Showing papers by "S. Sitharama Iyengar published in 1991"

Autonomous Mobile Robots

Abstract: In recent years, autonomous mobile robotics is increasingly becoming an effective way to carry out tasks that are difficult, dangerous, or simply boring for humans. Relevant examples include planetary exploration and search and rescue. While developing a system of autonomous robots, a designer should take care of fundamental issues that relate to locomotion, sensing, localization, and navigation. One of the most important and challenging aspects that could significantly impact on the system performance is the autonomous decision making to carry out the assigned task. This comprises the set of techniques that allow autonomous mobile robots to decide the next location to reach and to possibly coordinate among themselves, according to their current knowledge of the world they operate in. Consider for example the exploration problem, where mobile robots are employed for incrementally discovering and mapping the features of initially unknown environments, which could serve for more general missions, such as map building and search and rescue. The robots have to select the next locations where to perform sensing actions within the currently explored portion of the environment and who goes where. Clearly, the decisions made could have a significant impact on the performance of the exploration and, if made effectively, could really boost the robots’ autonomy. However, despite the importance of the development of techniques to make mobile robots more autonomous, general techniques are not mature yet. In this dissertation, we focus on some key activities for the multirobot exploration problem, namely the selection of interesting locations (exploration strategy) and their assignment to robots (coordination method). We

A 'retraction' method for learned navigation in unknown terrains for a circular robot

Abstract: The authors consider the problem of learned navigation of a circular robot R, of radius delta (>or=0), through a terrain whose model is not a priori known. The authors consider two-dimensional finite-sized terrains populated by an unknown (but finite) number of simple polygonal obstacles. The number and locations of the vertices of each obstacle are unknown to R; R is equipped with a sensor system that detects all vertices and edges that are visible from its present location. The authors deal with two problems: the visit problem and the terrain model acquisition problem. In the visit problem, the robot is required to visit a sequence of destination points, and in the terrain model acquisition problem, the robot is required to acquire the complete model of the terrain. The authors present an algorithmic network framework for solving these two problems based on a retraction of the free space onto the Voronoi diagram of the terrain. >

Information integration and synchronization in distributed sensor networks

Abstract: The computational, i.e. architectural, algorithmic, and synchronization, issues related to competitive information integration in a distributed sensor network (DSN) are addressed. The proposed DSN architecture consists of a set of binary trees whose roots are fully connected. Each node of the tree has a processing element and one or more sensors associated with it. The information from each of the sensors has to be integrated in such a manner that the communication costs are low and that the real time needs are met. An information integration algorithm that has a low message cost (linear in the number of nodes of the network) and a low distributed computation cost is presented. The problems associated with synchronizing information to be integrated in the presence of imperfect clocks is considered. The fault tolerant features of the network and the integration algorithm are discussed. >

A general greedy channel routing algorithm

Abstract: A general approach for the channel routing problem is presented as a framework for a class of heuristic routing algorithms. The algorithm is shown to possess a backtracking capability that increases the chance of completing the routing with a minimum number of tracks. Since the concepts described are general, they can be applied to other channel problems, such as switchbox routing, three-layer routing, and multilayer routing, or even to the overlap model, with only a few modifications. It is shown that track-oriented greedy algorithms can be modified to solve other channel routing problems. As examples, the algorithm is modified to solve the Manhattan switch-box problem and channel routing problems in the overlap and knock-knee models. Preliminary results show that the modified algorithms have good performance and show strong potential to outperform existing algorithms. Applying the algorithm MCRP-ROUT to the benchmark Deutsch's difficult problem and Burstein's difficult problem, routing solutions of 19 tracks and six tracks, respectively, were obtained. >

Functional characterization of fault tolerant integration in distributed sensor networks

Abstract: Fault-tolerance is an important issue in network design because sensor networks must function in a dynamic, uncertain world. A functional characterization of the fault-tolerant integration of abstract interval estimates is proposed. This model provides a preliminary version for a general framework that is hoped to develop to address the general problem of fault-tolerant integration of abstract sensor estimates. A scheme for narrowing the width of the sensor output in a specific failure model is proposed and given a functional representation. The main distinguishing feature of the model over the original model of K. Marzullo (1989) is in reducing the width of the output interval estimate significantly in most cases where the number of sensors involved is large. >

Functional characterization of sensor integration in distributed sensor networks

Abstract: Fault-tolerance is an important issue in network design because sensor networks must function in a dynamic, uncertain world. In this paper, the authors propose a functional characterization of the fault-tolerant integration of abstract interval estimates. This model provides a test bed for a general framework which the authors hope to develop to address the general problem of fault-tolerant integration of abstract sensor estimates. They further propose a scheme for narrowing the width of the sensor output in a specific failure model and give it a functional representation. The main distinguishing feature of the authors' model over the original Marzullo's model is in reducing the width of the output interval estimate significantly in most cases where the number of sensors involved is large. >

Neurocomputing formalisms for computational learning and machine intelligence

Abstract: Publisher Summary The chapter discusses the capabilities of neural-network learning that is central to the deeper question of its feasibility to artificial intelligence. It focuses on machine learning in the context of neural networks from the standpoints of computational complexity and algorithms information theory, and on the emerging area of learning theory in the context of dynamic systems. Engineered intelligent systems behave with remarkable rigidity when compared with, their biological counterparts, especially in their ability to recognize objects or speech, to manipulate and adapt in an unstructured environment, and to learn from past experience. A major reason for this limited technical success in emulating some of the more fundamental aspects of human intelligence lies in the differences between the organization and structuring of knowledge, and the dynamics of biological neuronal circuitry and its emulation using the symbolic-processing paradigm. The chapter rederives a theoretical framework for neural learning of nonlinear mappings, wherein both the topology of the network and synaptic interconnection strengths are evolved adaptively. The proposed methodology exploits a new class of mathematical constructs, terminal attractors, which provide unique information-processing capabilities to artificial neural systems. Terminal attractor representations are used not only to ensure infinite local stability of the encoded information, but also to provide a qualitative as well as quantitative change in the nature of the learning process. The chapter also draws from mathematical constructs in sensitivity theory for nonlinear systems to illustrate the notion of forward and adjoint-operators. The formalism exploits the concept of adjoint-operators to enable a fast global computation of the network's response to perturbations in all system parameters. This formalism eliminates the heuristic overtones of the proposed framework.

New results on channel routing

Abstract: Presents a new routing concept which guides the selection of wire segments in track-by-track fashion by inspecting the effects of the endpoints of each selected wire segment to column density and vertical constraint graph of the given channel routing problem. This new routing concept has been implemented in the two-layer and three-layer routers. The routing performance of the developed two-layer and three-layer routers has overwhelmingly outperformed all the currently existing two-layer and three-layer routers in most examples in the literature as shown in experimental results. >

An set refinement algorithm with applications

Abstract: Given two partitions A and B of a totally ordered set S, of size n, corresponding to the equivalence relations R A and R B respectively, the refinement problem calls for the computation of the partition C corresponding to the relation R A and R B We present an algorithm for this problem As an application, we present an algorithm that converts an inverted lists structure to a multiple attribute tree for a set of multi-dimensional points Using the idea of the proposed algorithm we present an alternate algorithm for an existing partition algorithm The alternate algorithm is very simple to implement and direct to prove

A Taxonomy on Event-Driven Production Systems

Abstract: Intelligent Computing Systems require an integration of the traditional Expert system technologies with real-time response and control capabilities. The purpose of this paper is to exploit the synergism between the use of an event-driven Al-based expert system and the issues in deriving parallel algorithms for various applications. The two issues that will be considered in this proposal are: (i)the effect of variations in the dynamics of the environment on overall system stability, and (ii) the structure and properties of event-driven parallel algorithms for some real time applications.