Researchers have developed an allocation-search-optimization (ASO) algorithm to address task allocation and path planning for truck-drone collaboration systems in complex, unstructured environments. The approach integrates task allocation, initial path search, safe corridor generation, and back-end trajectory optimization, tackling both obstacle avoidance and kinematic feasibility of heterogeneous agents.
Key takeaways
- The ASO algorithm combines a truck-drone enhanced conflict-based search (TD-ECBS) for initial path planning with a model predictive control (MPC) method for trajectory optimization, ensuring collision-free and kinematically feasible routes.
- In simulations with dense obstacles, the proposed method raises the critical obstacle density threshold by over 20% compared to existing algorithms (e.g., Boccia, Gonzalez-R) when planning success rates fall below 50%, while maintaining kinematic feasibility.
- Even without obstacle or kinematic constraints, the algorithm reduces planning time by 6.87% and 4.23% over Boccia's method for 10 and 20 agents, respectively, demonstrating efficiency in complex scenarios.
The work addresses a gap in current truck-drone coordination research, which often ignores obstacle handling and motion constraints. By explicitly incorporating these factors, the ASO algorithm improves mission success rates and execution efficiency, offering a practical solution for last-mile delivery and other applications in cluttered environments. The approach was validated in ROS simulations, confirming its effectiveness for real-world deployment.
Source: 《机器人》期刊 (robot.sia.cn) · Published 2026-05-12 · “复杂环境中车机协同系统的任务分配和路径规划”