Font Size: a A A

Pedestrian Detection Method And System Design For Intelligent Networ

Posted on:2022-11-17Degree:MasterType:Thesis
Country:ChinaCandidate:M X LvFull Text:PDF
GTID:2532307070452094Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
Intelligent network vehicle road coordination is an important direction for the future development of urban transportation,Its network communication technology includes V2V(vehicle vehicle),V2P(vehicle pedestrian)and V2I(vehicle infrastructure),etc.with the development of 5g technology,the intelligent network vehicle road coordination system can gradually meet the communication requirements of low delay and high reliability,and can be applied to a variety of scenarios.It can effectively improve traffic efficiency and provide convenient and timely information services for traffic travel.In the construction of smart city,pedestrian traffic accidents occur frequently,and pedestrian safety is valued by all parties Pedestrian detection is particularly important.This study establishes a new system to detect pedestrians through millimeter wave radar,realize real-time transmission with intelligent network technology,early warning pedestrian safety,improve pedestrian traffic safety,and promote the construction level of smart city.Pedestrian detection plays an important role in ensuring the safety of industrial production,traffic and life.Among them,pedestrian crossing safety is an important application scenario.Due to the complexity of road traffic,sometimes even if the driver pays attention,it is difficult to predict the sudden emergence of pedestrians in the blind area of the intersection,resulting in traffic accidents.At present,the methods of real-time early warning for drivers are limited,and it is difficult to realize the early detection,identification,tracking and early warning of pedestrians crossing the street.The system architecture constructed in this study is different from the traditional architecture.It combines artificial intelligence technology and Internet of things,and entrusts the computing work to the edge computing module to obtain better system performance.In addition,the designed roadside unit RSU(road site unit)can store,process and upload the data of the edge computing module at the intersection in real time,rather than directly transmit it to the cloud server,so as to improve the efficiency of communication transmission.The work done in this study includes the following three aspects:(1)A millimeter wave radar vehicle road cooperative pedestrian alarm device using edge computing module is proposed.The device can detect pedestrians in real time,efficiently and accurately,identify and track pedestrians,and alarm drivers.The device includes five modules: radar data acquisition module,edge computing module,data transmission module,cloud server and data display module.(2)In this study,the obj data returned by millimeter wave radar is preprocessed.The data preprocessing mainly uses DBSCAN algorithm for target clustering.After selecting the appropriate threshold,the training accuracy and verification accuracy of multiple features can be higher under the training conditions,and the recognition rate can be about better in the actual scene.Pedestrian tracking uses inter frame correlation algorithm combined with Kalman filter,and implements the algorithm of actual data,which can effectively track pedestrians and maintain good accuracy in multi pedestrian tracking scene.(3)An on-board terminal app is developed based on the Android platform,which can receive the real-time pedestrian warning information sent from the RSU,obtain the speed limit warning sent by the RSU,timely remind the driver in the form of voice broadcast,and complete the early warning.In addition,the app design combines the communication protocols of HTTP and mqtt to upload data to the cloud while receiving data to meet the driving requirements Real usage scenarios.
Keywords/Search Tags:Millimeter wave radar, Edge computing, Pedestrian detection, Pedestrian tracking, Artificial intelligence, Android client
PDF Full Text Request
Related items