Quantized consensus with finite data rate under directed topologies

This paper investigates the consensus problems in directed networks under communication constraints. Each agent has a real-valued state but can only exchange finitely many bits information with its neighbors at each time step. Based on dynamic difference encoding and decoding, a distributed algorithm is proposed to achieve quantized consensus asymptotically with as few as only one bit information exchange at each time step. The upper bound of asymptotic convergence rate and the group decision value are given. Furthermore, simulations are provided that demonstrate the effectiveness of the theoretical results.