Deep Long-Short Term Memory networks: Stability properties and Experimental validation

2023/04/06 by Fabio Bonassi, Alessio La Bella, Bonassi, Fabio +7
Engineering · Physics and Astronomy · #Fault Detection and Control Systems #Model Reduction and Neural Networks #Machine Fault Diagnosis Techniques

paper · pdf · doi:10.48550/arxiv.2304.02975

Abstract

The aim of this work is to investigate the use of Incrementally Input-to-State Stable (δISS) deep Long Short Term Memory networks (LSTMs) for the identification of nonlinear dynamical systems. We show that suitable sufficient conditions on the weights of the network can be leveraged to setup a training procedure able to learn provenly-δISS LSTM models from data. The proposed approach is tested on a real brake-by-wire apparatus to identify a model of the system from input-output experimentally collected data. Results show satisfactory modeling performances.

Related

• LSTM Neural Networks: Input to State Stability and Probabilistic Safety Verification
• Deep transfer learning for system identification using long short-term memory neural networks
• A critical look at deep neural network for dynamic system modeling
• Deep State Space Models for Nonlinear System Identification
• Formally Verified Neural Lyapunov Function for Incremental Input-to-State Stability of Unknown Systems