Confidence-rich grid mapping

Abstract

Representing the environment is a fundamental task in enabling robots to act autonomously in unknown environments. In this work, we present confidence-rich mapping (CRM), a new algorithm for spatial grid-based mapping of the 3D environment. CRM augments the occupancy level at each voxel by its confidence value. By explicitly storing and evolving confidence values using the CRM filter, CRM extends traditional grid mapping in three ways: first, it partially maintains the probabilistic dependence among voxels; second, it relaxes the need for hand-engineering an inverse sensor model and proposes the concept of sensor cause model that can be derived in a principled manner from the forward sensor model; third, and most importantly, it provides consistent confidence values over the occupancy estimation that can be reliably used in collision risk evaluation and motion planning. CRM runs online and enables mapping environments where voxels might be partially occupied. We demonstrate the performance of the method on various datasets and environments in simulation and on physical systems. We show in real-world experiments that, in addition to achieving maps that are more accurate than traditional methods, the proposed filtering scheme demonstrates a much higher level of consistency between its error and the reported confidence, hence, enabling a more reliable collision risk evaluation for motion planning.

A Novel Occupancy Mapping Framework for Risk-Aware Path Planning in Unstructured Environments

“…These methods still need to tune parameters that have great consequences on the quality of the resulting maps. An alternative algorithm for occupancy grids was presented in Agha-mohammadi et al [13] by storing richer data in a map, taking into account the estimation of the variance for each cell. Under these considerations, our framework and the Bayesian occupancy grid are very alike, as the environment is tessellated into cells.…”

“…Under these considerations, our framework and the Bayesian occupancy grid are very alike, as the environment is tessellated into cells. The Lambda Field also stores a confidence interval over each cell in the same fashion as Agha-mohammadi et al [13].…”

Autonomous Spot: Long-Range Autonomous Exploration of Extreme Environments with Legged Locomotion

“…a meanμ i and variance σ i in the case of a beta distribution). The "confidence" about µ i is captured in σ i , where fully unknown and fully known cells have the highest and lowest σ i values, respectively [25].…”

Perception‐aware autonomous mast motion planning for planetary exploration rovers

“…Similarly, in Strader et al (), cooperative maneuvers are performed to enable observability and uniquely solve for the relative pose of a pair of Unmanned Aerial Vehicle (UAVs). Additionally, active mapping (Agha‐mohammadi, Heiden, Hausman, & Sukhatme, ; Forster, Pizzoli, & Scaramuzza, ; Heiden, Hausman, & Sukhatme, ), which is often addressed along with active localization, aims to enhance the quality of the map by picking trajectories that maximize the information captured by the sensors. The topic of active perception has been investigated extensively for both range‐based sensors (Bachrach et al, ; Bourgault, Makarenko, Williams, Grocholsky, & Durrant‐Whyte, ; Feder, Leonard, & Smith, ; Vidal‐Calleja, Sanfeliu, & Andrade‐Cetto, ) and vision‐based sensors (Achtelik, Lynen, Weiss, Chli, & Siegwart, ; Costante, Delmerico, Werlberger, Valigi, & Scaramuzza, ; Davison & Murray, ; Inoue, Ono, Tamaki, & Adachi, ; Mostegel, Wendel, & Bischof, ; Rodrigues, Basiri, Aguiar, & Miraldo, ; Sadat, Chutskoff, Jungic, Wawerla, & Vaughan, ; Vidal‐Calleja et al, ).…”