Online Mapping and Motion Planning Under Uncertainty for Safe Navigation in Unknown Environments

Abstract

Safe autonomous navigation is an essential and challenging problem for robots operating in highly unstructured or completely unknown environments. Under these conditions, not only robotic systems must deal with limited localization information but also their maneuverability is constrained by their dynamics and often suffers from uncertainty. In order to cope with these constraints, this article proposes an uncertainty-based framework for mapping and planning feasible motions online with probabilistic safety guarantees. The proposed approach deals with the motion, probabilistic safety, and online computation constraints by: 1) incrementally mapping the surroundings to build an uncertainty-aware representation of the environment and 2) iteratively (re)planning trajectories to goal that is kinodynamically feasible and probabilistically safe through a multilayered sampling-based planner in the belief space. In-depth empirical analyses illustrate some important properties of this approach, namely: 1) the multilayered planning strategy enables rapid exploration of the high-dimensional belief space while preserving asymptotic optimality and completeness guarantees and 2) the proposed routine for probabilistic collision check-

Uncertainty-driven Planner for Exploration and Navigation

“…c) Uncertainty estimation: To navigate in partially observed maps, uncertainty has been estimated across nodes in a path [4], [40], via the marginal probability of landmarks [5], and with the variance of model predictions across predicted maps [24], [41]. Furthermore, uncertainty-aware mapping has been shown to be effective in unknown and highly risky environments [42], [43]. In our work, we use uncertainty differently for exploration and point goal navigation.…”

Sensor‐driven autonomous underwater inspections: A receding‐horizon RRT‐based view planning solution for AUVs

“…In contrast, a brief description of the others is presented at the end of the section. Moreover, a novel line of work of Sodhi et al (2019), Ho et al (2018), and Pairet et al (2020) based on the OctoMap framework tries to provide a unique map representation useful both localization, under the Simultaneous Localization And Mapping (SLAM) paradigm, and planning. Nevertheless, this is not the core of this study.…”

“…The employed mapping strategy resorts to the occupancy grid mapping paradigm (Burgard et al, 2005). Born as a robust representation of the surrounding environment, occupancy grid mapping (Moravec & Elfes, 1985) has encountered several marine robotics applications in the context of collision checking and obstacle/collision avoidance (Youakim et al, 2020), mapping (Franchi, Bucci, et al, 2020; Teixeira et al, 2016), planning (J. D. Hernández et al, 2019; Vidal et al, 2020), and navigation with planning (Ho et al, 2018; Pairet et al, 2020; Sodhi et al, 2019).…”

Planning with SiMBA: Motion Planning under Uncertainty for Temporal Goals using Simplified Belief Guides

“…We present a modular framework to solve Problem 1. The framework is inspired by [5], [21], [23] and consists of three main layers: task planning layer, SiMBA guide layer, and belief search layer, as depicted in Fig. 1.…”

“…This is a relaxation of maximizing success probability, but it allows for fast and scalable motion planning with robustness guarantees. Most works focus on motion uncertainty [14], [15], [21], but recent works [17]- [20] introduce extensions to account for sensor uncertainty. While they provide efficient planning under uncertainty, they are limited to simple task of A-to-B planning.…”