ProtoInnovations in collaboration with the University of Wisconsin-Madison proposes to architect, design, develop, and validate a high-fidelity modeling and sensor simulation software and perception algorithms for surface hazard detection in harsh lunar-like environments. The space robotics community currently lacks an end-to-end software suite that simulates the appearance and granular-terrain mechanics of lunar-like environments. As NASA prepares for progressively complex and longer future lunar surface missions, surface robot systems will require higher performance and autonomy capabilities to carry out mission-critical tasks. This includes performing reliably and traversing successfully through previously unexplored lunar terrain in harsh environments. To achieve this, perception algorithms will need to advance significantly to enable high-performance autonomy in dynamic lunar surface conditions.
For this project we aim to design and develop a software suite consisting of a simulation environment that replicates harsh lunar surface conditions, lunar-terrain mechanics, and exploration rovers with various sensing capabilities. The simulation platform will have interfaces similar to those present on current NASA prototype rovers. This will maximize infusion potential for the deployment of software on future missions utilizing various rover platforms and provide continuous development and monitoring of performance on their digital counterparts. In addition, we will leverage the simulation platform to research, develop and validate perception algorithms to enable object/hazard detection under harsh conditions which commonly occur on the lunar surface: low lighting, oblique lighting, long shadows, permanent shadows, irregular reflectance, and soft soil (lunar regolith), excessive slip and sinkage. Robust perception under these conditions will enable new autonomous capabilities for future surface mobility systems and enhance safe mission operations.
Cited as a need in subtopic T4.01, this project focuses on robust rover perception for the purpose of critical hazard detection under extreme conditions to prevent failure scenarios in future lunar surface missions. Additionally, the proposed framework contains some core elements of an end-to-end ground software. This provides a multitude of opportunities for the development of advanced autonomy capabilities pre-fight and could also be developed as an analysis and performance monitoring tool for in-flight use.
Advances made in the Chrono environment under this project will enable new simulation capabilities for all-terrain vehicles in defense and earthmoving applications. New perception algorithms for hazard detection in special environmental conditions such as low lighting, long shadows, etc., would be applicable to various field robotic applications in agriculture, defense, and mining.