In this paper we propose an active control strategy for scanning laser sensors on autonomous vehicles traveling offroad at high speeds. As speed increases the amount of sensor information about the terrain decreases. We address the problem of sensor control in the context of this speed-coverage trade off. The algorithm and testing methodologies are described with results comparing our active sensing method to a passive sensing method.

The extended Kalman filter (EKF) has been the de facto approach to the Simultaneous Localization and Mapping (SLAM) problem for nearly fifteen years. However, the EKF has two serious deficiencies that prevent it from being applied to large, real-world environments: quadratic complexity and sensitivity to failures in data association. FastSLAM, an alternative approach based on the Rao-Blackwellized Particle Filter, has been shown to scale logarithmically with the number of landmarks in the map. This efficiency enables FastSLAM to be applied to environments far larger than could be handled by the EKF. In this paper, we show that FastSLAM also substantially outperforms the EKF in environments with ambiguous data association. The performance of the two algorithms is compared on a real-world data set with various levels of odometric noise. In addition, we show how negative information can be incorporated into FastSLAM in order to improve the accuracy of the estimated map.

This paper describes two robotic systems developed for acquiring accurate volumetric maps of underground mines. One system is based on a cart instrumented by laser range finders, pushed through a mine by people. Another is a remotely controlled mobile robot equipped with laser range finders. To build consistent maps of large mines with many cycles, we describe an algorithm for estimating global correspondences and aligning robot paths. This algorithm enables us to recover consistent maps several hundreds of meters in diameter, without odometric information. We report results obtained in two mines, a research mine in Bruceton, PA, and an abandoned coal mine in Burgettstown, PA.

In this paper we describe our open-source robot control software, the Carnegie Mellon Navigation (CARMEN) Toolkit. The ultimate goals of CARMEN are to lower the barrier to implementing new algorithms on real and simulated robots and to facilitate sharing of research and algorithms between different institutions. In order for CARMEN to be as inclusive of various research approaches as possible, we have chosen not to adopt strict software standards, but to instead focus on good design practices. This paper outlines the lessons we have learned in developing these practices.

Presents a probabilistic algorithm for simultaneously estimating the pose of a mobile robot and the positions of nearby people in a previously mapped environment. This approach, called the conditional particle filter, tracks a large distribution of person locations conditioned upon a smaller distribution of robot poses over time. This method is robust to sensor noise, occlusion, and uncertainty in robot localization. In fact, conditional particle filters can accurately track people in situations with global uncertainty over robot pose. The number of samples required by this filter scales linearly with the number of people being tracked, making the algorithm feasible to implement in real-time in environments with large numbers of people. Experimental results illustrate the accuracy of tracking and model selection, as well as the performance of an active following behavior based on this algorithm.