For the purpose of the initiative an advanced robot was defined as a machine or system capable of accepting high level mission oriented commands, navigation to a work place and performing complex tasks in a semistructured environment with a minimum of human intervention. Such a semi-autonomous device would exhibit various attributes of intelligent operation the totality and integration of which would certainly challenge existing capabilities. The application areas envisaged usually encompassed the nuclear industry, space, underwater, construction, health care and general service functions such as surveillance and cleaning.

This chapter will focus on mechanisms and technologies for implementing machine intelligence. Nevertheless the ability to learn must be one criterion for describing intelligent behaviour. Another one closely related to robotics is a machine that is able to act autonomously in the presence of uncertainty.

This chapter discusses the advanced control systems for robotic arms. The robot control problem can be divided into two main areas: kinematic control, the co-ordination of the links of the kinematic chain to produce desired motions of the robot, and dynamic (or kinetic) control, driving the actuators of the mechanism to follow the commanded positions/velocities.

This chapter discusses intelligent gripping systems. The gripping system is composed of a new system of theories and methodologies, a novel fingertip sensor system, and a versatile user-oriented computer package. The new system of theories has shown great advantage in grasping arbitrary objects. The implementation of the gripping system has been carried out by a series of experiments on the grasping of practical objects.

Force control is a central requirement if robot arms are to use tools or interact with workpieces in an unstructured environment. This chapter concentrates on a form of force control called active compliance control. Experimental results are presented to illustrate the difficulties encountered when designing such a force controller on a practical robot. A new theorem on the stability of such a controller is also proved. The paradigm considered is the use of tele-robots' arms which are operated under human supervision while carrying out complex tasks. The application area considered is nuclear decommissioning but the technology is widely applicable to such diverse activities as bomb disposal and the 'virtual reality' simulators of drug interactions.

Robots are often required to function in environments that are too dangerous or expensive for direct human operation, but, computer control and intelligence are not sufficiently developed to permit them to operate under their own initiative. In these instances teleoperational control forms a popular solution set. For optimum performance under these conditions the operator would wish to control an advanced instrumented robot, comparable in function (where possible) with the human body, integrated into a sophisticated tele-presence system to provide the operator with a full range of motion inputs and sensory feedback data. This chapter considers the development of input, control and feedback (visual, audio and tactile) systems (man-machine interface) for a twin armed mobile robot to be used in tele-presence applications. Modules for leg, arm, hand and head motion monitoring have been developed to control the activities. Cameras, microphones and a multi-functional tactile sensing system provide feedback signals that give the operator a relatively realistic impression of the robot's activities. The signals from these sensors can be fed to the operator as direct stimuli, comparable in many instances with normal sensations. All the feedback modules have been made sufficiently light and compact that they can be worn by the operator during normal usage, without restricting motion or comfort. Comparisons of various visual, control, audio and tactile techniques have also revealed the optimum sensory configurations for specific tasks.

In this chapter sensors which are being developed for robotics, and how they might be used together to provide a robust facility for real-time sensing, planning and control are described. The main requirement for robot control and planning is to determine the range of nearby objects i.e. how far away they are from the robot.

The nuclear industry relies on telerobotics for inspection, refurbishment and to carry out basic plant operations, in radioactive facilities. A wide range of remote handling systems have been developed to carry out varied tasks. The remote handling systems can range from simple grab and release mechanically linked systems to complex hydraulically/electrically linked systems with special features such as programmability, resolved motion and collision detection. The driving forces to continue development of remote handling systems come from the need to make the plant more efficient, carry out complex tasks cheaply and quickly and to satisfy new regulations. This chapter is based on systems used in British Nuclear Fuels plc (BNFL).

The term 'robotic surgery' is usually regarded by researchers to mean a motorised reprogrammable computer controlled device which can carry a series of sensors and tools as an aid to diagnosis, therapy or surgery. However many clinical groups also use the term 'robot surgery' for the area which more properly should be considered 'computer assisted surgery'. The majority of applications to date have been in this area, partially because they are regarded as safer than active robot systems. For this reason a preliminary discussion of Computer Assisted Surgery is given below.

This chapter describes the difficulties of developing an intelligent autonomous car that can drive itself through the current traffic environment. Such systems will not appear for a long time although they may be introduced sooner if the traffic environment is greatly constrained. More viable for the medium term are semi-autonomous systems where the driver maintains responsibility for the overall control of the car. Examples of this type of system are autonomous intelligent cruise control, lane support and collision warning. For each of these examples, the benefits and the likely technologies are described and results shown from demonstrators under development at Lucas.

In this chapter design aspects and experiences on walking machines are summarised and critically analysed. Experiences are based on designing the MECANTI machine, which is a research and development test-bed for work machine applications in an outdoor environment. Considerations are focused to the mechatronics of the machine and especially to the control system. The control system is based on two level hierarchy and distributed control philosophy which follow the canonical layout. Differences between the requirements of engineering design and what nature has realised in animals are considered.

Flexible materials are in widespread use in the home and in industry. A particularly important subset is that of sheet-like limp materials such as fabrics which have low resistance to out-of-plane bending. Yet it is only recently that the automated handling of such materials has received much attention, arising from attempts by researchers to automate the assembly processes. The apparel and footwear industries rarely get the media, political and academic attention merited by their importance in economic and employment terms. With textiles, the apparel industry was the source of over 10% of total industrial output, and in many countries was in the top three industrial sectors in terms of value of production and employment. It should also be noted that many of the industrialised countries also have significant machinery and IT support industries. This threat of competition from low labour cost countries has given rise to a number of research programs which aim to reduce labour content and help ensure consistently high quality production which can adapt rapidly to market needs.

Many applications exist for robotics from ready meal assembly through to the packaging of cooked products in highly demanding conditions for this industry. The use of robotics may be more obvious in the meat, fish and poultry sector.

The Silsoe Automatic Milking System consists of a novel high speed pneumatic robot and associated hardware, software systems and dairy parlour equipment. The system will allow voluntary attendance by cows over 24 hours with minimal human supervision. The control software for the robot and end-effector has to guide the teat cup onto an unpredictable, moving target and cope with different cows, udders and teat sizes. The pneumatic actuators are compliant and not likely to damage the cow. Mechanical contact sensors pressed onto the cow by air springs monitor the position of the cow when in the milking stall. This information is combined with signals from position sensors in the end effector to guide the robot. The present robot has been used on several extended trials including voluntary milking. This chapter addresses system design, control problems associated with the pneumatic milking robot and discusses some milking trial results.

This chapter describes an on-going collaborative research project between the University of Brighton and Applied Materials Ltd., a manufacturer of semiconductor process equipment. A robot manipulator places semiconductor wafers with extreme reliability within this equipment. Advances in semiconductor technology have prompted demands for even higher levels of reliability in the manipulator, but have also led to greater diversity in the physical properties of the wafers, posing great problems for the manipulators' sensing systems. A comprehensive survey of suitable sensing techniques has found vision to have the potential to overcome these problems, and a systemic approach to the implementation of machine vision is presented.

Societies are formed as collaborating structures to execute tasks which are not possible or are difficult for individuals alone. There are many types of biological societies, but societies formed by machines or robots exist only in laboratories at the moment. The concept offers, however, interesting possibilities especially in applications where a long-term fully autonomous operation is needed and/or the work to be done can be executed in a parallel way by a group of individuals. This chapter introduces the basic control and communication structures for robot societies by using a model society. Also a mini-scaled mobile robot which is under construction, and will be duplicated for testing as a physical society demonstrator, is introduced. Simulation results illustrating the behaviour of the model society are given. Some potential applications are introduced and discussed.

There is a growing interest worldwide in the concept and possible applications of miniature, micro and even nano scale devices, including robots. This interest is motivated by the widespread perception that advances in microfabrication technology can lead, together with the progresses of robotics research, to the development of autonomous or tele-operated machines capable of carrying out useful tasks at a miniature, micro and perhaps nano scale, thus substantially extending the range of application of current machines and possibly leading to a technological breakthrough whose effects could be comparable with those determined by the micro-electronic revolution. As is often the case for new areas of application, and despite the increasing popularity of the field, the very same notions of 'micromachine' and of 'microrobot' have not yet been defined clearly. In fact, such terms as 'microsystems', 'microelectromechanical systems', 'micromechatronics', 'micromechanisms', 'micromachines' and 'microrobots' are used almost like synonyms to indicate a wide range of devices whose functioning is related to the 'fuzzy' concept of operation at 'small' scale. The first serious attempt to provide some guidelines for research in this area has been made in Japan: during the preparatory phase for the 'Large Scale Project on Micromachine Technology', the Japanese MITI has promoted the activity of panels of experts, who were asked to provide definitions, identify application domains, and highlight theoretical knowledge and technologies which are necessary to implement real devices. The results of this analysis are presented in papers and reports which represent perhaps the most accurate and comprehensive discussion so far available for approaching the field of micromachines and microrobots.

In this chapter we review some of the recent and more significant contributions to behaviour-based robotics. In doing so, we attempt to identify common characteristics of the various implementations and examine the relationship between potentially conflicting approaches. While most workers have concentrated on the mobile robot domain, our own interest is in exploiting the potential benefits of the behaviour-based approach in robotic handling and assembly applications in appropriate industrial sectors.

Research into the procedural control of robotic devices has seen a resurgence of interest in recent years. However, work has tended to focus upon solitary mobile robot scenarios as opposed to those situations where multiple mobile robots are required to communicate and co-operate with each other and use their combined functions to achieve a particular task. Researchers at the University of Salford are involved in a long-term research programme into the area of co-operant mobile robots and a novel behaviour synthesis architecture (BSA) has been designed and implemented for the specific purpose of controlling multiple co-operating mobile robots. To illustrate the application of the BSA, the chapter focuses upon a material handling operation, where two mobile robots are given the task of co-operatively relocating an object while having to avoid an obstacle en route. Details of the mobile robots' task achieving behaviour are provided firstly via simulation results, based upon our mathematical model of the BSA, and secondly via the results of having implemented the architecture on two real mobile robots. A discussion of our theoretical and experimental findings is presented and the chapter highlights a number of important areas for future research.

Soft automation in the form of robotics can be applied to low count production runs, variable tasks in assembly and testing and can achieve increased quality and quantity of production. These necessary abilities are due to the in-built repeatability and relative ease of adjustment and feedback. This chapter describes how, even with robotic execution of tasks, there are still considerable costs in fixed jigging and tools due to the limited dexterity of commercial arms or the cost of possible redundancy in the production of special purpose single-arm robots to carry out complex tasks. The multi-arm systems solution, described here, removes some of the limitations and peripheral costs of further application of robotics, although this is at the cost of greater complexity in intelligence, planning and control. Future developments in the areas outlined in this chapter will be the key to fully flexible automation and high technology production. Initially a description of the range of multi-arm systems is given. Then work in enabling research which is required for multi-armed systems is outlined. This is followed by details of research at the Queen's University of Belfast into task planning, path planning and collision detection, multi-arm sensing, hardware architectures for practical implementation and maintenance of reliability.

Human labour remains widely used in many production systems, although this is often not ideal due to such factors as inconsistencies, high costs, and physical dangers. However, replacing a specific individual is often expensive, even when the individual's role appears to be simple and relatively mundane. Tasks that appear to be very simple, such as simple visual inspection, can often be the most difficult to automate. Conversely tasks that to the layman appear very complex, such as the interpretation of mass spectography data, can be performed by a computer 'Expert System' relatively easily. In between these two domains lies a third area; the fuzzy grey terrain of the professional craftsman. It is in these first and third areas, which to date have proved intractable to conventional computer approaches to automation, that neural methods have shown particular promise. This chapter will describe two industrial problems and illustrate how neural computing is being used to help in their automation.

This chapter reports on recent research on advanced motion planning and control of articulated and mobile robotic systems. In addition to employing the principles of distributed and parallel processing to produce feasible real-time multi-processor networks, other new theories such as neural networks and genetic algorithms are deployed as possible solutions. All reported algorithms are implemented for either the PUMA 560 arm or the B12 mobile robot. The ultimate aim of the on-going research is to present working architectures for real-time robotic systems by augmenting all developed structures.