基于生物启发自组织神经网络的任务分配与路径规划

摘要

针对自治水下机器人(AUV)系统的任务分配与路径规划问题,提出一种生物启发自组织映射(GBSOM)神经网络任务分配与路径规划算法.根据AUV的水下工作环境建立二维生物启发神经网络(GBNN)模型,神经网络中每一个神经元的活性值与水下栅格地图中的位置单元一一对应;使用自组织神经网络(SOM)算法将水下目标分配给一组AUV并确定AUV访问目标点的顺序;在SOM任务分配的基础上根据神经网络中神经元的活性输出值分布情况自主规划AUV下一步的目标点.重复上述3步直至完成所有目标点的访问.最后,通过二维静态与动态环境下的仿真实验证明该算法的有效性.%For the task assignment and path planning of multiple autonomous underwater vehicle (AUV) system,a glasius bio-inspired self-organizing map (GBSOM)neural network algorithm was proposed based on two-dimension grid map.Firstly,a two-dimension glasius bio-inspired neural network (GBNN) model was established to represent two-dimension underwater working environment.The active value of each neuron in the neural network corresponds to the location unit in the grid map respectively.Secondly,a self-organizing map (SOM) was used to assign the targets to a set of AUVs and determine the order of AUV to access the target point.Thirdly,AUV next target point can be planned autonomously according to the distribution of the output value of the neuron in the neural network based on SOM task allocation.By repeating the above three steps,the access to all target points was completed.Simulation results verified the effectiveness of the proposed algorithm in two-dimension static and dynamic environment.