Out of Distribution Detection via Domain-Informed Gaussian Process State Space Models

TL;DR

This paper introduces a domain-informed Gaussian process state-space model for improved out-of-distribution detection in robotic navigation, enhancing safety by reliably identifying unseen scenarios using an online runtime monitor.

Contribution

It proposes embedding domain knowledge into the kernel of GPSSMs and develops an online OoD detection method based on receding-horizon predictions, improving detection accuracy.

Findings

Informed kernel improves regression with smaller datasets.

The OoD monitor reliably detects unseen terrains in real robot navigation.

Numerical results show enhanced OoD detection performance.

Abstract

In order for robots to safely navigate in unseen scenarios using learning-based methods, it is important to accurately detect out-of-training-distribution (OoD) situations online. Recently, Gaussian process state-space models (GPSSMs) have proven useful to discriminate unexpected observations by comparing them against probabilistic predictions. However, the capability for the model to correctly distinguish between in- and out-of-training distribution observations hinges on the accuracy of these predictions, primarily affected by the class of functions the GPSSM kernel can represent. In this paper, we propose (i) a novel approach to embed existing domain knowledge in the kernel and (ii) an OoD online runtime monitor, based on receding-horizon predictions. Domain knowledge is provided in the form of a dataset, collected either in simulation or by using a nominal model. Numerical results show that the informed kernel yields better regression quality with smaller datasets, as compared to standard kernel choices. We demonstrate the effectiveness of the OoD monitor on a real quadruped navigating an indoor setting, which reliably classifies previously unseen terrains.