Mabel: Extending Human Interaction and Robot Rescue Designs

AAAI Conferences

Mabel (the Mobile Table) is a robotic system that can perform waypoint navigation, speech generation, speech recognition, natural language understanding, face finding, face following, nametag reading, and localization. Mabel can interact intelligently to give information about the conference to patrons. Major additions to this year's design are Monte Carlo Localization, Filter-Cascade techniques for vision applications, and an improved robot search and rescue system using a 3D OpenGL mapping system. Mabel was the winner of the 2003 robot host event and tied for third place in the robot search and rescue event at IJCAI 2003 in Acapulco, Mexico.


Position Estimation for Mobile Robots in Dynamic Environments

AAAI Conferences

For mobile robots to be successful, they have to navigate safely in populated and dynamic environments. While recent research has led to a variety of localization methods that can track robots well in static environments, we still lack methods that can robustly localize mobile robots in dynamic environments, in which people block the robot's sensors for extensive periods of time or the position of furniture may change. This paper proposes extensions to Markov localization algorithms enabling them to localize mobile robots even in densely populated environments. Two different filters for determining the "believability" of sensor readings are employed. These filters are designed to detect sensor readings that are corrupted by humans or unexpected changes in the environment. The technique was recently implemented and applied as part of an installation, in which a mobile robot gave interactive tours to visitors of the "Deutsches Museum Bonn." Extensive empirical tests involving datasets recorded during peak traffic hours in the museum demonstrate that this approach is able to accurately estimate the robot's position in more than 98% of the cases even in such highly dynamic environments.


Teaching Localization in Probabilistic Robotics

AAAI Conferences

In the field of probabilistic robotics, a central problem is to determine a robot's state given knowledge of a time series of control commands and sensor readings. The effects of control commands and the behavior of sensor devices are both modeled probabilistically. A variety of methods are available for deriving the robot's belief state, which is a probabilistic representation of the robot's true state (which cannot be directly known). This paper presents a series of five assignments to teach this material at the advanced undergraduate/graduate level. The theoretical aspect of the work is reinforced by practical implementation exercises using ROS (Robot Operating System), and the Bilibot, an educational robot platform.


Monte Carlo Localization: Efficient Position Estimation for Mobile Robots Dieter Fox, Wolfram Burgard, Frank Dellaert, Sebastian Thrun

AAAI Conferences

This paper presents a new algorithm for mobile robot localization, called Monte Carlo Localization (MCL). MCL is a version of Markov localization, a family of probabilistic approaches that have recently been applied with great practical success. However, previous approaches were either computationally cumbersome (such as grid-based approaches that represent the state space by high-resolution 3D grids), or had to resort to extremely coarse-grained resolutions. Our approach is computationally efficient while retaining the ability to represent (almost) arbitrary distributions. MCL applies sampling-based methods for approximating probability distributions, in a way that places computation "where needed." The number of samples is adapted online, thereby invoking large sample sets only when necessary. Empirical results illustrate that MCL yields improved accuracy while requiring an order of magnitude less computation when compared to previous approaches. It is also much easier to implement.


Robust Global Localization Using Clustered Particle Filtering

AAAI Conferences

Global mobile robot localiz ation is the problem of determining a robot's pose in an environment, using sensor data, when the starting position is unknown. A family of probabilistic algorithms known as Monte Carlo Localization (MCL) is currently among the most popular methods for solving this problem. MCL algorithms represent a robot's belief by a set of weighted samples, which approximate the posterior probability of where the robot is located by using a Bayesian formulation of th e localization problem. This article presents an extension to the MCL algorithm, which addresses its problems when localizing in highly symmetrical environments; a situation where MCL is often unable to correctly track equally probable poses for the robot. The problem arises from the fact that sample sets in MCL often become impoverished, when samples are generated according to their posterior likelihood. Our approach incorporates the idea of clusters of samples and modifies the proposal distribution considering the probability mass of those cluste rs. Experimental results are presented that show that this new extension to the MCL algorithm successfully localizes in symmetric environments where ordinary MCL often fails.