The dynamics of extraterrestrial rovers is dependent on the terrain. The high-level terrain classification used in most current rovers is often not enough to ensure safe path selection, as the experience with NASA’s Curiosity and Spirit shows.
A recent paper suggests a model of the terrain parameters that govern wheel-terrain interaction. Knowing the terrain may help to predict whether the neighboring regions are traversable, plan the safest route, and prevent damage.
A linear model, which relates terrain parameters (namely cohesion and internal friction angle) and rover dynamics is supplemented by a meta-learned neural network. The interpretability of the model is enhanced by the orthogonality of nominal and meta-learned features. The model is capable of rapid adaptation and provides low estimation errors (the largest error is less than 5%).
Rovers require knowledge of terrain to plan trajectories that maximize safety and efficiency. Terrain type classification relies on input from human operators or machine learning-based image classification algorithms. However, high level terrain classification is typically not sufficient to prevent incidents such as rovers becoming unexpectedly stuck in a sand trap; in these situations, online rover-terrain interaction data can be leveraged to accurately predict future dynamics and prevent further damage to the rover. This paper presents a meta-learning-based approach to adapt probabilistic predictions of rover dynamics by augmenting a nominal model affine in parameters with a Bayesian regression algorithm (P-ALPaCA). A regularization scheme is introduced to encourage orthogonality of nominal and learned features, leading to interpretable probabilistic estimates of terrain parameters in varying terrain conditions.