Probabilistic Consensus of Multi-agent System under Joint Control of SMC and Minimum Entropy Compensation

Abstract

Due to the stochastic of multi-agent systems, it is difficult to achieve strict consensus. In this paper, consensus in the sense of probability is achieved by reducing the output error entropy of multi-agent system. Sliding mode controller is the core to keep the system stability and probability density function(PDF) compensator is used to reduce the chattering effect of sliding mode and compensate the random part of the system. Radial basis function neural network combined with the minimum entropy criterion is used to model the PDF compensator, and the output error entropy of the system is minimized through the training of weights, so as to optimize the control effect. Finally, the simulation results verify the effectiveness of the method.