Total: 1
Autonomous agents require the capability to identify dynamic objects in their environment for safe planning and navigation. Incomplete and erroneous dynamic detections jeopardize the agent's ability to accomplish its task. Dynamic detection is a challenging problem due to the numerous sources of uncertainty inherent in the problem's inputs and the wide variety of applications, which often lead to use-case-tailored solutions. We propose a robust learning-free approach to segment moving objects in point cloud data. The foundation of the approach lies in modelling each voxel using a hidden Markov model (HMM), and probabilistically integrating beliefs into a map using an HMM filter. The proposed approach is tested on benchmark datasets and consistently performs better than or as well as state-of-the-art methods with strong generalized performance across sensor characteristics and environments. The approach is open-sourced at https://github.com/vb44/HMM-MOS.