Computer Science and Engineerng

We present empirical results of an auction-based algorithm for dynamic allocation of tasks to robots. The results have been obtained both in simulation and using real robots. A distinctive feature of our algorithm is its robustness to uncertainties and to robot malfunctions that happen during task execution, when unexpected obstacles, loss of communication, and other delays may prevent a robot from completing its allocated tasks. Therefore tasks not yet achieved are resubmitted for bids every time a task has been completed. This provides an opportunity to improve the allocation of the remaining tasks, enabling the robots to recover from failures and reducing the overall time for task completion.

Clustering techniques have been used by many intelligent software agents in order to retrieve, lter, and categorize documents available on the World Wide Web. Clustering is also useful in extracting salient features of related web documents to automatically formulate queries and search for other similar documents on the Web. Traditional clustering algorithms either use a priori knowledge of document structures to de ne a distance or similarity among these documents, or use probabilistic techniques such as Bayesian classi cation. Many of these traditional algorithms, however, falter when the dimensionality of the feature space becomes high relative to the size of the document space. In this paper, we introduce two new clustering algorithms that can e ectively cluster documents, even in the presence of a very high dimensional feature space. These clustering techniques. which are based on generalizations of graph partitioning, do not require pre-speci ed ad hoc distance functions, and are capable of automatically discovering document similarities or associations. We conduct several experiments on real Web data using various feature selection heuristics, and compare our clustering schemes to standard distance-based techniques, such as hierarchical agglomeration clustering, and Bayesian classi cation methods, AutoClass.

We show how paths for articulated robots with many degrees of freedom can be generated in a few seconds or less using nonsystematic parallel search. We present experimental results obtained on a multicomputer for an accurate model of a 7-jointed manipulator arm operating in realistic 3D workspaces. We then present and discuss a fast method for smoothing the paths delivered by the parallel algorithm.

Clustering techniques have been used by many intelligent software agents in order to retrieve, lter, and categorize documents available on the World Wide Web. Clustering is also useful in extracting salient features of related web documents to automatically formulate queries and search for other similar documents on the Web. Traditional clustering algorithms either use a priori knowledge of document structures to de ne a distance or similarity among these documents, or use probabilistic techniques such as Bayesian classi cation. Many of these traditional algorithms, however, falter when the dimensionality of the feature space becomes high relative to the size of the document space. In this paper, we introduce two new clustering algorithms that can e ectively cluster documents, even in the presence of a very high dimensional feature space. These clustering techniques, which are based on generalizations of graph partitioning, do not require pre-speci ed ad hoc distance functions, and are capable of automatically discovering document similarities or associations. We conduct several experiments on real Web data using various feature selection heuristics, and compare our clustering schemes to standard distance-based techniques, such as hierarchical agglomeration clustering, and Bayesian classi cation methods, such as AutoClass.

We present a generalized market architecture that provides support for a variety of types of transactions, from simple buying and selling of goods and services to complex multi-agent contract negotiations. This architecture is organized around three basic components: the exchange, the market, and the session. We also present a negotiation protocol for planning by contracting that takes advantage of the services of the market. We show how the existence of an appropriate market infrastructure can add value to a multi-agent contracting protocol by controlling fraud and discouraging counterspeculation.

In this paper we show that parallel search techniques derived from their sequential counterparts can enable the solution of instances of the robot motion planning problem that are computationally infeasible on sequential machines. We present a simple parallel version of a robot motion planning algorithm based on "quasi best first" search with randomized escape from local minima and random backtracking and discuss its performance on two problem instances, one of which was computationaUy infeasible on a single processor of an nCUBE21 multicomputer. We then discuss the limitations of parallel robot motion planning systems, and suggest a course for future work.

We present a system for multi-agent contract negotiation, implemented as a generalized market architecture called MAGNET. MAGNET provides support for a variety of types of transactions, from simple buying and selling of goods and services to complex multi-agent contract negotiations. In the latter case, MAGNET is designed to negotiate contracts based on temporal and precedence constraints, and includes facilities for dealing with time-based contingencies. In contrast with other approaches to multi-agent negotiation, we introduce an explicit intermediary into the negotiation process, which helps in controlling fraud and discouraging counterspeculation.

We propose an agent for exploring and categorizing documents on the World Wide Web. The heart of the agent is an automatic categorization of a set of documents, combined with a process for generating new queries used to search for new related documents and ltering the resulting documents to extract the set of documents most closely related to the starting set. The document categories are not given a-priori. We present the overall architecture and describe two novel algorithms which provide signi cant improvement over traditional clustering algorithms and form the basis for the query generation and search component of the agent.

We present a generalized market architecture that provides support for a variety of types of transactions, from simple buying and selling of goods and services to complex multi-agent contract negotiations. This architecture is organized around three basic components: the exchange, the market, and the session. We also present a negotiation protocol for planning by contracting that takes advantage of the services of the market. We show how the existence of an appropriate market infrastructure can add value to a multi-agent contracting protocol by controlling fraud and discouraging counterspeculation.

Complex tasks are usually described as high-level goals, leaving out the details on how to achieve them. However, to control a robot, details must be provided. Having the robot move itself to and through an unknown, and possibly narrow, doorway is an example of such a task. We illustrate the difficulty of such a task using actual data from a real robot. We show how the transformation from high-level goals to primitive commands can be performed at execution time and we propose an architecture based on reconfigurable objects that contain domain knowledge and knowledge about the sensors and actuators available. We then show how our approach is used in solving the illustrated task.

We have designed and built a set of miniature robots, called Scouts, and have developed a distributed software system to control them. This paper addresses the fundamental choices we made in the design of the control software, describes experimental results in a surveillance task, and analyzes the factors that affect robot performance.

In this paper we study the performance of multiple robots at a search and retrieval task. The robots have all the same capabilities and perform the same task without any explicit communication. The sensing capabilities of the robots are quite limited, yet the robots succeed at performing the task. We show how the performance is affected by the number of robots working in the same area, the distribution of the targets the robots are searching for, and the complexity of the environment.