AI强化学习算法在陶瓷雕刻机器设备中的应用研究

Application Research of AI Reinforcement Learning Algorithm in Ceramic Engraving Machine Equipment

为了快速适应陶瓷雕刻产品的个性化需求,研究分析人工智能(Artificial Intelligence,AI)强化学习算法在陶瓷雕刻机器设备中的应用。在陶瓷雕刻机器人铣削仿真模型的前提下,创建包括动作空间、环境状态的雕刻强化学习环境,同时介绍了深度确定性策略梯度(Deep Deterministic Policy Gradient,DDPG)算法单次走刀和全局走刀训练智能体。对比于常见的单次走训练智能体,随机偏差形成的错误走刀导致误差的取值范围为-1.876~2.265。智能体控制刀具不仅能按照给出的路径完成移动训练,同时能极大程度减少走刀错误的次数,进而能防止陷入局部最优走刀状态。在雕刻设计强化学习环境中,陶瓷雕刻机器人DDPG具有较强的学习能力,可将其应用于实际陶瓷工艺作品的生产。

In order to quickly adapt to the personalized needs of ceramic carving products,the application of Artificial Intelligence(AI) reinforcement learning algorithms in ceramic carving machines and equipment was studied and analyzed.Based on the milling simulation model of a ceramic engraving robot,a sculpting reinforcement learning environment is created that includes action space and environmental status.At the same time,the Deep Deterministic Policy Gradient(DDPG) algorithm is introduced for single tool walking and global tool walking training agents.Compared to common single walk training agents,the value range of error caused by wrong walk caused by random deviation is-1.876-2.265.The intelligent agent control tool can not only complete the movement training according to the given path,but also greatly avoid the number of tool walking errors,thereby preventing falling into the local optimal tool walking state.In the intensive learning environment for carving design,the ceramic carving robot DDPG has strong learning ability and can be applied to the production of actual ceramic craft works.