This was a second project following on from a PCFC Project to investigate new ways of modelling in real time.
The project investigated robot spot welding processes and the use of knowledge based systems and vision systems with them. As part of that, a new system was created that could program several machines from a single graphical resource.
New real-time geometric modelling methods were introduced using models in an actuator space and cartesian space to representing objects and the environment in a computer. That led to some work on perception in robotics and the demonstration of a spot welding process using knowledge based systems.
That all led to an automatic system for simple spot welding tasks and recognition using artificial neural networks and Fourier descriptors.
Knowledge based systems were demonstrated being used for spot welding.
A new robot vision system was created.
More than one machine was programmed from a single graphical resource.
Real-time geometric modelling was achieved using models in an actuator space and cartesian space.
Obstacles and the environment were efficiently represented in a computer.
Perception in robotics was investigated.
An automatic system was created for simple spot welding tasks.
Recognition was achieved using artificial neural networks and Fourier descriptors.