连续动作强化学习及其在机器人中的应用研究

讨论了连续动作的强化学习系统实现及学习方法。首先介绍了连续动作的强化学习系统的组成原理，讨论了采用神经网络实现强化学习系统的方法，然后，介绍了强化学习机制在智能机器人避碰行为学习系统中的应用，并给出了系统的仿真结果。仿真结果表明机器人具有较好的避碰能力。

Research on Reinforcement Learning with Continuous Actions and its Application to Intelligent Robot

The realization method and learning algorithm for reinforcement learning system with continuous actions was presented. First, the compositive principle of reinforcement learning system with continuous actions was introduced and the implement of reinforcement learning system adopting neural network was discussed. Then, the application of reinforcement learning mechanism to obstacle avoidance behavior learning of intelligent robot was given. Finally, the simulation experiments were carried out for this system. The simulation results show that the robot has better avoidance collision ability .