Autonomous Robot using Q-Learning
In order to make intelligent machines, it is important to understand that cognitive development of human. In which, autonomous and cognitive decision making with environment interaction plays a key role. This paper proposes Q Learning algorithm based on an Online Sequential Extreme Learning Machine (OS-ELM). This system is based on reinforcement learning to find the exact path and avoid the collision in the indoor environment by path planning functions. The mobile robot can traverse the room and autonomously search the target in an absolutely strange environment. Back propagation is traditional, Instead of using back propagation technique, ELM neural network updates the output weight in real time and improve the speed and efficiency of Q value approximation. In which the redundancy and stacking of the samples are reduced and greatly improve the efficiency of Q-network.
Keywords - Markov Decision Process (MDP), Extreme Learning, Machine(ELM), Online Sequential Extreme Learning Machine(OSELM), Reinforcement Learning(RL).