Gary K Shaffer

A robotic system for underground coal mining

The authors describe a system that automates a continuous miner, enabling it to maneuver in highly constrained environments and cut coal without a human operator onboard. The system consists of a modified continuous miner, a laser range sensor, a SPARCstation, and control software. To date, the system has been tested on a mobile robot and on a continuous miner both above ground and in a real coal mine. The authors note that this system is the first instance of an intelligent robotic system for cutting coal.<<ETX>>

Position estimator for underground mine equipment

A 2-D perception system that exploits the accuracy and resolution of a laser range sensor to determine the position and orientation of a mobile robot in a mine environment is described. The perception system detects features from range sensor data and matches the features to a map of the mine to compute the sensor position. The features used are line segments and corners, which represent the typical geometry of the mine walls and intersections found in room-and-pillar-type mining. The position estimate is refined by minimizing the error between the map and sensed features. This position information can be used for autonomous navigation when a map of the mine is available or to survey the mine in order to build such a map. The technique is applied to robot navigation in a mine mockup. A refinement of this system could guide machines to yield productive, safe mining operations. >

Comparison of two range-based pose estimators for a mobile robot

Pose estimation, or self-location, is a fundamental requirement for a mobile robot which enables it to navigate along a planned path or to position itself within its environment. This paper compares two 2D pose estimation algorithms: one is feature-based, and the other is iconic. Both techniques match range data to a map to calculate x, y, and (theta) (the pose) of the range sensor within its 2D environment. The metrics for the comparison include accuracy, processing time, number of range data points, robustness to errors in the map, sensitivity to the initial rough pose estimate, and environmental models. The feature-based approach is better with respect to processing time and the initial estimate; the iconic method has the advantage with respect to accuracy, the number of data points, and the environment model. For the test environment used in this paper, both methods are robust to errors in the map. A hybrid approach is proposed which combines the two estimators in a way that exploits the advantages of both.