There are a number of problems associated with using robots for automation, but the fundamental issue is cost. If robots were cheaper to buy, install, maintain and use, they would provide a cost-effective solution for a wider range of applications. This work surveys a range of previous work in robotics which has particular relevance to reducing the cost of robotic automation. Robotics is a diverse discipline, and the literature survey brings together research from a variety of areas often considered in isolation, including mechanical design, sensory feedback and control. In particular, visual feedback enables tasks to be monitored as they are executed, thereby removing uncertainty. In the past, most vision systems have been used in conjunction with standard (i.e. expensive) industrial robots. This work proposes the use of visual feedback to compensate for the inaccuracies inherent in a cheaply constructed robot.To speed up the visual feedback loop, a new image processing algorithm has been developed. Partial summation allows a number of conventional image processing techniques to be implemented very efficiently, which enables a standard computer to be used without introducing excessive delays. The associated vision system hardware is currently also available at a low cost. It seems eminently sensible to shift the emphasis in a robotic system away from costly, high-specification mechanics, to more sophisticated control systems, when the cost of the latter continues to fall.A real application, automatic PCB manufacture, has been addressed by this work. This is currently only practical for large volumes, which ensure high machine utilization. In response to the needs of a low-volume producer, who currently relies on manual labour, very low-cost prototype PCB drilling and surface mount component placement machines have been built and tested. Results support the efficacy of the approach proposed in this work, and suggest that further improvements in performance may be possible with the use of a learning controller.
|Institution||Cambridge University Engineering Department|
|Address||Trumpington Street, Cambridge, UK|