Motion Generation Around Obstacles: A Multidimensional Sampling-Based Planner

ZW Xie and ZP Wang and BS Cao and Y Liu and WD Sun and GH Xie and YM Ji and BY Ma, IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, 72, 9314-9322 (2025).

DOI: 10.1109/TIE.2025.3539387

The quintic polynomial trajectory of the space manipulators holds significant value for the transmitting of ground-based data. In practice, it is frequently employed for large-scale transfer motions. However, this approach leads to an uncontrollable end-effector trajectory, which in turn restricts the applicability of various trajectory planning techniques reliant on end-effector trajectories. Sampling-based motion planner can be used to plan polynomial joint trajectories and optimize specific objectives, such as obstacle avoidance. Nevertheless, existing approaches either fix the sample dimensionality or neglect dimensional optimization. This article presents an enhancement based on genetic algorithms, introducing coding and crossover mechanisms for samples of varying dimensions. The trajectory planner performs search and optimization on a multidimensional sample set, thereby expanding the search domain and overcoming dimensional constraints. The mutual inspiration among samples of different dimensions enhances the search capabilities of the algorithm, yielding superior results. This work focuses on an actual engineering problem, and through simulation and validation experiments, it demonstrates that the proposed trajectory planner can generate a safe and coordinated quintic polynomial joint trajectory for the manipulator.

Return to Publications page