摘要
针对机器人在家庭服务的应用过程中,由于环境复杂度较高而难以精确抓取目标物体的问题,文中提出一种改进的深度强化学习算法。采用Yolov3使机器人从多个物体中识别出目标物,并利用单目测距结果获取目标物体的相对位置,进而通过深度强化学习规划出机器人手臂抓取路径。结合真实机器人在预设场景中的运行结果得出,改进的深度强化学习算法在复杂环境中的目标抓取准确率最高可达到0.96,且具有较高的鲁棒性。
In the process of robot application in home service,it is difficult to accurately grasp the target object due to the high complexity of the environment.In this paper,an improved deep reinforcement learning algorithm is proposed.Yolov3 is used to recognize the target from multiple objects,and the relative position of the target object is obtained by monocular ranging results,and then the grasping path of the robot arm is planned by deep reinforcement learning.Combined with the running results of real robot in the preset scene,the improved deep reinforcement learning algorithm has the highest target grasping accuracy of 0.96 in complex environment,and has high robustness.
作者
刘天宇
陈晔
刘雪峰
LIU Tianyu;CHEN Ye;LIU Xuefeng(Electrical and Control Engineering,North University of China,Taiyuan 030051,China)
出处
《电子设计工程》
2021年第22期184-188,共5页
Electronic Design Engineering
