Many walking machines have six legs, since this allows an efficient and stable gait in which the machine is supported by two tripods of legs alternately. It is common to give each leg three powered joints so each foot can be moved up and down, backwards and forwards, and from side to side. This requires a total of 18 actuators, each with a driving device such as an amplifier or servovalve, and the associated cables or hoses. The weight, cost and complexity of all this equipment presents an obstacle to the widespread use of legged vehicles. Further, some manoeuvres such as turning, or even straight-line walking in some designs, requires the co ordinated action of several joints, presenting a considerable problem of computing, in real time, the joint angles and velocities needed, and driving the actuators to produce them. For many purposes there is no alternative to the fully servo-controlled 18 degree of freedom vehicle. However, at this early stage in the development of walking robots some investigations can be done with cruder machines. For example, there has been little research on the interaction of legged vehicles with different surfaces such as sand, mud and loose earth. Studies of the interaction of vehicle loading and foot design with surface conditions do not need a high degree of adaptability. Another example is the development of aids to stable locomotion such as balancing devices. This chapter describes such a simple vehicle. The next sections cover the methods used to reduce complexity: of these the most unusual is the steering geometry and mechanism.

As part of a joint project between UMIST and Leyland Vehicles Limited under the Teaching Company Scheme, the feasibility of using a robot to automate the asssembly of components for commercial vehicles was investigated. The company were very keen to integrate the C.A.D. and C.A.M. aspects of the project. The envisaged system has many similarities to the system developed elsewhere for the manufacture of electrical looms, described by Gibbons. However, because of the low-volumes and variety of looms the production time lost with conventional teach-and-repeat methods would be significant. The recent availability of an off-line robot programming package for the company's C.A.D. system meant that both of these aspects could be tackled. The use of off-line programming from a C.A.D. system is a relatively new concept with associated high technical risk, and so a development robot-assembly cell was set up to demonstrate that the technique was practicable. Many problems were encountered but were overcome, and the project has been very successful in confirming the viability of automating this assembly process.

This chapter examines the programming task required by a robot workstation within a flexible assembly system. A system is described which optimises programming by use of off-line programming and automatic location editing. The system uses a robot equipped with two simple two state touch sensors which are manipulated to provide redundant data capable of accurately describing various types of 5 or 6 degree of freedom locations. Locations are determined by automatic search routines and edit algorithms, which enables automatic calibration of the sensors. The implications of the results are discussed and it is concluded that automatic location editing can increase both productivity and flexibility of assembly workstation.

The article discusses the collision avoidance between robots operating in the same cell. The cell controller operates the cell by sending instructions to the individual (local) hardware and robot controllers, the latter as a sequence of subroutines. For multi-robot cells the Route Consultant is an additional requirement. Its function is to interpret the overall status of the cell and thence to advise the cell controller accordingly.

This chapter will concentrate on the research work being carried out at the University of Wales Institute of Science and Technology (U.W.I.S.T), sponsored by the Science and Engineering Research Council, concerned with the development of intelligent robotic devices to enhance automatic assembly systems. Industrial robots have recently been introduced to the assembly environment with the aims of doing the work of several work stations. This results in a cheaper assembly system that is economically viable at lower production volumes and can be modified as product developments are introduced. The authors work on adaptable handling systems shows that relatively simple devices can be manufactured to accommodate misalignments between parts. By the use of appropriate control strategies chamfered pegs can be assembled whether or not initial contact occurs on the chamfer.

This chapter discusses robot force sensing. The stochastic force sensing can be useful, inexpensive and straightforward technique for integrating force monitoring in a robot. More study is needed to transfer this technique to an efficient and universal force sensing methodology, suitable for all robot control applications.

In order for an industrial robot to handle components correctly during an assembly process, the components are usually presented to the robot in a known orientation using some form of jigging. If the component cannot be orientated using jigging or simple sensory devices then the assembly process is usually performed by a human operator. In order to overcome the problem of component orientation a vision system can be interfaced to an industrial robot which can identify and calculate the position and orientation of the component. This chapter discusses the application and flexibility of such a robot vision system.

The potential benefits of equipping arc wielding robots with vision to provide adaptive control for seam tracking have been widely accepted. A limited survey of industrial users' requirements suggested a more robust and cost effective solution was needed. This chapter describes continuing work at UMIST to build a practical industrial device to meet this requirement, based on earlier work using a single linescan, and avoiding the use of laser illumination .

Imperial College and Lucas Research Centre have a collaborative research program whose aim is to investigate the feasibility of automating the assembly of motor car headlamps. It was decided that a Flexible Manufacturing System (FMS) is required which would not only identify and assemble various types of headlamps but also be easily reprogrammed to identify and assemble new or modified headlamps. This is an essential requirement since headlamps are cosmetic items with a limited product life cycle. The FMS would include independent Robot Cells executing independent tasks of the headlamp assembly (the insertion of fixing 'tulips' onto reflectors, the insertion of lamps, casing assembly and so on). A cheap, fast and flexible Vision System was required within six months to identify headlamp reflectors prior to assembly. This chapter will describe such a system for the purpose of 'tulip' insertions.