Wei Chen and Yuting Shang. Neural Network Based Sensor Dynamic System Compensation Algorithm and DSP Implementation. Dynamic Systems and Applications 30 (2021) No.5, 735-752

https://doi.org/10.46719/dsa20213055

ABSTRACT.
Due to the continuous progress of science and technology, the stability of the sensor is required by the detection system, so the research on the dynamic system of the sensor has been paid more and more attention. At present, there are many research results on sensor dynamic compensation algorithm, but its compensation effect is not very ideal. To put forward an effective dynamic compensation algorithm and improve the dynamic characteristics of the sensor, this study is based on BP neural network, carries out forward fitting of parameters, and uses a genetic algorithm to continuously optimize and select the best value. After analyzing and mastering the dynamic compensation principle, the network model and excitation signal are used to build the sensor dynamic compensation model, and the compensation filter is designed and simulated. To realize real-time compensation, DSP with powerful data processing function is selected as the experimental platform in this paper, with the TMS320 processor as the core, to process data signals and finally complete the effect verification of sensor dynamic compensation. The experimental results show that the compensation algorithm of the sensor dynamic system based on neural network and DSP can complete the real-time dynamic compensation, which is beneficial to improve the dynamic characteristics of the sensor and maintain the stable performance of the sensor.

KEYWORDS: Neural Network, Dynamic Compensation, Digital Signal Processor, Sensor Dynamic Characteristics, Dynamic Characteristics