论文部分内容阅读
智能机器人行为学习可以看成是从感知空间到动作空间中寻找一种映射,文中采用遗传算法来实现寻找一种最佳映射,使机器人具有较好的避碰能力.把机器人旋转角度和平移距离分别划分成7个等级和4个等级,每一个等级看成机器人的一种行为;把机器人发生碰撞之前所走的路径长度作适应函数.通过遗传算法中的选择、交叉和变异等操作使机器人学会避碰行为.在文中的最后给出了仿真的实验结果.
Intelligent robot behavior learning can be seen as looking for a mapping from perceptual space to motion space. In this paper, genetic algorithm is used to find the best mapping, which makes the robot better avoid collision. The robot rotation angle and translation distance are divided into seven levels and four levels, each level as a robot behavior; the robotic collision path length before going to adapt to function. Through the selection of genetic algorithms, crossover and mutation and other operations so that robots learn to avoid collision behavior. At the end of this paper, the simulation results are given.