In this paper, a robot assembly wrist, which is able to assemble chamferless parts, has been developed. The RCC structure is used as a basic structure, 5 position sensors and 4 pneumatic actuators are intalled additionally to measure the deformation of RCC structure and correct the errors actively. Due to the restricted direction of actuation, a decision rule which selects the suitable actuator according to the position seneor signal is needed. for this purpose, a neural network is used and it is experomentally shown that neural network overcome systems nonlinearity. This paper presents fundamental experiment results for the inspection of parts with several clearance.