Improved PBFT Algorithm Based on K-Means Clustering for Emergency Scenario Swarm Robotic Systems

In order to solve the problem of data security and sharing efficiency caused by the complex and uncertain environment faced by swarm robotic systems in emergency scenarios, this paper designs a data security consensus algorithm based on blockchain technology. Aiming at the problems of high communication overhead, long consensus delay and low throughput when the traditional Practical Byzantine Fault Tolerance (PBFT) algorithm is applied to this scenario, a grouped practical Byzantine Fault Tolerance consensus algorithm based on K-means clustering is proposed. Firstly, the K-means clustering algorithm is used to group robotic nodes according to the location distribution of robotics in the emergency scenario. Secondly, a reputation mechanism is designed to dynamically evaluate the behaviour of robotic nodes in each group during the consensus process. Each consensus node is divided into master and slave chains according to its reputation score. The consensus task is firstly decomposed, and slave chains participate in the consensus in parallel. Finally, the master chain completes the global consensus, so as to reduce the communication times between nodes. The experimental results show that compared with the traditional PBFT, SG-PBFT and P-PBFT consensus algorithm, the proposed consensus algorithm effectively improves the system throughput, reduces the delay, reduces the communication overhead, and has a higher success rate of consensus.