Design And Implementation Of Temperature Sensing System Based On Smartphone Acoustic Sensors

Environmental temperature monitoring plays an important role in agriculture,meteorology,smart home and other fields,such as automatic temperature control of greenhouse,weather forecast,intelligent regulation of central air conditioning,etc.Although the temperature measurement equipment is simple to use and low cost,most of them do not have network function,so it is difficult to carry out large-scale distributed temperature monitoring.In recent years,the number of smart phones is growing rapidly.If we can use smart phones to measure the ambient temperature,we can get large-scale temperature data with low cost.However,the temperature sensors inside the smart phone are used to monitor the running temperature of the device,which can't measure the ambient temperature;with the development of mobile computing,there are innovative applications based on smart phone sensors,such as using camera for visible light communication,using Inertial measurement unit(IMU)for inertial navigation and son on.These innovative applications bring an opportunity to environmental temperature measurement based on smart phones.In this context,this paper first uses the smart phone acoustic sensor to realize the ambient temperature measurement.Besides,innovative algorithms are also proposed to solve the background noise and multipath effect.The main contributions of this paper are as follows:1.In this paper,high-precision ambient temperature measurement is realized on the smart phone for the first time.The median error of the system is 0.3?.It has strong antinoise and anti-multipath interference ability.Combined with GPS and multi-source information fusion indoor positioning technology,large-scale indoor/outdoor ambient temperature monitoring can be realized,which completely changes the current large-scale distributed temperature monitoring.2.This paper innovatively proposes a mask based signal denoising algorithm.According to the linear time-frequency characteristics of chirp signal,the original signal is fitted by RANSAC to generate a denoising mask,and an improved spectral subtraction method is introduced to suppress the residual noise.This denoising algorithm improves the anti-noise performance of the system.The experimental results show that the system based on this algorithm can achieve a median error of 0.69?under the interference of 70 d B background noise.3.This paper innovatively proposes a multipath processing algorithm,which based on structure-borne signal and air-borne signal.According to the propagation speed and energy difference of structure-borne signal and air-borne signal,the probability that the energy peak in the mixed signal spectrum is generated by the line of sight path is modeled,which effectively avoids the interference of coherent multipath on the extraction of the main peak of the mixed spectrum.It is suitable for various multipath environments and improves the anti-multipath interference ability of the system.In addition,the system solves multipath solution by reducing the signal strength and averaging the adjacent data frames.The experimental results show that the system has high measurement accuracy in different multipath environments.