A Novel Low-Cost Detection Scheme Based on Multi-Sensor Array Optimization

In this paper, a novel low-cost odor detection scheme for tea grade identification is proposed, including a negative pressure odor information acquisition module with a low-cost sensor array and the corresponding conditioning circuits. To solve the redundant information brought by the broad-spectrum response characteristics of the sensors and the mutual interference of the hardware circuits and further reduce the manufacturing cost, a new array optimization algorithm based on independent classification information with low class-related redundancy (ICLCR) is developed, which further clarifies the independent classification information and the redundant information contained in the features. To verify the effectiveness of the designed system and the array optimization scheme, the typical odor detection system, i.e, PEN3, and the existing array optimization algorithms are exploited for comparison. The classification results demonstrate the effectiveness of the designed system and the superior performance of the designed array optimization algorithm compared to existing array optimization algorithms.