Closing the Loop Inside Neural Networks: Causality-Guided Layer Adaptation for Fault Recovery Control
Year: 2025
Type: article
Abstract: This paper studies the problem of real-time fault recovery control for nonlinear control-affine systems subject to actuator loss of effectiveness faults and external disturbances. We develop a two-stage framework that combines causal inference with selective online adaptation to achieve an effective learning-based recovery control method. In the offline phase, we introduce a causal layer attribution technique based on the average causal effect (ACE) to evaluate the relative importance of each layer in a pretrained deep neural network (DNN) controller compensating for faults. This provides a principled approach to select the most causally influential layer for fault recovery control in the sense of ACE, and goes beyond the widely used last-layer adaptation approach. In the online phase, we deploy a Lyapunov-based gradient update to adapt only the ACE-selected layer to circumvent the need for full-network or last-layer only updates. The proposed adaptive controller guarantees uniform ultimate boundedness (UUB) with exponential convergence of the closed-loop system in the presence of actuator faults and external disturbances. Compared to conventional adaptive DNN controllers with full-network adaptation, our methodology has a reduced computational overhead in the online phase. To demonstrate the effectiveness of our proposed methodology, a case study is provided on a 3-axis attitude control system of a spacecraft with four reaction wheels.
Source: IEEE Control Systems Letters
Authors: Mahdi Taheri, Soon-Jo Chung, Fred Y. Hadaegh
Institution: California Institute of Technology
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7123335396