Colloquium


DATE2022-05-12 16:10-17:00

PLACE數學系館 1F3174教室

SPEAKER曾國師 教授(國立中央大學數學系

TITLE3D Map Exploration using Topological Fourier Sparse Set
線上專題演講

ABSTRACT 線上專題演講
3D map exploration is one of key technologies in robotics. However, finding an optimal exploration path is a challenge due to unknown environments. This research proposed the Topological Fourier Sparse Set (TFSS) algorithm to enable an unmanned aerial vehicle (UAV) to explore 3D environments with theoretical guarantees. The algorithm combines the Rips complex with Fourier sparse set representation to take the advantages of topological and submodular approaches. More specifically, the Rips complex is used for expanding the exploration subgoals, while the Fourier sparse set encodes a learned representation of the subgoal selection problem in the form of a submodular optimization problem. Since the objective function of spatial exploration is reformulated as a maximizing submodular function with path constraints, greedy algorithms can achieve 0.5(1 - e^(-1)) of the optimum. Experiments conducted with this algorithm demonstrates that the TFSS explores unknown environments 25% to 127% more than the NBV algorithm does. The TFSS exploration performance is close to the SFSS but it is 50 times faster than the SFSS.
會議鏈結: https://nckucc.webex.com/nckucc/j.php?MTID=m6cb97ca14494cdcf291f71f0b8a0fa7b
會議號: 2511 759 8689
會議密碼: 238172