Skip to content

Virtual and Task orientated programming and Total Quality manufacture

Project: Research

Description

The project analysed and identified some critical materials used in the manufacture of energy cables and proposed and developed methods to promote design, materials and manufacture to achieve Total Quality in cable manufacture. Then moved on to investigate motion planning and mage modelling for real time manufacturing applications using virtual task machines. Neural networks and parallel programming methods were investigated for control of a robot and for automatic programming of production machinery (especially for de-flashing plastic parts).

Dr Giles Tewkesbury MBE introduced the concept of influence analysis as a new technique for identifying design focuses within a framework for the automatic programming of advanced production machinery. That work achieved the programming of more than one machine (three were demonstrated) from a single graphical resource (specifically for the automatic programming of advanced production machinery AND de-flashing of plastic parts). A new robot command library which includes simulation was created and a new simulation based robot command library was applied to three robots. The use of distributed intelligence within advanced production machinery for design applications was demonstrated. The work generally aimed to achieve Total Quality Manufacture.

As part of the project, Dr Paul Strickland completed his PhD in 1993 after studying Task Oriented Robotics. The PhD was initially directed by J Billingsley and then later by J Hollis when J Billingsley left for Australia. DA Sanders was added as a supervisor for the final two and a half years. In addition, Dr Giles Tewkesbury completed his PhD in 1994 after studying design using distributed intelligence within advanced production machinery (PhD directed by DA sanders).

Key findings


- GRASP could be used as a visual programming environment for motion planning through complicated (but mainly static) environments.
- Image modelling for real time manufacturing applications was investigated and a new method using 2-D slices in joint space and simple polyhedra was proposed.
- The novel idea of virtual task machines was introduced and demonstrated.
- The effect of varying the ANN net size on the control of a robot was investigated.
- The automatic programming of production machinery for de-flashing plastic parts was achieved.
- Neural networks were used as both feed forward estimators and as feedback controllers to control machinery (mainly industrial robots).
- The new concept of Influence Analysis was introduced as a technique for identifying design focuses.
- A new framework was created to automatically program advanced machinery.
- That allowed a single graphical resource to program numbers of machines.
- Advanced production machinery was automatically programmed.
- The new concept of Task Rules was demonstrated for de-flashing plastic parts.
- A new robot command library was created that included simulation.
- The new simulation based robot command library was successfully applied to three robots.
- The use of distributed intelligence within advanced production machinery for design applications was successfully demonstrated.
- Suitable materials were identified for use in manufacturing energy cables.
- New methods were proposed to improve design, materials and manufacture.
- Total Quality Manufacture was achieved and demonstrated.
StatusFinished
Effective start/end date7/10/9213/03/96

Collaborative partners

  • University of Portsmouth (lead)
  • Army Air Corps
  • University of Southern Queensland
  • The ClockTrust
  • Cranfield University
  • Chailey Heritage Clinical Services
  • Pirelli
Relations

ID: 9724645