Subliminal calibration for machine operation with prediction based filtering

This paper addresses a skill assist technique without awareness. Generally, human assist system is by adding autonomous input to the operation command as force from obstacles. Although these assists are suitable in a particular task, they may bring about hindering human learning process. This paper focused on the human learning to utilize for suitable human-machine systems. The assist is carried out without human awareness, namely “subliminal”, and it does not hinder the operator´s learning ability. Furthur, since the technique is based on prediction of operator´s input command, subliminal filtering method is applied to improve its prediction accuracy. Then, in spite that the experimental results show the proposed technique make NOT the operators aware of, the operation performance is improved especially on low skilled operators.