Paul Freedman

Mapping in unknown graph-like worlds

We consider the problem of constructing a map of an unknown environment by an autonomous agent such as a mobile robot. Because accurate positional information is often difficult to ensure, we consider the problem of exploration in the absence of metric (positional) information. Worlds are represented by graphs (not necessarily planar) consisting of a fixed number of discrete places linked by bidirectional paths. We assume the robot can consistently enumerate the edges leaving a vertex (that is, it can assign a cyclic ordering). A mobile robot is assigned the task of creating a topological map, i.e. a graph-like representation of the places in the world and their connectivity, by moving from place to place along the paths it encounters. It can detect edges and count them, but cannot directly sense the labels associated with a place or an edge. In principle, this type of representation could be used for non-spatial environments such as computer networks. We present an approach to the exploration of unknown environments for which local sensing information alone is insufficient to distinguish distinct places, based simply on the structure of the world itself. Places are identified by a non-unique signature and by using a collection of such non-unique local signatures, an extended signature may be constructed that determines the robots position (although in certain “degenerate” worlds additional information is required). We describe the “exploration tree” as a representation of a collection of alternative descriptions of the environment. In addition, heuristics are presented that can accelerate the search for the correct world model.

Surface sensing and classification for efficient mobile robot navigation

Mobile robot navigation and localization is frequently aided by, or even dependent upon, a good estimate of the rate of dead-reckoning error accumulation. Sensor data can be used for position estimation, but this often involves overheads in acquiring and processing the data. By sensing and then classifying the surface type, an estimate of the rate of error accumulation for dead-reckoning allows one to estimate accurately how often localization, including sensor data acquisition, must be performed. The authors describe experiments in which a boom-mounted microphone is tapped on different floor materials, much as a blind man might tap his cane. The acoustic signature arising from the contact is then used to classify the floor type by comparing a windowed power spectrum of the acoustic signature with one of a family of prototypical signatures generated statistically from the same material. The technique is low-cost, involves limited computational expense, and performs very well.

The analysis and optimization of repetition within robot workcell sequencing problems

The authors introduce a classification of sequencing problems according to the presence of nondeterminism, whether due to the structure of the problem itself (internal) or to online events (external). They develop a theory, in the context of timed Petri nets, of how repetition can be described and predicted. They conclude with a description of a PROLOG-based decision-support system which manages this nondeterminism by exploring the consequences of alternate orderings of workcell operations, and ultimately constructs the time-optimal sequence. The analysis and optimization are illustrated with an example.<<ETX>>

Analyzing the temporal behavior of real-time closed-loop robotic tasks

ORCCAD (Open Robot Controller Computer Aided Design), is a development environment for robotic controllers which provides verification and simulation tools along with a graphical human-machine interface, in order to bridge the gap between control laws as understood by the control systems community, and real-time computing as understood by the computer science community. In this paper, we introduce some recent extensions to ORCCAD which make it possible to perform certain kinds of temporal verification of real-time closed loop robotic tasks.<<ETX>>

SAGE: a decision support system for the sequencing of operations within a robotic workcell

Abstract This paper describes a new approach to decision support based on logic programming. SAGE, Sequence Analysis by Graphical Evaluation, is a both a logic framework and a set of Prolog programs to express and then solve the problem of repetitive sequencing of operations within a robotics workcell. We create a Timed Petri Net from the user description consisting of the initial state of the workcell, plus the operations to be performed. Each operation is defined by (i) the command to be executed, (ii) a list of enabling conditions, (iii) a list of changes made to the state of the workcell by executing the command, and (iv) its relative integer duration. After some preliminary analysis, this net is used to generate the Task Space of all feasible time histories of workcell events. Repetitive sequences of all the workcell operations are then constructed from cycles in the Task space. From the feasible sequences thus identified, the time-optimal one is selected. SAGE clearly demonstrates both the power of symbolic expression, and the power of pattern matching as a basis for decision support systems.

A database design for the runtime environment of a robotic workcell

Abstract Intelligent robotic workcell activities have come to require a database framework for arranging, storing and accessing information in real-time about the workcell environment in a standard way. After a brief introduction to the theory of databases (DB), we examine the general topic of robotic workcells, identify characteristics typical of robotic applications, and then present a survey of DB-related work in the robotics domain. We then construct a set of design constraints based on our analysis of robotic applications, and describe a suitable software architecture. The paper concludes with a discussion of experience gained with two generations of DB implementations.

An integrated programming environment for a generic robotic workcell

The complexity of modern robotic workcells, consisting of many cooperating elements (e.g., robots, sensing systems) makes programming applications enormously difficult. Even when the application can be easily decomposed into workcelll operations, the tight coupling among these operations due to constraints on precedence and resource sharing makes it difficult to identify and take advantage of possible concurrency. Since robotic workcells are typically configured to repetitively perform an application, we seek to minimize the ‘cycle’ time required, by optimizing the order in which operations take place. Eventually, a sequential program must be written for each workcell element, and they must then execute concurrently. Commercial robot languages and systems provide neither the necessary tools for programming such a distributed collection of elements nor a mechanism for inter-element communication. In this article the authors propose a two-part solution to this workcell programming problem: SAGE/WRAP. A graphics interface is provided to the user in order to define the application. Then, from the users input, SAGE builds a Petri Net description of the workcell application; this is used to obtain a time-optimal sequence of operations for each workcell element. The time-optimal sequence then becomes the basis of the runtime program that performs the application. This program is interpreted and executed by the hierarchical workcell runtime environment called WRAP. An example is provided to illustrate the approach.

Modelling the actions of an intervention robot

Reviews the inadequacies of models of robot actions, proposed in the literature for intervention tasks. The STRIPS-inspired model of robot actions defined in terms of pre-conditions and post-conditions is extended by defining conditions which qualify the execution contexts of the actions. It is demonstrated that when all of the conditions are defined in terms of state variables and their possible values it becomes possible to study the soundness of the specifications of the possible actions to be performed, and the soundness of the interrelationships between these specifications. To this end, a particular kind of Petri net is proposed. An example of a mobile robot is also presented.<<ETX>>