Intelligent Balance Vehicle For Tracking Based On Neural Network

With the rapid development of various technologies such as sensing technology,control technology and computer technology,the research on mobile robots is gradually developing into a wider range of application fields.For example,machine robots can independently and flexibly perform corresponding tasks in specific scenarios to achieve some engineering purpose.The two-wheeled balance vehicle studied in this paper is a kind of wheeled robot.The platform is a comprehensive system that integrates functions of environment perception,planning decision,automatic driving,etc,and uses computers,sensors,information,communication,navigation,and automatic control technologies.The technical research based on the intelligent vehicle platform can be applied to many scientific research and practical engineering fields,such as automatic parking,cruise control and other assisted driving and unmanned driving technology in the field of intelligent vehicles,pipeline damage and crack detection in the manufacturing field.In this paper,an intelligent balance vehicle for tracking based on neural network is designed.By summarizing the research status at home and abroad and researching related technology,the basic control principle and characteristics of the intelligent balance vehicle are analyzed,and the overall scheme of the system is proposed,and the hardware,software and algorithms of the system are designed.In terms of hardware,the system uses Freescale's K60 series chip as the main controller,and collects the track information through the OV7620 digital camera.The system's circuit of power supply,motor drive,attitude acquisition module,speed acquisition and other auxiliary debug circuits are designed.In terms of software and algorithms the system's related programs of image acquisition,speed measurement,attitude acquisition and PID control are designed.Based on BP neural network,an adaptive PID controller is designed for speed control,which realizes adaptive adjustment of control parameters,reduces the difficulty of parameter debugging in actual engineering and enhances the robustness of the system.Based on the LeNet-5 model,a track element recognition algorithm based on convolutional neural network is designed.The algorithm which assists the image acquisition and processing program effectively reduces the difficulty of extracting track features from image data,and the misjudgment of the intelligent vehicle in the tracing driving.Finally,the simulation and test of the intelligent vehicle program and related algorithms are carried out.The results show that the neural network-based tracing racing balance car is feasible and has better control and operation effects.