Force control in robotic assembly under extreme uncertainty using ANN

Robotic assembly operations can be performed by specifying an exact model of the operation. However, the uncertainties involved during assembly make it difficult to conceive such a model In these cases, the use of a connectionist model may be advantageous. In this paper, the design of a robotic cell based on the adaptive resonance theory artificial neural network and a PC host-slave architecture that overcame these uncertainties is presented. Different sources of uncertainty under real conditions are identified and their contribution in a typical assembly operation evaluated. The robotic system is implemented using a PUMA 761 industrial robot with six degrees of freedom (DOF) and a force/torque (F/T) sensor attached to its wrist which conveys force information to the neural network controller (NNC). Results during assembly operations are presented which validate the approach. Furthermore, the method is generic and can be implemented onto other manipulators