Mobile robots detect faults quickly using adaptive lightweight models
Continuous Online Fault Detection for Mobile Robots via Adaptive Edge Models
RoboticsMachine Learning
Summary
Mobile robots need to spot problems with their parts quickly and while running on small, limited computers inside them. The authors created a two-step system where a complex model teaches a simpler, faster model to recognize faults in robot movements. This simpler model can run on the robot in real time and learn new problems over time without forgetting older ones. The system also asks humans for help only when it encounters completely new issues, reducing the amount of work for operators.
What this means in practice
- •For robotics engineers: Run fast and continuously adapting fault detectors directly on mobile robots’ onboard computers to catch mechanical issues early and maintain operation.
- •For industrial maintenance teams: Use adaptive models that require minimal human intervention to detect novel machine faults and reduce downtime across various equipment.
Authors
Jordan Levy, Nicolas Verstaevel, Vincent Talon, Benoit Gaudou
Abstract
Mobile robots require robust, real-time fault detection capable of continuous adaptation on constrained edge hardware. While deep time-series models excel at unsupervised anomaly detection, their computational cost prohibits high-frequency onboard execution. This paper bridges this gap via a Teacher-Student distillation framework. An offline foundation model (TSPulse) generates pseudo-labels from unlabeled time series augmented with fault injections. A lightweight MiniRocket Student, adapted with a Recursive Least Squares estimator, approximates this complex decision boundary to execute real-time inference onboard. Evaluations on the TSB-AD benchmark and a physical mobile robot demonstrate the Student achieves a 4.30 ms CPU inference latency. During real-world domain shifts, online adaptation enables the Student to recover from unseen mechanical degradation, improving VUS-PR scores from 0.26 to 0.75 without catastrophic forgetting. Crucially, an uncertainty-guided active learning strategy minimizes operator cognitive load, requesting sparse interventions only when encountering novel fault distributions. These results validate the deployment of state-of-the-art anomaly detection on resource-constrained robotics through offline-to-online distillation.