This paper analyzes the consistency of the classical extended Kalman filter (EKF) solution to the simultaneous localization and map building (SLAM) problem. Our results show that in large environments the map quickly becomes inconsistent due to linearization errors. We propose a new EKF-based SLAM algorithm, robocentric mapping, that greatly reduces linearization errors, improving map consistency. We also present results showing that large-scale mapping methods based on building local maps with a local uncertainty representation (Tardós et al., 2002) have better consistency than methods that work with global uncertainties.

Abstract-This article describes a rigorous and complete framework for the simultaneous localization and map building problem for mobile robots: the symmetries and perturbations map (SPmap), which is based on a general probabilistic representation of uncertain geometric information. We present a complete experiment with a LabMate TM mobile robot navigating in a human-made indoor environment and equipped with a rotating two-dimensional (2-D) laser rangefinder. Experiments validate the appropriateness of our approach and provide a real measurement of the precision of the algorithms.Index Terms-Correlations, probabilistic model, simultaneous localization and map building.

In this paper 1 we study the Extended Kalman Filter approach to simultaneous localization and mapping (EKF-SLAM), describing its known properties and limitations, and concentrate on the filter consistency issue. We show that linearization of the inherent nonlinearities of both the vehicle motion and the sensor models frequently drives the solution of the EKF-SLAM out of consistency, specially in those situations where uncertainty surpasses a certain threshold. We propose a mapping algorithm, Robocentric Map Joining, which improves consistency of the EKF-SLAM algorithm by limiting the level of uncertainty in the continuous evolution of the stochastic map: (1) by building a sequence of independent local maps, and (2) by using a robot centered representation of each local map. Simulations and a large-scale indoor/outdoor experiment validate the proposed approach.

We address the problem of online path planning for optimal sensing with a mobile robot. The objective of the robot is to learn the most about its pose and the environment given time constraints. We use a POMDP with a utility function that depends on the belief state to model the finite horizon planning problem. We replan as the robot progresses throughout the environment. The POMDP is high-dimensional, continuous, non-differentiable, nonlinear, non-Gaussian and must be solved in real-time. Most existing techniques for stochastic planning and reinforcement learning are therefore inapplicable. To solve this extremely complex problem, we propose a Bayesian optimization method that dynamically trades off exploration (minimizing uncertainty in unknown parts of the policy space) and exploitation (capitalizing on the current best solution). We demonstrate our approach with a visually-guide mobile robot. The solution proposed here is also applicable to other closely-R. Martinez-Cantin ( )

Abstract-This paper proposes a simulation-based active policy learning algorithm for finite-horizon, partially-observed sequential decision processes. The algorithm is tested in the domain of robot navigation and exploration under uncertainty, where the expected cost is a function of the belief state (filtering distribution). This filtering distribution is in turn nonlinear and subject to discontinuities, which arise because constraints in the robot motion and control models. As a result, the expected cost is non-differentiable and very expensive to simulate. The new algorithm overcomes the first difficulty and reduces the number of simulations as follows. First, it assumes that we have carried out previous evaluations of the expected cost for different corresponding policy parameters. Second, it fits a Gaussian process (GP) regression model to these values, so as to approximate the expected cost as a function of the policy parameters. Third, it uses the GP predicted mean and variance to construct a statistical measure that determines which policy parameters should be used in the next simulation. The process is iterated using the new parameters and the newly gathered expected cost observation. Since the objective is to find the policy parameters that minimize the expected cost, this active learning approach effectively trades-off between exploration (where the GP variance is large) and exploitation (where the GP mean is low). In our experiments, a robot uses the proposed method to plan an optimal path for accomplishing a set of tasks, while maximizing the information about its pose and map estimates. These estimates are obtained with a standard filter for SLAM. Upon gathering new observations, the robot updates the state estimates and is able to replan a new path in the spirit of openloop feedback control.