非处方助听产品造型批量定制设计方法

Mass customization design method for over-the-counter hearing aids

依据概率神经网络算法构建耳甲腔形状识别模型,在满足用户佩戴非处方助听产品舒适性的基础上,实现该产品造型的批量定制设计。构建以耳甲腔曲面型值点为输入层,耳甲腔形状类别为输出层,径向基函数为模式层的耳甲腔形状识别模型。通过曲面偏差分析和有限元分析,分别对识别结果的可靠性进行验证。通过训练和测试得到模型的识别精度为94.3%。偏差分析结果显示:新样本耳甲腔曲面与其识别组曲面间的最大差值均小于1.9 mm。当新样本佩戴以识别结果为依据而设计的非处方助听产品造型时,接触压力云图分布均匀,且最大压力值仅为0.22 MPa。结果表明:以耳甲腔形状差异为视角,进行识别与批量定制设计的非处方助听产品造型具备较高的佩戴舒适性。

A shape recognition model of auricular concha was constructed according to probabilistic neural network algorithm to realize mass customization modeling design of this product based on user comfort of over-the-counter hearing aids products. The conchal shape recognition model was constructed with the shape points of auricular concha as input layer, the shape category of auricular concha as outer layer and the radial basis function as pattern layer. Curved surface deviation analysis and finite element simulation analysis were performed reliability validation of recognition results. The training and test results showed that model recognition accuracy was 94.3%. The deviation analysis showed that the maximum deviation between conchal shapes of new samples and the shapes in their cluster were all less than 1.9 mm. When the new samples wear the OTC hearing aids designed based on the recognition results, the stress distribution was uniform, and the maximum pressure was only 0.22 MPa. The OTC hearing aids product modeling mass customization was in high comfort recognized with the differences of auricular concha.