一、Towards Transferring Skills to Flexible Surgical Robots with Programming by Demonstration and Reinforcement Learning
Tendon-driven flexible serpentine manipulators(TSM,肌腱驱动蛇形机器人) 算法
提出一种基于期望最大化的强化学习算法PoWER,改进TSM机器人逆运动学模型提升追踪精度。与策略梯度法相比,PoWER不需要学习率参数(learning rate parameter),采样合并( sampling incorporated)便于利用经验,使快速易于实现。
使用高斯噪声模拟模型不确定性和干扰。PoWER搜索参数空间改进干扰模型并在线最大化奖励函数(maximize the reward function on line ),使收敛到标准逆运动学模型。
二、Web of Science 检索 Da Vinci surgical robot + motion control
1、Review of emerging surgical robotic technology
医疗机器人综述
2、A sequential windowed inverse reinforcement learning algorithm for robot tasks with delayed rewards
一种增强学习算法,以达芬奇为例
3、五页加