Research On Target Tracking And Positioning Algorithm For OCC System

Wireless communication technology is widely used in the field of autonomous driving,which can provide driving assistance information for vehicles.However,there also exist problems such as scarce spectrum resources,serious interference,and high cost simultaneously.Optical Camera Communication(OCC)technology,as a new type of Optical Wireless Communication(OWC)technology,has the advantages of rich spectrum resources,environmentally friendly,economical and practical features,etc.As an effective supplementation with the existing driving-assisted technologies,OCC can help realize mutual coordination to ensure vehicle driving safety,and therefore,boasts a very broad application prospect.The problems existing in the current OCC technology research are as follows:most theoretical algorithm research,lack experimental platform based investigations;the tracking algorithm is difficult to guarantee the accuracy of the targets identification in complex environments,and also suffers a poor real-time performance;Besides,most studies only consider one single experimental scene and lack outdoor environment analysis as well;Positioning algorithm calculation is complicated and has a high measurement error;the information output by the algorithms has not been effectively integrated.Aiming at these problems,this paper will conduct investigations on OCC’s target tracking,positioning and ranging,and coding communication algorithms.Main contributions of this research are as follows:(1)OCC target tracking process optimization and algorithm performance analysisThrough the adjustment of the detection frame,the error tolerance rate of the target tracking process is increased;the online training mode is adopted to integrate the machine learning algorithm into the OCC system.In order to test the performance of the algorithm under mobility scenarios,a mobile slide rail test platform is designed and practically applied to quantify the speed index together with enriching the experimental scene.A variety of experimental scenarios have been considered,such as deformation,occlusion,and interference in the target area.By calculating the tracking accuracy,comparisons between performance of traditional tracking algorithms and machine learning tracking algorithms have been realized,and the superiority of machine learning algorithms in complex mobile scenarios has been proved.Based on the results of the tracking accuracy,tracking stability,algorithm overhead and other indicators under the variations of the target moving speed,comparative performance comparisons of machine learning algorithms were carried out,and the MOSSE algorithm was selected as the final applied tracking algorithm due to its best performance.Target tracking tests have been carried out in outdoor environments.Under such conditions when the complex interfering light sources exist within the target area which moves in random routes,the tracking accuracy of the MOSSE algorithm is 99.11%.The average frame rate of the algorithm is 507.23 fps,and the average time-consuming per frame is 1.97 ms,which meets OCC system reliability and real-time requirements.(2)Research on target positioning and coding algorithm for OCC systemBased on the realizations of accurately identifying and tracking the target,technical route of target LED positioning and coding communication algorithm process aiming at the current set up OCC system is designed.By improving the monocular imaging algorithm of the camera,a positioning algorithm based on the pixel deviation ratio of the target area is proposed to realize the measurement of the actual distance and orientation information between the transceivers.By adopting a combination of selfdefined coding and Manchester coding method,the content of the data frame is rearranged,and a complete coding and decoding communication process is designed to help transmit vehicle speed,steering,braking and other information between the transceivers.Combining positioning and encoding algorithms can deliver driver assistance information and further expand the practicability of OCC technology.The experimental results show that the maximum ranging relative error of the positioning algorithm does not exceed 5%,and the actual distance and deviation angle of the target can be calculated more accurately;the codec communication algorithm can accurately distinguish between conventional data frames and self-defined data frames,effectively eliminate interference information under the condition of the same communication rate,and increase the system throughput.Research work in this paper has a certain reference value for future performance improvement investigations and may provide some technological suggestions for practical applications in the fields of intelligent transportation and automatic driving of the OCC system.