Research And Design Of Travel Mode Recognition System Based On Smart Phone

With the development of science and technology,mobile Internet appear more and more frequently in people's field of vision.All kinds of applications related to the Internet have a huge commercial value.Travel mode recognition is brought forward with the development of intelligent technology or social services,and has a broad application prospect.Nowadays,the smart phone is no longer a communication tool that only makes phone calls and sends text messages.It is more like a microcomputer with strong operation perception ability and large storage space.It can almost meet the demand of all basic necessities of life.In the study of travel mode recognition based on smart phones,the limited data sources restricts the performance of the algorithm.In addition,the adaptability of the algorithm in different scenarios varies considerably.These limitations impede the improvement of travel mode recognition accuracy and need to be solved one by one.This paper uses an improved random forest algorithm and a smart phone based travel mode recognition system,which mainly includes the following contents: this paper introduces a variety of data to the existing travel mode recognition algorithm,and the effects of various data on the accuracy of the recognition system are analyzed through experiments.By setting parameters to compare and analyze the performance differences of different algorithms,this paper studies the impact of different algorithms on the accuracy of travel mode recognition,and has carried out some analysis and Research on the random forest algorithm.Compared with other algorithms,it is more suitable for travel mode recognition if the problem of unbalanced classification is overcome.This paper deals with the lack of negative samples by artificial increase in an innovative way.This paper has completed the design and implementation of a complete system and provided a complete set of design ideas on data acquisition,feature extraction and identification model construction.This paper creates six core modules: system interaction module,data acquisition module,storage module,main control module,feature extraction module and mode discrimination module.The seven specific travel modes involved in this article include state of rest,walking,running,bicycles,buses,self driving and subway.The test results show that the precision of the algorithm can achieve the desired effect,which is good application value and practical significance.