C.R. Weisbin

Autonomous mobile robot navigation and learning

Research focused on the development and experimental validation of intelligent control techniques for autonomous mobile robots able to plan and perform a variety of assigned tasks in unstructured environments is presented. In particular, an autonomous mobile robot, HERMIES-IIB intelligence experiment series, is described. It is a self-powered, wheel-driven platform containing an onboard 16-node Ncube hypercube parallel processor interfaced to effectors and sensors through a VME-based system containing a Motorola 68020 processor, a phased sonar array, dual manipulator arms, and multiple cameras. Research on navigation and learning is examined.<<ETX>>

Robot navigation in an unexplored terrain

Navigation planning is one of the most vital aspects of an autonomous mobile robot. Robot navigation for completely known terrain has been solved in many cases. Comparatively less research dealing with robot navigation in unexplored obstacle terrain has been reported in the literature. In recent times this problem has been addressed by adding learning capability to a robot. The robot explores terrain using sensors as it navigates, and builds a terrain model in an incremental manner. In this article we present concurrent algorithms for robot navigation in unexplored terrain. The performance of the concurrent algorithms is analyzed in terms of planning time, travel time, scanning time, and update time. The analysis reveals the need for an efficient data structure to store an obstacle terrain model in order to reduce traversal time, and also to incorporate learning. A modified adjacency list is proposed as a data structure for storing a spatial graph that represents an obstacle terrain. The time complexities of the algorithms that access, maintain, and update the spatial graph are estimated, and the effectiveness of the implementation is illustrated.

Concurrent algorithms for autonomous robot navigation in an unexplored terrain

Navigation planning is one of the most vital aspects of an autonomous mobile robot. The problem of navigation in a completely known obstacle terrain is solved in many cases. Comparatively less number of research results are reported in literature about robot navigation in a completely unknown obstacle terrain. In recent times, this problem is solved by imparting the learning capability to the robot. The robot explores the obstacles terrain using sensors and incrementally builds the terrain model. As the robot keeps navigating, the terrain model becomes more learned and the usage of sensors is reduced. The navigation paths are computed by making use of the existing terrain model. The navigation paths gradually approach global optimality as the learning proceeds. In this paper, we present concurrent algorithms for an autonomous robot navigation in an unexplored terrain. These concurrent algorithms are proven to be free from deadlocks and starvation. The performance of the concurrent algorithms is analyzed in terms of the planning time, travel time, scanning time, and update time. The analysis reveals the need for an efficient data structure for the obstacle terrain in order to reduce the navigation time of the robot, and also to incorporate learning. The modified adjacency list is proposed as a data structure for the spatial graph that represents the obstacle terrain. The time complexities of various algorithms that access, maintain, and update the spatial graph are estimated, and the effectiveness of the the implementation is illustrated.

HERMIES-III: A step toward autonomous mobility, manipulation and perception

HERMIES-III is an autonomous robot comprised of a seven degree-of-freedom (DOF) manipulator designed for human scale tasks, a laser range finder, a sonar array, an omnidirectional wheel-driven chassis, multiple cameras, and a dual computer system containing a 16-node hypercube expandable to 128 nodes. The current experimental program involves performance of human-scale tasks (e.g., valve manipulation, use of tools), integration of a dexterous manipulator and platform motion in geometrically complex environments, and effective use of multiple cooperating robots (HERMIES-IIB and HERMIES-III). The environment in which the robots operate has been designed to include multiple valves, pipes, meters, obstacles on the floor, valves occluded from view, and multiple paths of differing navigation complexity. The ongoing research program supports the development of autonomous capability for HERMIES-IIB and III to perform complex navigation and manipulation under time constraints, while dealing with imprecise sensory information. 10 refs., 4 figs.

On terrain acquisition by a finite-sized mobile robot in plane

The terrain acquisition problem deals with the acquisition of the complete obstacle terrain model by a mobile robot placed in an unexplored terrain. This is a precursory problem to many well-known find-path and related problems which assume the availability of the complete terrain model. In this paper, we present a method for terrain acquisition by a finite-sized robot operating in plane populated by an unknown (but, finite) number of polygonal obstacles; each obstacle is arbitrarily located and has unknown (but, finite) number of vertices. The robot progressively explores newer vertices of the obstacles using sensor equipment. We show that the complete terrain model will be built by the robot in a finite time. We also show that at any point of time the partially acquired terrain suffices for the navigation of the robot during the exploration. Hence we conclude that the navigation techniques for known terrains can be applied for the robot navigation during exploration.

Learning by an autonomous robot at a process control panel

The authors describe initial work in autonomous learning, using HERMIES-IIB, their current robotic experimental testbed. The integrated system in HERMIES-IIB (the hostile environment robotic machine intelligence experiment, series IIB) is the latest in a series of autonomous, intelligent machines designed to perform in environments that humans cannot readily enter. They describe a system that learns the control panels system dynamics and remembers the most efficient series of responses to shut down a control process in case of future encounters with similar (but not necessarily identical) situations. Ultimately, this system could infer a classification scheme for panel categories, enabling it to hypothesize about correct response sequences for panels not yet encountered. They discuss preliminary work on the systems inferencing section.<<ETX>>

The development of an intelligent nuclear maintenance robot

The US Department of Energy has provided support to four universities and the Oak Ridge National Laboratory (ORNL) to pursue research leading to the development and deployment of robotic system(s) for advanced nuclear power stations. The scope of the program and the R&D effort contributed by ORNL are reported.<<ETX>>

Asynchronous production systems

Intelligent computing systems for aerospace, robotics and process control applications require an integration of traditional expert system technologies with real-time response and control capabilities. To address such requirements, the Asynchronous Production System (APS) is presented, which is a rule-based inference engine capable of monitoring and processing asynchronous, real-time information. The primary motivation of the proposed APS is to provide a unique and convenient mechanism for designing rule-based expert systems capable of dynamic and rapid interaction with their environments. The paper elaborates on the architectural and operational characteristics of the APS and addresses issues pertinent to the the implementation of the APS in a multi-processor environment.

3-D world modeling based on combinatorial geometry for autonomous robot navigation

In applications of robotics to surveillance and mapping at nuclear facilities the scene to be described is three-dimensional. Using range data a 3-D model of the environment can be built. First, each measured point on the object surface is surrounded by a solid sphere with a radius determined by the range to that point. Then the 3-D shapes of the visible surfaces are obtained by taking the (Boolean) union of the spheres. Using this representation distances to boundary surfaces can be efficiently calculated. This feature is particularly useful for navigation purposes. The efficiency of the proposed approach is illustrated by a simulation of a spherical robot navigating in a 3-D room with static obstacles.

Asynchronous production system for control of an autonomous mobile robot in real-time environment

Abstract Autonomous robotic systems designed for hazardous environments require the development of onboard real-time knowledge-based systems capable of generating inference-driven responses to asynchronous, external stimuli. Building such real-time expert systems involves the integration of traditional-knowledge engineering methodologies with time-constrained response and control capabilities. To address such requirements, the Asynchronous Production System (APS), which is a rule-based inference engine capable of dynamic and rapid interaction with its environment, is presented. The APS uses a concurrent-event-driven execution mechanism and an external input-data structure to facilitate the monitoring and processing of real-time asynchronous information. The enhanced conflict-resolution strategies and rule-interruptibility features of the APS execution mechanism allows the APS to make efficient use of the onboard system resources. The implementation of the APS involves the development of a shared-memory, m...