Research On Low-Power Vehicle Detection Method Based On Multi-source Fusion Technology With Time-Domain Sparse Signals

The intelligent parking system is an important part of smart city construction,which is used to manage existing parking space resources to alleviate the problem of urban parking difficulties.To provide data support for intelligent parking system,the magnetometer-based wireless vehicle detectors(WVDs)are widely used in urban parking space occupancy information collection.However,the geomagnetic sensor is susceptible to adjacent vehicles,temperature,and strong magnetic interference,resulting in a decrease in the accuracy of the WVDs.Moreover,since the WVDs powered by the non-rechargeable batteries,the energy limitation results in a short life cycle.Therefore,the overall performance of the WVDs cannot satisfy the intelligent parking system.In order to improve the overall performance of the WVDs,a low-power vehicle detection method based on multi-source fusion technology is proposed in this dissertation,which fuses the signature of magnetic signals and radio waves attenuation and vehicle vibrations to achieve vehicle detection.The main work of this dissertation is summarized as follows:(1)Firstly,a temperature compensation algorithm is proposed to correct the temperature drift of the geomagnetic data.Then,a geomagnetic multi-characteristics based vehicle detection algorithm(GMCBA),which combines signatures such as the similarity and amplitude of geomagnetic signals is proposed.Therefore,the robustness of vehicle detection algorithm is improved due to the introduction of geomagnetic multi-dimensional information fusion technology.(2)According to Electromagnetic Theory,the vehicle in the wireless environment can attenuate the energy of radio waves.Hence,the attenuation signature of radio waves is introduced into aforementioned GMCBA,and then a multi-source heterogeneous information fusion vehicle detection algorithm(MSHIFA)combing the data feature of received signal strengths(RSSs)and that of magnetic signals based on Fuzzy theory is proposed.Therefore,the vehicle detection accuracy can be improved due to the introduction of heterogeneous data.(3)To reduce the energy loss of redundant geomagnetic data collection when the parking space is occupied or idle,a self-powered geophone and a low-power circuit with the vibration event extraction are introduced into the WVDs.Then a micropower multi-source heterogeneous information fusion vehicle detection algorithm(μp MSHIFA)is proposed,in which the vibration event excited by the vehicles is used to wake up the aforementioned MSHIFA to achieve vehicle detection.Therefore,energy consumption can be optimized by reducing the acquisition frequency of redundant geomagnetic data.Finally,four validation scenarios were built to evaluate the detection accuracy and power consumption of the vehicle detection method proposed in this dissertation.The experiments show that the detection accuracy of the proposed method is up to 98.8%,which is 13.4% higher than that of existing algorithms.Moreover,under the condition that the parking space turnover rate reaches 30 vehicles,the average energy consumption of the proposed method is as low as 17μA,which is only 10.6% of that of existing algorithms.Therefore,the research has high theoretical significance and scientific value.