Research On The Building Of Flexible Fabric Pressure Sensing System And Data Preprocessing Method

With China’s entry into an aging society,the proportion of the elderly living alone in the total population is growing.However,there are few pension related products for the elderly living alone,especially for the elderly living alone who fall down unexpectedly,which leads to missing the best treatment time.Therefore,a flexible pressure sensing array and a pressure distribution detection system,referred to as flexible pressure sensing system,are designed for this background.The system is embedded in the home environment of the elderly living alone to collect the plantar pressure data of the elderly living alone in real time.Finally,the collected data are preprocessed for outliers and crosstalk,so as to facilitate the subsequent analysis of the physical state of the elderly.Firstly,the research status of flexible pressure sensor and its measurement system at home and abroad is studied,including the structure of flexible pressure sensor,preparation process,application scenarios and the crosstalk problem of sensor array.Then,a kind of mesh flexible pressure sensor array with fabric as the substrate and single layer structure is proposed,and the process of preparing flexible pressure sensor array is studied,including the selection of conductive materials,polymer matrix materials and organic solvents,as well as the specific process of preparing conductive composite solutions.In this experiment,conductive carbon black is used as conductive material,polyurethane as polymer matrix material,N,N-dimethylformamide as organic solvent,and flexible pressure sensor array is prepared by screen printing.The conductive principle of flexible pressure sensing unit is further explored,mainly involving the percolation theory and tunnel current effect when carbon black is mixed in polymer materials.Then the experiments are carried out on the fabricated flexible pressure sensor array,and the piezoresistive characteristics,sensitivity,pressure resistance response range and stability of the four key indicators are tested and analyzed.The test results show that the flexible pressure sensing unit has good piezoresistive characteristics under the conditions of 0-0.3 N pressurization and 0.3-0 N depressurization.Then,the piezoresistive characteristic curve is fitted and derivative is obtained.The smaller the pressure is,the higher the sensitivity is.On the contrary,the greater the pressure is,the lower the sensitivity is.Calculate the corresponding sensitivity values when the pressure is 0,11.6 and 25.0 kPa under the condition of pressurization and depressurization respectively.The maximum sensitivity is 263.57kPa^-1 and the minimum sensitivity is 9.50 kPa^-1.The pressure resistance response range of the flexible pressure sensing unit is that when the pressure gradually increases from 0kPa to 25 kPa,the resistance value of the sensing unit rapidly decreases from about 2500Ω to the lowest response resistance around 500Ω and then remains unchanged.For the stability of the flexible pressure sensing unit,under the cyclic action of 16 periodic alternating pressures of 0-0.3-0 N,its resistance always changes with the change of pressure,and it maintains good stability on the whole.The corresponding pressure distribution detection system is developed for the prepared flexible pressure sensor array.In order to meet the functional requirements of real-time data acquisition and wireless transmission,the PCB design,hardware design,circuit principle and control program design are studied.The electrical connection between the hard circuit board and the flexible fabric is designed.Under the condition of ensuring the stability of the circuit connection,the circuit board can also be easily disassembled.In terms of control,Arduino UNO development board was selected as the control module,ADG731 as the multiplexer,and NodeMCU as the wireless communication module for the pressure distribution detection system,and the data acquisition program and wireless transmission program were respectively written for the system to address and measure the resistance of the sensor units in each row and column of the flexible pressure sensor array,and the collected resistance value was wirelessly transmitted through NodeMCU,The maximum transmission rate is 150Mbps,which meets the speed requirements of real-time transmission.When the flexible pressure sensor array has a large area,there will be some bad sensing units and crosstalk,which will lead to inaccurate data collected from the pressure distribution detection system.Due to the problem of abnormal data caused by bad sensor units,this paper uses the median filtering algorithm to process the pressure cloud with abnormal values,and calculates the value of the peak signal to noise ratio（PSNR）of the pressure cloud before and after processing as the standard to evaluate the effect of the algorithm.According to the calculation results,the PSNR of the processed pressure nephogram is between 30 and 40 d B,and the image quality is good.Aiming at the crosstalk phenomenon,the method based on U-net convolution neural network is used to process the collected pressure nephogram.Under the same pressure state,the pressure nephogram collected from the flexible pressure sensor array is used as the input of the algorithm model,and the pressure nephogram collected from the commercial pressure pad without crosstalk is used as the output to train the input and output image regression prediction model.The root mean square error（RMSE）is used as the standard to evaluate the training results of the algorithm.The calculation shows that the final root mean square error RMSE of the algorithm after training is 7.108,and from the practical effect,it meets the standard of eliminating crosstalk.