NN-ETM: Enabling safe neural network-based event-triggering mechanisms for consensus problems
In networked control applications, event-triggering mechanisms (ETMs) reduce the communication load while ensuring performance guarantees. However, the design of ETMs is becoming increasingly complex, particularly for decentralized multi-agent and consensus setups, where the condition used to trigger communication might incorporate the agent's local information and the information received from neighbors. This typically results in ad-hoc solutions, which may only work for the consensus protocols under consideration. In this work, we aim to safely incorporate neural networks in the ETM to provide a general and flexible solution while ensuring guaranteed performance. To decouple the stability analysis of the consensus protocol from the abstraction of the neural network in the ETM, we first derive design criteria for the consensus and ETM pair, which allow independent analysis of each element under mild constraints. As a result, we propose NN-ETM, a novel ETM featuring a neural network, which provides an all-purpose solution to optimize communication in consensus problems while preserving the stability guarantees of the consensus protocol.
PDF Abstract