Design Study Of A Multi-Sensor Fusion Physics Experiment Demonstration System

Physics experiment demonstration system can help teachers to show students the physics experiment process,the physics experiment phenomenon system.With the digital transformation of education and teaching,the digital transformation of traditional physics experiment demonstration system has become the trend,but the existing improved system generally has such defects as single sensor,complicated equipment,high cost,poor data accuracy and lack of compatibility.To address these shortcomings and the requirements of the new curriculum reform for teachers’ demonstration teaching,this paper designs and implements a new physics experiment demonstration system using low-cost Micro-Electro-Mechanical Systems(MEMS)sensors and multi-sensor fusion technology.The key issues in the design of the demonstration system are how to fuse the multi-sensor data and how to effectively remove the noise from the sensors.The specific research contents are as follows:(1)A certain initialization strategy is analyzed to address the initialization problem of the Kalman filter algorithm used for fusing accelerometer sensor data and gyroscope sensor data in MEMS inertial sensors.At the same time,a combination of Exponential Moving Average(EMA)and Kalman filtering is proposed for denoising to improve the effect of initialization parameters.By evaluating the denoising results of different algorithms,the root mean square error(RMSE)of the improved Kalman filtering algorithm is improved by no less than 6.6%compared with other algorithms.Also,the denoising results of the improved Kalman filter algorithm are compared with the traditional Kalman filter algorithm and the complementary filter algorithm in practical applications.The results show that the improved Kalman filtering algorithm has a significant noise removal effect,and the improved Kalman filtering algorithm can track the actual data better than other algorithms.(2)To address the problem of noise in the audio signal captured by the MEMS silicon mike sensor.Firstly,wavelet threshold denoising algorithm is used to pre-process the audio data.The wavelet threshold denoising algorithm can produce pseudo-Gibbs phenomenon due to the threshold function,which has a large impact on the denoising results.An improved adaptive weighting algorithm is proposed to fuse the multi-sensor data to suppress the effect of this phenomenon.By analyzing the denoising results,the improved adaptive weighting algorithm improves the Signal to Noise Ratio(SNR)by about 12.9%,the Normalized Cross-Correlation(NCC)by about 6‰ and the Mean Squared Error(MSE)by about 1.5% compared with the traditional wavelet threshold denoising algorithm.Squared Error(MSE)is improved by about54.1%.In addition,the stability of the improved adaptive weighting algorithm is better than other algorithms when the noise of the sensors is inconsistent in the simulated real environment.(3)Experiments are conducted using the designed experimental demonstration system,and eventually,with the application of improved algorithms,the experimental system can accurately acquire experimental data,visualize the experimental process and physical laws.The research results of this paper have been applied to an educational technology company in Ningbo,and the feasibility and reliability of the system have been verified by the company and user feedback.In addition,the system provides new ideas for informatization experimental demonstration equipment into the classroom,provides new possibilities for teachers to use informatization tools for physics experimental demonstration,and also has some reference value for the improvement of experimental demonstration teaching aids.