Efficient and Robust RFS-SLAM with Probabilistic Extended Feature Models

Localization within an environment is often critical to mission success, though it typically requires a priori information, such as an environmental feature map, that can be expensive and time-consuming to obtain. By minimizing the dependency on a priori information, budgets and timelines can be reduced while, in tandem, expanding the capabilities of exploration systems. Recent developments in random finite set (RFS)-based simultaneous localization and mapping (SLAM) show promise for autonomous navigation and exploration. Creating an RFS-based SLAM framework for space-based navigation will minimize the dependence upon a priori information and external communications, while simultaneously producing an architecture that is robust to common issues, like erroneous measurements and mismatches in the expected and observed features. Unfortunately, existing feature modeling approaches utilize simplistic techniques that often assume features to be point-sources with non-representative probabilistic models. These modeling decisions only partially reflect the underlying, physics-based processes and severely limit the information processed. Developing improved modeling techniques that incorporate probabilistic and extended feature considerations will allow autonomous systems to be more thorough in representing environmental features, producing richer information for navigation, exploration, and interaction. RFS-based SLAM approaches, however, are not traditional for navigation and thus lack much of the robustness and efficiency that is paramount to successful navigation. By extending the architectures to include factorized forms and consider parameters while also examining methods to reduce the burden of a multitarget update, robustness and efficiency can be addressed. The proposed research will develop an efficient and robust state-of-the-art RFS-based SLAM framework for autonomous navigation and extend the capabilities of current feature modeling techniques by introducing novel physics-based models for features of finite extent. Carrying out the proposed research will lead to a navigation framework that is applicable to all manner of space platforms will be realized.

The proposed research will develop an efficient and robust state-of-the-art RFS-based SLAM framework for autonomous navigation and extend the capabilities of current feature modeling techniques by introducing novel physics-based models for features of finite extent. Carrying out the proposed research will lead to a navigation framework that is applicable to all manner of space platforms will be realized.

This is early/mid-stage (TRL 2) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.