In the deep reinforcement learning method, the Intrinsic Curiosity Model (ICM) guides the agent to obtain the opportunity to learn unknown strategies in the sparse reward environment, but the curiosity reward is a state difference value, which will make the agent pay too much attention to the exploration of new states. Then the problem of blind exploration arises, and an Intrinsic Curiosity Model algorithm based on Knowledge Distillation (KD-ICM) is proposed. Firstly, the method of knowledge distillation is introduced to make the agent acquire more abundant environmental information and strategy knowledge in a short time and accelerate the learning process. Secondly, by pre-training teachers' neural network model to guide the forward network to obtain a forward network model with higher accuracy and performance, reducing the blind exploration of agents. Two different simulation experiments are designed on the Unity simulation platform for comparison. The experiments show that in the complex simulation task environment, the average reward of KD-ICM algorithm is 136% higher than that of ICM, and the optimal action probability is 13.47% higher than that of ICM. The exploration performance of the agent can be improved while the exploration quality can be improved, and the feasibility of the algorithm is verified.

[2024-05-11] Accepted Paper /doc/2024.01.0014/v1