抄録
In this paper, we considered an assessment function for effective optical elements placement of the Mesh of proximity sensor and searched a simulation result which maximizes the assessment function. Output characteristic of this sensor depends on relative placement of the optical elements. Since the placement has large amount of freedom, designing of the sensor which meets desired specification is difficult even using the optical simulator. Thus we proposed the simulator-supported method for effective element placement using Genetic Algorithm, and attempted to improve the sensor output characteristic over applying this method to the sensor design. In this article, we applied this method to the fingertip sensor which has unidirectional curvature, and confirmed the effect.
