Rao-Blackwellized Particle Filters multi robot SLAM with unknown initial correspondences and limited communication

Abstract

Multi robot systems are envisioned to play an important role in many robotic applications. A main prerequisite for a team deployed in a wide unknown area is the capability of autonomously navigate, exploiting the information acquired through the on-line estimation of both robot poses and surrounding environment model, according to Simultaneous Localization And Mapping (SLAM) framework. As team coordination is improved, distributed techniques for filtering are required in order to enhance autonomous exploration and large scale SLAM increasing both efficiency and robustness of operation. Although Rao-Blackwellized Particle Filters (RBPF) have been demonstrated to be an effective solution to the problem of single robot SLAM, few extensions to teams of robots exist, and these approaches are characterized by strict assumptions on both communication bandwidth and prior knowledge on relative poses of the teammates. In the present paper we address the problem of multi robot SLAM in the case of limited communication and unknown relative initial poses. Starting from the well established single robot RBPF-SLAM, we propose a simple technique which jointly estimates SLAM posterior of the robots by fusing the prioceptive and the eteroceptive information acquired by each teammate. The approach intrinsically reduces the amount of data to be exchanged among the robots, while taking into account the uncertainty in relative pose measurements. Moreover it can be naturally extended to different communication technologies (bluetooth, RFId, wifi, etc.) regardless their sensing range. The proposed approach is validated through experimental test.

Distributed Multi - Robot Rao - Blac k wellized Particle Filt er ing Simultaneous Localization and Mapping with Consensus Calculation of Particle Weight and Posterior Parameters

“…We refer to [5] in order to calculate global frame . Based on global frame calculation in [5], robot calculates local frame reference ′ of robot ′ and applies roto-translation using formulation that is written as equation (18).…”

“…Based on global frame calculation in [5], robot calculates local frame reference ′ of robot ′ and applies roto-translation using formulation that is written as equation (18). = ⊕ ′ ⊖ ′ (18) Each robot processes the global frame locally; thus, in every consensus iteration the robots try to find the convergence of the global frame ( ) which can be expressed as equation (19).…”

“…Meanwhile, mapping is a computation problem of integrating environmental information obtained from sensor installed on the robot into map representation. There are several map representations that can be built using SLAM; those are: (1) grid map, a map that represents the environment into occupied grid [1,2,3,4,5], (2) topological map, a map that represents the environment in a simplified graph, which only provides vital information such as robot's position and geometric structure of environment remains [1,6,7], (3) line map, a map that represents the environment using line and curve model [1,8,9,10,11], and (4) feature based map, a map that represents the environment into a set of features called landmark [1,12,13,14].…”

“…This situation may occur due to a considerable difference in initial position of each robot. Multi-robot SLAM with unknown initial position by applying distributed Rao-Blackwellized Particle Filter SLAM is then proposed in [5]. Thisis no connection with other robots or when relative position data is unavailable within the network.…”

Coordinated multi-robot exploration: Out of the box packages for ROS

“…Distributed simultaneous localization and mapping (SLAM) has been discussed in the literature [8], [9], [10], though only few implementations are available on ROS. The cg mrslam package implements distributed SLAM using condensed graphs while the mapstitch package allows to compute the overlap between maps and combines them accordingly.…”