Unverified50% confidenceFactExact time
Meda modeled the moving sofa problem as a reinforcement learning environment, in which the agent controls the sofa's translation and rotation with the goal of maximizing the area that can pass through the corridor.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/24/2026
First Seen
Valid until: 1/22/2027
Sources
Related Entities
Related Claims
UnverifiedCompared to manually designed controllers, reinforcement learning can automatically discover counter-intuitive gaits that exploit spinal elasticity.59% similarUnverifiedModel-based reinforcement learning can leverage physical priors to conduct imagination rollouts, reducing the number of interactions with the real environment and avoiding dangerous states.58% similarUnverified受限空间运动规划的解决思路包括冗余自由度细长型机械臂、基于优化的规划方法（如CHOMP、TrajOpt）以及学习型策略57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1092477API
curl https://kongchang.com/api/v1/knowledge/claims/1092477MCP
get_claim(id=1092477)