基于任务规划的多机器鱼协作策略
内容摘要：
为解决2D 仿真机器鱼比赛中多鱼对抗项目的任务分工、整体合作及实时动态调整问题，提出基于多机器鱼协作任务规划原理的协作策略。根据机器鱼自身完成任务的能力和效率，将给定的任务分解和分配，按照各比赛项目的特征分别采取不同的协作策略：对2vs2 采用群协作分配方式，5vs5 项目采用集中分配方式，并设计了三鱼协作队形以提高动作的有效性。仿真结果表明：该协作任务规划方法可增强多鱼的协作能力、动态调整能力和进球率，较好地满足了多机器鱼系统的控制要求。
关键词：
机器鱼；协作；任务规划；策略
中图分类号：TP242 文献标志码：A
Abstract: To solve the problems of task division, cooperation and real-time dynamic adjustment in the 2D simulation robot fish game, this paper proposes the strategy based on the principle of multiple robot cooperative mission planning. Decompose and assign the task according to the capacity and efficiency of each robot fish. In terms of the characteristics of competition, we take corresponding strategy: in 2 vs 2 competition, the strategy is assigning the tasks among the fishes in cooperation pattern. The strategy of 5 vs 5 competition is assigning the tasks among fishes in master-slave pattern. At the same time, we design collaborative formation of three fishes to improve the efficient. The result shows that these strategies can enhance fishes' cooperation ability, dynamic adjustment ability and the rate of scoring, meet the control equirements of multiple robot fishes system.
Keywords: robot fish; cooperation; mission planning; strategy
