Archived video

Abstract
Simultaneous localization and mapping (SLAM) is the problem of estimating the state of a moving agent with sensors on it while simultaneously reconstructing a map of its surrounding environment, which has been a popular research field due to its wide applications. As many state-of-the-art SLAM algorithms can already achieve high accuracy in both state estimation and mapping, improving the robustness of SLAM systems has become a research focus in both academia and the industry in recent years.

One common challenge to robustness is the ambiguity problem, which is the situation when more than one interpretation could be plausible for the same observations. Common sources of ambiguities in SLAM include insufficient information, conflicts among different sensor measurements, uncertain data association, loop closing based on appearance only, etc. However, most of the state-of-the-art SLAM systems only estimate a single solution without considering multiple highly likely hypotheses resulting from ambiguous measurements, which can fail easily when ambiguities occur.

Therefore, we introduce a novel multi-hypothesis back-end optimization solver called MH-iSAM2 to take ambiguities into account and output multi-hypothesis solutions when the ambiguities are temporarily unsolvable. Our novel solver allows nonlinear incremental updates in all hypotheses while avoiding redundant computations across different hypotheses, which results in better efficiency than computing each hypothesis individually. Then, we develop a passive ambiguity-aware planar-inertial SLAM (API-SLAM) system based on MH-iSAM2 to reconstruct dense 3D models of indoor environments in real-time, which provides an example of applying MH-iSAM2 in a multi-hypothesis SLAM (MH-SLAM) framework for better robustness. Finally, we propose an ambiguity-aware active SLAM framework to make use of the multi-hypothesis state and map estimates from the MH-SLAM system in decision making and path planning, which demonstrates a complete and interactive usage of the multi-hypothesis estimations in a real-world robotic system. The experimental results show that MH-iSAM2 can be applied properly to improve the robustness of both passive and active SLAM systems, especially for handling the ambiguities in real-world tasks.

More Information

Thesis Committee Members
Michael Kaess, Chair
Sebastian Scherer
George A. Kantor
Christian H. Debrunner, Lockhead Martin Corporation