改进型BINN算法应用在无人船优先区域覆盖路径规划的研究

Research on Improved BINN Algorithm for Coverage of Prioritized Area in Path Planning of Unmanned Surface Vessel

对于无人船在复杂海域执行的全区域覆盖任务,用原生物激励神经网络算法规划的路径欠佳,为此提出以下改进措施:为解决两个方向上的活性值相同而导致下一位置决策困难的问题,提出新的最优点决策公式;为解决障碍物毗邻区和海图边界区域活性值较低而导致无法完全覆盖的问题,重新定义生物激励神经网络模型中的输入激励。同时,通过在神经元输入中加入高激励选项,有效地实现了对区域的优先覆盖。仿真结果表明,提出的改进型算法无论在静态还是在动态海况下,都有很好的适应性,可实现对任务海域的100%覆盖,且能有效规避障碍物和快速脱离死区。

With respect to the non-ideal planning path of the original biological excitation neural network algorithm,this paper intends to solve the problem in decision making for the next position in the case of same activity values in two directions,and a new optimal decision formula is proposed.With respect to the incompleteness of coverage due to low activity values in the adjacent area of an obstacle and in the boundary area,excitation input of the neural network model is redefined.Simulation results show that the improved algorithm has good adaptability in both static and dynamic sea conditions.The 100%coverage of the mission sea area can be achieved,and obstacles can be avoided effectively and quickly break away from trapped points.Full coverage path planning under the premise of preferential coverage of a high-valued area can be realized.