基于粗糙集与遗传算法的采摘机器人路径规划

为了提高采集机器人路径规划速度和自主导航的智能化水平,提出了一种基于粗糙集和遗传算法的路径规划方法,从而有效地提高了路径规划的速度和精度。采摘机器人根据实际果实采摘环境,利用图像分割技术,对果实目标进行识别,在二维栅格地图环境下,制定出决策表,并使用粗糙集对决策表进行约简,得到最小决策表,将其作为遗传算法初试种群,进行遗传交叉和复制操作,优化路径规划算法。为了验证采摘机器人算法性能的可靠性,对采摘机器人的性能进行了测试,包括果实图像的识别和机器人路径规划能力。通过测试发现:采摘机器人可有效地分割提取出成熟果实,并可完成多目标任务。对粗糙集和遗传算法的性能进行了测试,结果发现:使用粗糙集可以大大降低所需训练种群的数目,减少平均迭代次数;增加障碍物的复杂程度后,使用粗糙集遗传算法可以明显地提高路径规划的速度,从而提高了机器人采摘作业的效率。

Path Planning of Picking Robot Based on Rough Set and Genetic Algorithm

In order to improve the speed of acquisition robot path planning , improve the intelligent level of the robot au-tonomous navigation ,it proposed a path planning method based on rough set and genetic algorithm , thus it can effectively improve the speed and precision of path planning .Based on the actual fruit picking environment , by using the technology of image segmentation robot on fruit target identification , in a 2 D grid map of the environment , it can make a decision ta-ble ,and the use of rough set attribute reduction of decision table is ,to get the minimal decision table .As the initial popu-lation of genetic algorithm ,it introduced genetic crossover and copy operation , optimal path planning algorithm .In order to verify the picking robot algorithm performance reliability , the picking robot performance were tested , which test items include fruit image recognition and robot path planning ability .It was found by testing that picking robot can effectively segment extracted fruit ripening , and it can complete the objectives and tasks .The performance of rough set and genetic algorithm was tested by using the rough set ,which can greatly reduce the required number of training population .To re-duce the average number of iterations , increase the complexity of the obstacle .It can significantly improve the path plan-ning speed , thereby improve the efficiency of picking robot by using rough set genetic algorithm .