中国科学院自动化研究所机构知识库

In previous work, we demonstrated the creation of bespoke tactile elements on 3-D surfaces through the “Embedding Soft material into Structure ENabling Tactile sensing” (ESSENT) approach. Each tactile sensing channel is filled with soft material, transforming applied force at one end of the channel into a microdeformation at the other end of the channel, measured by means of emitting light and sensing the intensity of the reflected light. However, due to the large number of design parameters of the sensing channel structure, it is difficult to predict the optimal parameters for the desired tactile sensing specifications. In such case, consistent performance among multiple tactile sensing channels requires a tedious trial-and-error approach. Such problem deteriorates rapidly with the increase in the number of tactile sensing elements, hindering the widespread adoption of ESSENT technology toward high-density arrays and complex surface layouts. To address this challenge, this article presents a multiobjective optimization method for designing the parameters of ESSENT sensing channels. Theoretical predictions and experimental results show that by identifying the channel design parameters, sensing consistency can be significantly improved. Describing by mean square deviation, the value based on the best optimization result (0.0394) is seven times better than nonoptimized sensing channels (0.2865). Moreover, by calculating the sensitivity of different channels in an array sensor, we found that the variance of sensitivity across sensing channels after optimization is one order of magnitude smaller than that without optimization.