Bayesian Optimisation for Safe Navigation Under Localisation Uncertainty

Abstract

In outdoor environments, mobile robots are required to navigate through terrain with varying characteristics, some of which might significantly affect the integrity of the platform. Ideally, the robot should be able to identify areas that are safe for navigation based on its own percepts about the environment while avoiding damage to itself. Bayesian optimisation (BO) has been successfully applied to the task of learning a model of terrain traversability while guiding the robot through more traversable areas. An issue, however, is that localisation uncertainty can end up guiding the robot to unsafe areas and distort the model being learnt. In this paper, we address this problem and present a novel method that allows BO to consider localisation uncertainty by applying a Gaussian process model for uncertain inputs as a prior. We evaluate the proposed method in simulation and in experiments with a real robot navigating over rough terrain and compare it against standard BO methods.

A Bayesian Optimization Approach for Water Resources Monitoring Through an Autonomous Surface Vehicle: The Ypacarai Lake Case Study

“…A Bayesian Exploration-Exploitation approach for online planning was presented in [25], using a POMDP utility function. A recent work [26] shows the development of save navigation procedures for mobile robots using BO to predict high movement vibrations in a defined space.…”

“…Gaussian Processes (GPs) are generally used as a surrogate model in the Bayesian inference for both the prior and posterior model. Since they are based on multivariate Gaussian distributions, they can fit data with ease (some examples include [5], [21], [26]). GPs are defined with a mean function µ(x) and a covariance function.…”

A Bayesian Optimization Approach for Multi-Function Estimation for Environmental Monitoring Using an Autonomous Surface Vehicle: Ypacarai Lake Case Study

“…As the AF depends directly on the surrogate model, first, this model should be appropriately defined. Many authors [17,[25][26][27] use Gaussian processes as probabilistic surrogate models since they present important advantages for modeling unknown functions.…”

Informative Path Planning for Active Field Mapping under Localization Uncertainty

“…GPs are a popular non-parameteric Bayesian technique for modeling spatio-temporal phenomena [14]. They have been applied in various active sensing scenarios [1,3,5,16] to gather data based on correlations and uncertainty in continuous maps. However, most of these works assume that the training data for prediction is inherently noise-free, which may lead to inaccuracies if measurements are incorporated at wrong locations and mislead predictive planning algorithms.…”

“…Rapid technological advancements are inciting the use of autonomous mobile robots for exploration and data acquisition. In many marine [1,2], terrestrial [3,4], and airborne [5,6] applications, these systems have the ability to bridge the spatiotemporal divides limiting traditional measurement methods in a safer and more cost-effective manner [7]. However, to fully exploit their potential, algorithms are required for planning efficient informative paths in complex environments under platform-specific constraints.…”