An Investigation into a Combined Visual Servoing and Vision-Based Navigation System Robot for the Aerospace Manufacturing Industry

High-Value manufacturing, such as aerospace manufacturing, has been less impacted by the mainstream use of robotic automation compared to other manufacturing industries. This is due to the cost factor required when creating robotic systems which can successfully interact with such high tolerance work�pieces. This research aims to investigate the gap of robotics within high-value aerospace manufacturing, with the goal of creating a generic robotic algorithm which can effectively and optimally detect and trace a variety of aerostructure inspired workpieces.

This goal was achieved by firstly developing a vision system for detecting and tracing particular features of partially-known workpieces. These workpieces var�ied in size and spatial profile, having both obtuse and acute edges. Once an effective vision system was developed, a variety of distribution-of-labour algorithms were developed, with the aim of dividing the task of tracing a work�piece between the kinematic arm and mobile base. The results showed that different distribution-of-labour algorithms performed differently, depending on the type of detected feature, specifically how vertically inclined the feature was.

These results were used to develop an optimal distribution-of-labour algorithm, which could dynamically and optimally switch between different distribution-of-labour systems, to trace a workpiece both quickly and accurately. Results showed that an optimal distribution-of-labour algorithm decreased tracing time and increased accuracy in realistic aerostructure-inspired workpieces compared to just using one major algorithm, and could dynamically trace workpieces regardless of previous knowledge or spatial profile.