Richard Goodwin

Lessons learned from Xavier

We have been running an experiment in web-based interaction with an autonomous indoor mobile robot. The robot, called Xavier, can accept commands to travel to different offices in our building, broadcasting camera images as it travels. The experiment, which was originally designed to test a new navigation algorithm, has proven very successful. This article describes the autonomous robot system, the web-based interfaces, and how they communicate with the robot. It highlights lessons learned during this experiment in web-based robotics and includes recommendations for putting future mobile robots on the web.

Experience with rover navigation for lunar-like terrains

Reliable navigation is critical for a lunar rover, both for autonomous traverses and safeguarded remote teleoperation. This paper describes an implemented system that has autonomously driven a prototype wheeled lunar rover over a kilometer in natural, outdoor terrain. The navigation system uses stereo terrain maps to perform local obstacle avoidance, and arbitrates steering recommendations from both the user and the rover. The paper describes the system architecture, each of the major components, and the experimental results to date.

Xavier: experience with a layered robot architecture

Office delivery robots have to perform many tasks such as picking up and delivering mail or faxes, returning library books, and getting coffee. They have to determine the order in which to visit locations, plan paths to those locations, follow paths reliably, and avoid static and dynamic obstacles in the process. Reliability and efficiency are key issues in the design of such autonomous robot systems. They must deal reliably with noisy sensors and actuators and with incomplete knowledge of the environment. They must also act efficiently, in real time, to deal with dynamic situations. To achieve these objectives, we have developed a robot architecture that is composed of four layers: obstacle avoidance, navigation, path planning, and task planning. The layers are independent, communicating processes that are always active, processing sensory data and status information to update their decisions and actions. A version of our robot architecture has been in nearly daily use in our building since December 1995. As of January 1997, the robot has traveled more than 110 kilometers (65 miles) in service of over 2500 navigation requests that were specified using our World Wide Web interface.

Formalizing Properties of Agents

There is a wide gulf between the formal logics used by logicians to describe agents and the informal vocabulary used by people who actually build robotic or software agents. In an effort to help bridge the gap, this report applies techniques borrowed from the field of formal software methods to develop a common vocabulary. Terms useful for discussing agents are given formal definitions. A framework for describing agents, tasks and environments is developed using the Z specification language. The terms successful, capable, reactive, reflexive, perceptive, predictive, interpretive, rational and sound are then defined in terms of this framework. In addition, a hierarchy for characterizing tasks is given. The aim of this report is to develop a precise vocabulary for discussing and comparing agents.

Robot Navigation with Markov Models: A Framework for Path Planning and Learning with Limited Computational Resources

Navigation methods for mobile robots need to take various sources of uncertainty into account in order to get robust performance. The ability to improve performance with experience and to adapt to new circumstances is equally important for long-term operation. Real-time constraints, limited computation and memory, as well as the cost of collecting training data also need to be accounted for. In this paper, we discuss our evolving architecture for mobile robot navigation that we use as a test-bed for evaluating methods for dealing with uncertainty in the face of real-time constraints and limited computational resources. The architecture is based on POMDP models that explicitly represent actuator uncertainty, sensor uncertainty, and approximate knowledge of the environment (such as uncertain metric information). Using this model, the robot is able to track its likely location as it navigates through a building. Here, we discuss additions to the architecture: a learning component that allows the robot to improve the POMDP model from experience, and a decision-theoretic path planner that takes into account the expected performance of the robot as well as probabilistic information about the state of the world. A key aspect of both additions is the efficient allocation of computational resources and their practical application to real-world robots.

Rational handling of multiple goals for mobile robots

The mobile robot planning domain is dynamic, with goals becoming active asynchronously. In order to successfully operate in this environment, a robot must be able to interrupt and reformulate its plan of action on-the-fly. This report investigates a method for incorporating the accomplishment of a new goal into a partially executed plan. A decision theoretic approach using net present value as the decision criterion serves as the basis for determining goal ordering dynamically. The appropriateness of net present value over other criteria is argued. The approach has been implemented on a robot operating in an office setting. Examples from this domain and a planetary exploration domain are used to show the advantages of the approach with respect to fixed priority and heuristic approaches.

Using neural networks in agent teams to speedup solution discovery for hard multi-criteria problems

Evolutionary population-based search methods are often used to find a Pareto-optimal set of solutions for hard multicriteria optimization problems. We utilize one such agent architecture to evolve good solution sets to these problems, deploying agents to progressively add, modify and delete candidate solutions in one or more populations over time. Here we describe how we assign neural nets to aid agent decision-making and encourage cooperation to improve convergence to good Pareto optimal solution sets. This paper describes the design and results of this approach and suggests paths for further study.