Measurement Error Analysis And Non-linear Compensation Methods Of Temperature And Humidity Sensor

In the observations of actual business in AWS,the measurement accuracy of HMP-45D temperature and humidity sensors is susceptible to the non-target environmental which leads to the actual measurement error large.Especially,the measured data in high temperature and humidity environments is greater.Therefore,we should take measures to compensate the data error which is caused by the temperature and humidity sensors.We analyze all the influence of environmental parameters on sensors to study which parameter is the main factor,and then do experiments to verify our theory according to the development of precision instrumentation standards experimental program.Then,in order to compensate the main factor for the humidity sensors,we use the improved GA to optimize the SVM algorithm.At the same time,we take the way to compensate the humidity of the temperature sensor by using the Cuckoo Search algorithm to optimize the BPNN.Study on the influence factors of temperature and humidity sensors.From the analysis of measurement data,the main impact factors of environmental parameters were obtained that the temperature has a great influence on humidity sensors of the humidity sensors while humidity has a great influence on the temperature sensors.Pointing at the influence environmental parameters in the actual measurement of temperature and humidity sensors,identify the main influencing factors by theoretical analysis.And we make a detailed experimental program to confirm the theoretical analysis.The compensation model of the temperature influence on humidity sensor,which is based on the experimental data measured by humidity sensor,uses the improved GA-SVM model to compensate the temperature effects on humidity measured by humidity sensor.The fitness function in GA is reconstructed and the operations of selection.Crossover and mutation are optimized to improve the GA,and then use the improved GA to optimize the parameters of SVM.Based on the multiple sets of experimental data under different temperature and humidity data,we used this improved GA-SVM to compensate the temperature for humidity sensor and the results were compared with SVM.The experimental results show that the improved GA-SVM can compensate the effects of temperature and improve the accuracy.The compensation model of the humidity influence on temperature sensor,which we establish the CS-BP model to compensate the effects.The CS algorithm is used to estimate the training error of the BP neural network as a fitness function,and its weights and thresholds are optimized.The CS algorithm is established to optimize the weights and thresholds by using training error of BPNN as fitness function,and then use the CS-BP model to compensate the measured data of the temperature sensor which is influenced by humidity.And in order to improve the randomness of CS,make the algorithm jump out of local extremity and improve the performance of CS-BP,we introduce adaptive step-size into the process of cuckoo searching for nests.The established compensation models compensate for the nonlinear error of the disturbance factor better and improve the accuracy of the observation data.