Litcius/Paper detail

Data-Based Output Synchronization of Multi-Agent Systems With Actuator Faults

Yingying Liu, Zhanshan Wang, Yuan Wang

2022IEEE Transactions on Neural Networks and Learning Systems14 citationsDOI

Abstract

In this brief, the output synchronization of multi-agent systems (MAS) with actuator faults is studied. To detect the faults, a backward input-driven fault detection mechanism (BIFDM) is presented for MAS. Different from previous works, the system operation can be monitored without system model by the proposed BIFDM. Then to tolerate the faults, a novel fault-tolerant controller (FTC) based on reinforcement learning (RL) and backward information (BI) is proposed. Particularly, by the combination of BI, the design of additional parameters for faults is avoided. Furthermore, the proposed FTC overcomes the shortcoming that the previous FTCs cannot be applied to heterogeneous MAS. Finally, two simulation examples are given to verify the effectiveness of the proposed methods.

Topics & Concepts

ActuatorComputer scienceSynchronization (alternating current)Control theory (sociology)Reinforcement learningFault (geology)Controller (irrigation)Fault detection and isolationFault toleranceControl engineeringDistributed computingEngineeringControl (management)Artificial intelligenceChannel (broadcasting)Computer networkBiologySeismologyGeologyAgronomyDistributed Control Multi-Agent SystemsNonlinear Dynamics and Pattern FormationNeural Networks Stability and Synchronization
Data-Based Output Synchronization of Multi-Agent Systems With Actuator Faults | Litcius