The aim of ANNIE was to find out which of several generic problem areas, such as pattern recognition, sensor fusion, and adaptive control and optimisation, are best approached using neural networks.
Application of neural networks for industry in Europe (ANNIE) aimed to find out which general problem areas are best approached using neural networks. Applications in pattern recognition, autonomous vehicle control and optimisation of airline crew scheduling were developed.
Specific application areas have now grown in conditioning monitoring of complex mechnisms by neural network analysis of conventional instrument data. Benefits include the following: earlier warning of failures; analysis of human signatures, leading to a socially acceptable form of automatic identification (biometrics) through automatic signature verification; analysis of finanacial transaction data to identify target classes of behaviour to allow simple quick analysis of customer preference.
A sensor device has been designed and constructed which generates 3-dimensional images of surface structures (eg after honing processes) and stores the data in a fast image memory for further automatic data evaluation.
The aim of the project was to find out which of several generic problem areas, such as pattern recognition, sensor fusion, and adaptive control and optimization, are best approached using neural networks.
A survey on the systematics and capabilities of neural networks, their architecture and applications and hardware available for neural network simulation has been performed. Areas of applications where neural networks might complete or out perform conventional techniques have been selected: image processing and inspection, robotics, optimization and technology transfer.
With respect to the inspection work, 2 problem areas related to nondestructive testing were chosen, visual inspection of solder joints and ultrasonic testing of pressure vessels. The work stimulated a large internal programme on graph matching, and now a package called MatchFinder, which enables automatic matching and comparison of gas liquid chromatography (GLC), is being marketed.
Neural networks can be of considerable help in the visualization and analysis of multivariate data. Further work has been carried out on NETVision (renamed RENDER) for use as a tool for the pharmaceutical industry to help analyse structure activity relationships in drug design.
The robotics work developed a prototype of an autonomous roving vehicle. This work has been found very useful in approaching the problem of collision avoidance.
Neural networks have also been found to be an ideal means of monitoring the oscillations of flexible structures and controlling them using a damping system based around a neural network.
Work on optimization involved an application for airline end user and crew scheduling problems. A first prototype has been produced, and the follow up work planned aims to produce a crew scheduling product. Applications in banking and finance are also planned to exploit the project's results.
Technology transfer has involved the production of a handbook and a neurocomputers and hardware bo ok should be published soon.
To do this, the partners prototyped neural network solutions in simulated problems chosen from automatic inspection, condition monitoring, robotics, control and scheduling.
WD2 4JP Watford
WC2N 5JT London
60311 Frankfurt Am Main