Design And Implementation Of Multi-bus Acquisition Embedded Real-Time System

With the rapid development of the Internet of Things and the constant new demands from all walks of life,the functions that IoT terminals need to implement are not only growing but also more and more complex.However,at present,the collection function of the IoT terminal that belong to data collection category is single,resulting in increased repetitive work and low efficiency in the terminal development and production process.So,for the IoT terminal of data collection,it is very urgent to develop a complete,efficient,scalable,and real-time,multi-bus acquisition embedded real-time system that is applicable to different acquisition scenarios.It also has great significance with improving the economic efficiency of enterprises and promoting the rapid development of the national Internet of Things industry.For the system hardware design,this paper divides the hardware into four modules: MCU module,acquisition module,external storage module and communication module,and other interfaces are reserved to facilitate expansion.In this paper,the system software design scheme adopts the object-oriented design idea,aiming at universality and practicality.The system software is divided into: data acquisition layer,task control layer,data link layer,protocol layer,communication layer.And layers are connected through common interfaces,avoiding deep coupling and facilitating development and expansion.In order to transplant the system to the IoT terminal based on STM32F4 series chip,this paper plans the system space,designs the bootloader.For the multi-bus acquisition part of the system,this paper introduces the USART,RS232,RS485,CAN,IIC bus and related peripheral circuit schematic design.Then the implementation of some data acquisition tasks is introduced,including the BD/GPS positioning task and the GPRS state machine task base on the USART interface,the UPS data acquisition task on the RS232 bus,the MODBUS protocol task on the RS485 bus,and the CAN bus protocol task and IC card operation task on the IIC bus.In addition,this paper also designs a multi-bus strategy,in the face of the environment with multiple acquisition methods,multiple communication protocols and a variety of sensor,a configurable way is given,which expands the scope of use of the terminal and reduces the repetitive work in development and production.In order to solve the problem of packet loss in the multi-bus acquisition process,this paper constructs a BP neural network model using artificial intelligence technology,uses offline training style to train a neural network that could predictive multi-bus interrupt priority scheme.Then use the test set to test,it is concluded that prediction of multi-bus interrupt priority scheme using BP neural network has certain accuracy,and it has certain use value but still needs improvement.Finally,a physical map of an IoT terminal with multi-bus acquisition embedded realtime system is given,then test the integrated system,introduce the function of bus policy configuration,and give an application example.