Resource Mapping with a Mobile Exploration Robot Using Spectral Mixture Ergodic Search
Margaret Hansen, Ananya Rao, Abigail Breitfeld, David Wettergreen
AI summary
Problem
Standard ergodic search relies on static, predetermined frequencies and ignores spatial correlations, limiting its efficiency and adaptability for dynamic resource mapping tasks.
Approach
The method pairs Gaussian process regression with a spectral mixture kernel to extract spatial frequencies from the map, which are then fed into an ergodic search planner to iteratively update trajectories as new observations are collected.
Key results
- Introduced SM-ES, dynamically linking Gaussian process-learned frequencies to ergodic trajectory planning
- Demonstrated iterative map updates and trajectory re-planning across synthetic, lunar ice, and real mineral datasets
- Enabled per-dimension frequency adaptation, reducing planning complexity compared to standard ergodic search
- Validated improved exploration-exploitation balance and higher-fidelity resource maps with fewer samples
Why it matters
Provides a scalable, adaptive planning framework for mobile robots to efficiently prospect for critical resources in extreme, inaccessible environments like the lunar south pole and Mars.
Abstract
Resource mapping and prospecting has become the focus of a number of proposed planetary exploration missions, particularly to locate water ice at the lunar south pole. Mobile robots, which are employed for exploration tasks in environments that are inaccessible to humans, collect the information in such missions. In these scenarios, intelligent and adaptive trajectory planning algorithms increase the accuracy of the resulting resource map, along with the efficiency with which information is gathered. In this work, we use ergodic search to generate a mobile robot trajectory that balances exploration and exploitation, while simultaneously mapping the spatial distribution of a resource by using Gaussian process regression with a spectral mixture kernel. The spatial correlation structure learned via Gaussian process regression informs the ergodic search about regions of high information, as well as the frequency components that appear in the map distribution. We call this method spectral mixture ergodic search (SM-ES) and demonstrate how it learns a map and updates the trajectory accordingly on three datasets: synthetic maps, an ice favorability index map for the lunar south polar region, and real mineral data from Cuprite, Nevada.