An autonomous mobile robot intelligent decision-making method and system based on reinforcement learning
By optimizing state-space modeling and designing a composite reward function, and combining an improved DQN algorithm with an attention mechanism, a distributed collaborative decision-making framework is constructed. This framework solves the path planning and cluster scheduling problems of AMRs in complex dynamic environments, achieving efficient and stable autonomous decision-making. It is suitable for AMR operation in various scenarios such as warehousing, power, and chemical industries.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-14
- Publication Date
- 2026-07-17
AI Technical Summary
Existing AMR decision-making technologies suffer from poor path planning flexibility, delayed obstacle avoidance response, and low cluster scheduling efficiency in complex dynamic environments. Furthermore, the training and convergence speed of reinforcement learning models is slow, which fails to meet the requirements for efficient and stable operation of industrial-grade, large-scale AMR clusters.
We adopt an intelligent decision-making method for autonomous mobile robots based on reinforcement learning. By optimizing state space modeling, designing multi-dimensional composite reward functions, and constructing a distributed multi-agent collaborative decision-making framework, combined with an improved DQN algorithm and attention mechanism, we can achieve autonomous optimal path planning and efficient cluster scheduling of AMRs in complex dynamic environments.
It improves the flexibility of AMR path planning and obstacle avoidance response speed, enhances environmental adaptability, reduces training time, avoids cluster conflicts, improves overall operation efficiency, and is suitable for industrial-grade applications in multiple scenarios.
Smart Images
Figure CN122411028A_ABST