Recent Findings in Robotics Described by a Researcher from Henan University of S cience and Technology (Active Collision Avoidance for Robotic Arm Based on Artif icial Potential Field and Deep Reinforcement Learning)
Recent Findings in Robotics Described by a Researcher from Henan University of S cience and Technology (Active Collision Avoidance for Robotic Arm Based on Artif icial Potential Field and Deep Reinforcement Learning)
河南科技大学研究员介绍的机器人学最新发现(基于人工势场和深度强化学习的机器人手臂主动避碰)
扫码查看
点击上方二维码区域，可以放大扫码查看
摘要
由一名新闻记者-机器人与机器学习的工作人员新闻编辑每日新闻-机器人的新研究是一份新报告的主旨。根据NewsRx编辑在洛阳的新闻报道,研究表明:“为了解决人工势场(APF)主动避碰中经常遇到的局部最小ISUE问题,提出了一种将APF与Dee P强化学习(DRL)相结合的机器人手臂主动避碰算法。”新闻记者从河南科技大学的研究中得到一句话：“首先，为了提高DRL避碰问题的训练效率，通过调整障碍物的位置来增强事后经验回放(HER)，产生事后经验回放(HER-CA)。”基于双延迟深决定论策略梯度(TD3)和her-ca方法,建立了机器人手臂避碰行为网络模型,并基于人工势场建立了机器人手臂的全身避碰势场模型。将训练好的动作网络模型用于APF实时避碰规划的GUI de APF，并与HER CA进行了对比实验。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-New research on robotics is the subjec t of a new report. According to news reporting out of Luoyang, People's Republic of China, by NewsRx editors, research stated, "To address the local minimum iss ue commonly encountered in active collision avoidance using artificial potential field (APF), this paper presents a novel algorithm that integrates APF with dee p reinforcement learning (DRL) for robotic arms." The news journalists obtained a quote from the research from Henan University of Science and Technology: "Firstly, to improve the training efficiency of DRL for the collision avoidance problem, Hindsight Experience Replay (HER) was enhanced by adjusting the positions of obstacles, resulting in Hindsight Experience Repl ay for Collision Avoidance (HER-CA). Subsequently, A robotic arm collision avoid ance action network model was trained based on the Twin Delayed Deep Determinist ic Policy Gradient (TD3) and HER-CA methods. Further, a full-body collision avoi dance potential field model of the robotic arm was established based on the arti ficial potential field. Lastly, the trained action network model was used to gui de APF in real-time collision avoidance planning. Comparative experiments betwee n HER and HER-CA were conducted."
Key words
Henan University of Science and Technolo gy/Luoyang/People's Republic of China/Asia/Emerging Technologies/Machine Le arning/Reinforcement Learning/Robotics/Robots
NETL
NSTL