Research/ SPACE AI & AUTONOMY
Terrain Classification Enhanced with Uncertainty for Space Exploration Robots from Proprioceptive Data
Mariela De Lucas Álvarez, Jichen Guo, Raul Domínguez, Matias Valdenegro-Toro
Learning-based MethodsarXiv: cs.LG · Perception & State EstimationOpenAlex topic: Robotics and Sensor-Based Localization · Surface Mobility & TerramechanicsOpenAlex topic: Soil Mechanics and Vehicle Dynamics
arXivcs.RO · cs.LGOpenAlexReinforcement Learning in Robotics · Robotics and Sensor-Based Localization · Soil Mechanics and Vehicle Dynamics
Terrain Classification is an essential task in space exploration, where unpredictable environments are difficult to observe using only exteroceptive sensors such as vision. Implementing Neural Network classifiers can have high performance but can be deemed untrustworthy as they lack transparency, which makes them unreliable for taking high-stakes decisions during mission planning. We address this by proposing Neural Networks with Uncertainty Quantification in Terrain Classification. We enable our Neural Networks with Monte Carlo Dropout, DropConnect, and Flipout in time series-capable architectures using only proprioceptive data as input. We use Bayesian Optimization with Hyperband for efficient hyperparameter optimization to find optimal models for trustworthy terrain classification.
Comments: 6 pages, 4 figures. LatinX in AI Workshop @ ICML 2023 Camera Ready
Journal ref: International Conference on Machine Learning Conference: LatinX in AI (LXAI) Research Workshop 2023