Exploration and map-building under uncertainty with multiple heterogeneous robots

Exploration and map-building under uncertainty with multiple heterogeneous robots,10.1109/ICRA.2011.5979686,Lourdes Munoz-Gomez,Moises Alencastre-Mira

Exploration and map-building under uncertainty with multiple heterogeneous robots  
BibTex | RIS | RefWorks Download
In this paper, we present a multi-robot exploration strategy for map-building. We consider a team of robots with different sensing and motion capabilities. We combine geometric and probabilistic reasoning to propose a solution to our problem. We formalize the proposed solution using dynamic programming in states with imperfect information. We apply the dynamic programming technique in a reduced search space that allows us to incrementally explore the environment. We propose realistic sensor models and provide a method to compute the probability of the next sensor reading given the current state of the team of robots based on a Bayesian approach. I. INTRODUCTION Automatic environment exploration and map building is an important problem in mobile robotics. Autonomous robots must posses the ability to explore their environments, build representations of those environments (maps), and then use those representations to navigate effectively. Maps built upon exploration can be used later by the robot to perform other tasks such as object finding. A strategy for exploring an unknown environment and building an environment representation with a mobile robot can be performed as follows: (i) the robot builds a local map with the sensor readings (ii) the robot moves to an intermedi- ary goal, which is defined based on suitable properties (iii) a global map is updated merging the information between the current global map and the new local map. In the last two decades several approaches have been proposed for map-building, for instance (5), (8), (10), (19) just to name a few. Most previous research has focused on developing techniques to extract relevant information from raw data and to integrate the collected data into a single model. However, a robot motion strategy to explore the environment has been less studied. In this work, we deal mainly with this latter problem. In this paper an exploration strategy is proposed. Our exploration strategy considers sensing and motions capa- bilities of each robot. Our algorithm outputs the sensing configurations to be visited. The sensing configurations are associated to the borders between the known and unknown space. Our method assigns a robot of the team for visiting a selected sensing configuration, according to its capabilities and without considering predefined roles.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.