Robot Motion Planning Around Dynamic Obstacles

The patent transforms the traditional 2D spatial motion planning problem into a 3D state space problem by adding the time dimension. The state space is defined as S = (x, y, t) where (x, y) represents spatial position and t represents time. This dimensional transformation allows the robot to plan trajectories that avoid dynamic obstacles by selecting appropriate time stamps for each spatial position, thereby reducing collision risk while maintaining manageable planning complexity through structured state representation.

The patent implements dynamic motion planning by making the state space and trajectories adaptive to changing obstacle positions. The robot observes dynamic obstacles and continuously updates its planned trajectories based on current obstacle positions and predicted future positions. This dynamic approach allows the motion plan to adapt in real-time, reducing collision risk with moving obstacles while maintaining operational efficiency.

The patent dynamically adjusts the safety distance parameter based on the robot's speed, obstacle type, and environmental context. Rather than maintaining a fixed conservative distance that would reduce efficiency, the system modifies the safety threshold parameter adaptively - using larger distances when risk is high and smaller distances when safe, thereby maintaining both collision avoidance reliability and motion efficiency.

The patent applies safety distance maintenance selectively rather than uniformly throughout the entire trajectory. The robot maintains full safety distance only in regions with high collision risk or uncertain obstacle behavior, while allowing closer approaches in regions with low risk or well-predicted obstacle paths. This partial application of safety measures optimizes the balance between collision avoidance and motion efficiency.