Indoor Parking Positioning System Based On Machine Learning

China's automotive data surge in recent years,leading to a parking space becomes scarce resources.Parking rational management can improve the utilization and reduce costs.In response to this demand,the paper launched a study of indoor parking positioning systems and algorithms.Positioning technology is now widely used is the GPS system,however,GPS indoors impossible to achieve reliable positioning.Therefore,a new system must be designed to achieve positioning.Indoor positioning respectively from hardware and algorithm research at home and abroad.At the same time there are a lot of research in this area.By comparing the different programs and the actual test,mainly in the hardware layer uses Zigbee wireless network system and algorithm layer selection is now more popular in machine learning methods,which from probability statistical methods to start positioning judgment.The main work in this system focus on the choice of hardware,develop protocols,implement the logic layer code,the selection of positioning system model and algorithm realization.Hardware selection is made by way of comparison and finally select Zigbee wireless network system as the primary hardware infrastructure.Protocol help achieve the machine to understand the data,which make judgments and response for different data.Application logic code is C language and design of communication flow based on Zigbee protocol stack.While designed socket network communications and a server under linux.Algorithm design is the core of this paper,which early on data collection and later analysis of data and training.Finally,established the hypothesis space and used statistical in cross-validation approach to find the optimal model.Reliability of the model is verified.Stability of the system and the system of regional positioning accuracy obtained by simulation and testing.Finally,proved the reliability of the system and the wide applicability.