
You May Also Like