水下机器人最优逃生线路规划研究与仿真  

Research and Simulation on Optimal Escape Route Planning of Underwater Robot

作　　者：刘莹[1] 祝毅鸣[1] 

机构地区：[1]郑州大学西亚斯国际学院,河南郑州451150

出　　处：《计算机仿真》2015年第3期352-355,411,共5页Computer Simulation

摘　　要：在水下机器人的逃生路线规划设计中,由于深海条件十分复杂,水下机器人在遇到各种险情逃生路线规划存在避障难题。传统的水下机器人逃生路线规划算法因受到海水连续波动问题,导致水下机器人速度和位置出现较大的扰动,摆脱路径规划结果存在较大偏差。为保障水下机器人作业安全,提出一种基于神经优化网络及遗传算法的水下机器人视觉最优逃生的线路规划,把机器人视觉仪器采集复杂障碍特征,归一化到视觉信息,融入规划模型中进行最佳路径的选择,将机器人摆脱复杂障碍以及最短路径的要求融合成一个适应度函数,通过遗传算法搜索获取最佳机器人逃生线路。仿真结果说明,神经网络优化遗传算法对于危险复杂海下情况,水下机器人最优逃生线路规划长度以及效率都优于传统模型。In order to guarantee the operation safety of underwater robot,an optimal escape route planning of underwater robot vision based on neural optimization network and genetic algorithm is proposed. The complex obstacle characteristics are collected by robot visual instrument to normalize to the visual information,and the planning model is integrated to make the optimal path selection. The requirements of the robot getting rid of complex obstacle and the shortest path are fused into a fitness function. The genetic algorithm is applied to search and obtain the best robot escape routes. The results of simulation show that the length and efficiency of the optimal escape route planning of underwater robot by using the neural network optimization and genetic algorithm are superior to the traditional model under the situation of dangerous and complex under sea.

关 键 词：水下机器人 逃生路线 神经优化算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]