Design Of Wearable Cattle Positioning And Behavior Detection Device

The significant increase in the number of cattle in grassland pastures has increased the workload of grassland grazing.It is time-consuming and labor-intensive to manually complete grassland cattle breeding,rational grazing,disease prevention and cattle loss prevention.It is urgent to use automation and information technology to complete these tasks.Work.With the rapid development of the Internet of Things technology in recent years,some wearable smart devices have also been used in the pastures of our country,but most of them are imported from foreign countries,and the cost is relatively high.Therefore,this thesis studies and designs a wearable cattle positioning and behavior detection device,which reduces the input of ranch staff,improves the efficiency of ranch operation,and reduces work costs.First,the wearable devices used to monitor the physiological condition of cattle are analyzed,including detection devices based on cattle sound,detection devices based on cattle body temperature,detection devices based on pressure,and detection devices based on accelerometers and GPS.After comprehensive analysis,location information was selected to monitor grazing intensity and prevent loss of grassland cattle,and sound information was selected as a detection method for grassland cattle behavior classification.Secondly,the positioning and behavior detection device of grassland cattle is designed.The low-power L80-R positioning module is used to collect the cow’s position information and the MAX4466 sound amplifier is used to collect the cow’s voice information,which is processed by the lowpower processor STM32L151 and sent to the gateway through the Lo Ra wireless wide-area low-power network.The gateway then forwards the data to the server,and the ranch staff can view it through the client.Finally,the voice recognition and classification algorithm of grassland cattle is designed,and the collected voice is analyzed by Python program.The sound is first denoised,including pre-emphasis,endpoint detection,framing and windowing.Then,the frequency domain information is analyzed through the fast Fourier transform,and the characteristics of cattle calls are found out by comparing other sounds that may exist in grassland grazing.Finally,it is obtained that the fundamental frequency of prairie cattle calls is around 500 Hz,most of the calls last for more than 0.8s,and the sound energy is relatively large.Some calls have an octave relationship in the frequency domain curve,so as to distinguish other sound signals,and distinguish cattle according to the frequency of vocalization.In starvation or calf hunting.According to the spectrogram,grassland cattle will intermittently appear high-energy segments when grazing,and the high-energy segments last for about 0.15 s.According to this feature,they are different from other sound signals.The device studied in this thesis has been tested for positioning accuracy,main power consumption module power consumption,and sound detection accuracy in the laboratory.The test results show that the positioning error of the device is about 5 meters,which can meet the requirements of preventing cattle loss and grasping grazing intensity.Information;the main power consumption module of the collar node has low power consumption,which can meet the needs of not replacing the battery for a long time;the sound frequency is at 1500 Hz,and the error does not exceed 0.65%.The designed device can monitor the position,sound and acceleration information of cattle in real time,accurately and effectively,and meets the needs of low power consumption and wearability,and is expected to be widely used in animal husbandry.