John S. Bay

Toward the development of a material transport system using swarms of ant-like robots

A decentralized method for controlling a homogeneous swarm of autonomous mobile robots that collectively transport a single palletized load is proposed. The small tank-like robots have no advanced sensory or communications capabilities. They have no information on the position or number of other robots transporting the small pallet. Instead, all information needed by the robots is derived from the dynamics inherent when the system of robots is contacting a common rigid body. Each robot derives the required local information from a force sensor mounted at the point at which it contacts the pallet. A distributed control law is derived, and the resulting stable behavior of the system is verified by computer simulation.<<ETX>>

Distributed control of simulated autonomous mobile robot collectives in payload transportation

A behavior-based control paradigm that allows a distributed collection of autonomous mobile robots to control the lifting and lowering processes of payload transportation is proposed and then tested with computer simulations. This control paradigm, which represents an approach to solving the cooperative load-bearing problem inherent in multi-agent payload transportation, is based upon a control structure we term thebehavior pathway controller. The behavior pathway controller emphasizes simple, feasible methodologies over complex, optimal methodologies, although we show that with some global self-organization of the collective, the feasible solutions approach and become optimal solutions. Using this controller in simulated environments, our robots demonstrate an ability to function with inaccurate sensor data, which is an important consideration for real world implementations of an autonomous mobile robot control paradigm. The simulated robots also demonstrate an ability to learn their place, or role, within the collective. They must learn their relative roles because they possess no predetermined knowledge about pallet mass, pallet inertia, collective size, or their positions relative to the pallets center of gravity.

Optimal control for cooperating mobile robots bearing a common load

Coordinated motion of a group of mobile robots bearing a common load is investigated. By extending the results of research with similar multi-robot material handling systems, a supervisory control is derived that provides an optimal distribution of work among the robots while accounting for the robots nonholonomic constraints.<<ETX>>

Tactile shape sensing via single- and multifingered hands

For mechanical hands, control of contact points is facilitated by fingertip force sensors which are capable of measuring the local surface normal of an unknown object. In addition to this information, joint angle sensors and knowledge of the kinematic parameters of the hand commonly provide the physical location of the surface contact point. This enables a surface recovery scheme to utilize both kinds of contact information. Here, a simple multifingered surface-exploration procedure is proposed which allows surface data to be obtained from a kinematic simulation. Using a linear formulation of the surface normal data, the shape estimator can be a simple sequential least-squares routine or the more versatile Kalman filter. Estimator performance is then evaluated for hands with no more than five fingers, with and without the surface normal data.<<ETX>>

Geometry and prediction of drift-free trajectories for redundant machines under pseudoinverse control

In this article, classical differential geometric techniques are used for two purposes: (1) to describe in geometric terms the trajectories to which redundant robots under pseudoinverse control will drift and (2) to formulate from this geometric description a practical and quick method for predicting these limit trajectories, thereby suggesting sets of initial conditions under which no drift will occur. Through this analysis, several useful results were obtained. First, it is demonstrated for a 3R example that the drifting trajectories are geodesics in the constrained joint space, and that as a result of this, all drift-free tra jectories are space curves of zero torsion. Second, it is found that each self-motion manifold in the joint space crosses the limiting (drift-free) trajectories at points where they themselves have zero torsion. Because self-motion manifolds are often more easily computed than exhaustive kinematic simulation of manipulator drift, these results provide a fast, new technique for locating and character izing drift-free trajectories based only on geometry of these self-motion manifolds. This work extends the results of previous work in pseudoinverse control to the point that the drift problem may be circumvented without resort to either exhaustive simulation or symbolic computation of Lie brackets of vector fields.

Learning classifier systems for single and multiple mobile robots in unstructured environments

The learning classifier system (LCS) is a learning production system that generates behavioral rules via an underlying discovery mechanism. The LCS architecture operates similarly to a blackboard architecture; i.e., by posted-message communications. But in the LCS, the message board is wiped clean at every time interval, thereby requiring no persistent shared resource. In this paper, we adapt the LCS to the problem of mobile robot navigation in completely unstructured environments. We consider the model of the robot itself, including its sensor and actuator structures, to be part of this environment, in addition to the world-model that includes a goal and obstacles at unknown locations. This requires a robot to learn its own I/O characteristics in addition to solving its navigation problem, but results in a learning controller that is equally applicable, unaltered, in robots with a wide variety of kinematic structures and sensing capabilities. We show the effectiveness of this LCS-based controller through both simulation and experimental trials with a small robot. We then propose a new architecture, the Distributed Learning Classifier System (DLCS), which generalizes the message-passing behavior of the LCS from internal messages within a single agent to broadcast massages among multiple agents. This communications mode requires little bandwidth and is easily implemented with inexpensive, off-the-shelf hardware. The DLCS is shown to have potential application as a learning controller for multiple intelligent agents.

Simulation study of learning automata games in automated highway systems

We propose an artificial intelligence technique called stochastic learning automata to design an intelligent vehicle path controller. Using the information obtained by on-board sensors and local communication modules, two automata are capable of learning the best possible actions to avoid collisions. Although the learning approach taken is capable of providing a safe decision, optimization of the overall traffic flow is required. This can be achieved by studying the interaction of the vehicles. The design of the adaptive vehicle path planner based on local information is extended with additional decision structures by analyzing the situations of conflicting desired vehicle paths. The analysis of the situations and the design of these structures are made possible by treatment of the interacting reward-penalty mechanisms in individual vehicles as automata games.

Navigation of an autonomous ground vehicle using the subsumption architecture

The subsumption architecture is used to provide an autonomous vehicle with the means to stay within the boundaries of a course while avoiding obstacles. A three- layered network has been devised incorporating computer vision, ultrasonic ranging, and tactile sensing. The computer vision system operates at the lowest level of the network to generate a preliminary vehicle heading based upon detected course boundaries. The networks next level performs long-range obstacle detection using an array of ultrasonic sensors. The range map created by these sensors is used to augment the preliminary heading. At the highest level, tactile sensors are used for short-range obstacle detection and serve as an emergency response to obstacle collisions. The computer vision subsystem is implemented on a personal computer, while both ranging systems reside on a microcontroller. Sensor fusion within a subsumption framework is also executed on the microcontroller. The resulting outputs of the subsumption network are actuator commands to control steering and propulsion motors. THe major contribution of this paper is as a case study of the application of the subsumption architecture to the design of an autonomous ground vehicle.

Intelligent control of vehicles: preliminary results on the application of learning automata techniques to automated highway system

We suggest an intelligent controller for an automated vehicle to plan its own trajectory based on sensor and communication data received. Our intelligent controller is based on an artificial intelligence technique called learning stochastic automata. The automaton can learn the best possible action to avoid collisions using the data received from on-board sensors. The system has the advantage of being able to work in unmodeled stochastic environments. Simulations for the lateral control of a vehicle using this AI method provides encouraging results.

Mechatronics education at Virginia Tech

The advent of more complex mechatronic systems in industry has introduced new opportunities for entry-level and practicing engineers. Today, a select group of engineers are reaching out to be more knowledgeable in a wide variety of technical areas, both mechanical and electrical. A new curriculum in mechatronics developed at Virginia Tech is starting to bring students from both the mechanical and electrical engineering departments together, providing them wit an integrated perspective on electromechanical technologies and design. The course is cross-listed and team-taught by faculty from both departments. Students from different majors are grouped together throughout the course, each group containing at least one mechanical and one electrical engineering student. This gives group members the ability to learn from one another while working on labs and projects.