Distributed Consensus-Based Online Monitoring of Robot Swarms With Temporal Logic Specifications

In this letter, we develop a distributed consensus-based online monitoring framework for robot swarm with fixed graph structure. Each agent can monitor whether the satisfies specifications given in form of Swarm Signal Temporal Logic (SwarmSTL) formulas. SwarmSTL formulas describe temporal properties swarm-level features represented by generalized moments (GMs), e.g., centroid and variance. To deal measurement noise, propose moment consensus algorithm (GMCA) Kalman filter (KF), allowing each to estimate GMs. Besides, prove convergence GMCA derive an upper bound error between agent's GMs actual This is derived be dependent on maximal allowed velocity but independent agents' exact motion. A set rules are proposed based estimation bound. As result, agents satisfaction over during execution. The applied supply transportation example, where efficacy KF also shown.