Multi-objective Task Assignment and Autonomous Approach Research Based on Multiple Unmanned Vehicles

To explore the decision optimization strategy of multiple unmanned ground vehicles (UGVs) for multi-objective tasks in a ground-air cross-domain collaborative unmanned system, a new four-stage multi-unmanned vehicles hybrid dynamic-static task assignment and autonomous approach architecture is proposed. A heuristic A* search algorithm is used to complete movement cost estimation among target points in the pre-task assignment phase. The task is modeled as a multiple traveling salesmen problem in the static assignment phase and a centralized genetic iterative optimization assignment scheme is designed. Distributed contract network mechanism is used in the dynamic allocation phase for collaborative allocation among UGVs for additional target points. In the task execution phase, a distributed local path planning control strategy for unmanned vehicles is proposed for obstacle avoidance and collision avoidance in the autonomous approach process. Validation tests were conducted based on three UGVs, and the results demonstrated the feasibility and real-time performance of multi-target tasking and autonomous approach architectures.