The improved Q-Learning algorithm based on pheromone mechanism for swarm robot system

The reinforcement learning of the robot learning have general applicability in path planning, motion control and other aspects of mobile robot, which not only converges of reinforcement learning but also attributes to the simple implementation of the reinforcement learning, the typical reinforcement learning method is Q-Learning. Some improvements of the shortcomings of the Q-Learning is proposed by using the pheromone mechanism of the ant colony algorithm to solve the information sharing problem in the reinforcement learning system. Finally, the improved Q-Learning algorithm is simulated in the platform of Player/Stage. The results are compared with Q-Learning algorithm and PSO algorithm, which prove that the improved Q-Learning has high efficiency in the path planning of swarm robotics.