Design And Implementation Of FPGA Edge Computing Based On USB Interface Of Agricultural Intetnet Of Things

With the development of the Internet of Things,big data and cloud computing,the development of smart agriculture is of great significance to improve crop yields and protect the ecological environment.Owing to the world's population increasing and arable land decreasing and deteriorating,the gap in world food reserves is still growing day by day.Therefore,in order to ensuring the green development of the ecological environment,it is particularly important to develop smart agriculture through new science and technology.The collection of agricultural environmental information has a great impact on the agricultural digital farming strategy.Sensor is the sensing device for environmental information acquisition,and there are various sensor interfaces with different functions.The different usages multi-type interfaces increase the difficulties of the installation,operation and maintenance of agricultural technological equipment,and the realization of real-time and online processing of massive environmental data requires"high"cost.In this thesis,the sensor interfaces based on USB architecture,data analysis algorithm and its hardware acceleration are studied,aiming to design and implement a sensor data acquisition and edge calculation system based on FPGA.The Experimental results show that the designed system can stably operate and the processed results based on hardware acceleration are effective and reliable,and can provide real-time data collection and online processing technical support for intelligent agricultural Io T systems.Firstly,I~2C sensor is standardized according to IEEE 1451 standard,That is,to choose the appropriate USB data converter to unify the I~2C sensor interface.The Raspberry Pi development platform is used as the data acquisition module to collect and store the data of various sensors of environmental information.Then,Vivado HLS is used to migrate the wavelet decomposition algorithm for implementing edge computing in Agriculture Io T.After optimizing the hardware algorithm,it is packaged into IP core for Zynq-7000 series development platform to accelerate the wavelet decomposition processing of the original environmental data.Finally,the TCP application program based on Socket is developed to transfer the data of the Raspberry Pi acquisition module to the FPGA edge computing module.Among them,the interface of the upper computer is written using QT.In order to simplify the design and reduce the cost,the data transfer of the upper computer also uses TCP protocol.At the same time,Raspberry Pi development board and FPGA development board are used to access the router to communicate with PC.Vivado HLS uses C/C++language for IP core development owing to the characteristics of strong portability.Doing so can lay a foundation for the subsequent algorithm realization of the system for edge node data prediction module.