Research On Clonal Selection Algorithm In The Application Of Vehicle Classification

In the work of the highway modernized management, the automatic vehicle classification may assist toll collector charging and realize the automatic statistics of the transportation data, which is advantageous for the reduction of the charge outflow caused by the manual charge and the integration management of the transportation administration department and improves the using efficiency of the highway.Firstly the vehicle detector based on the induction coil is used to gather vehicle type data in this paper and the work of the digital filtering, the sampling and the characteristic extracting are carried out to the gathered vehicle type data. Then the clonal selection algorithm based on the artificial immune principle is adopted to realize the vehicle automatic classification, which provides a new research method for vehicle classification.Aiming at the problems existed in the clonal selection algorithm that the binary coding causes the troubles of complex computation and"the mischance of the dimensions"and the fixed variation mode brings the problem of slow convergent speeding, an improved clonal selection algorithm according to the clonal selection principle based on the artificial immune system is adopted in this paper. In the improved clonal selection algorithm, the antibody is the candidate solution of the real coding, and it is introduced the thoughts of the clonal selection expanding, the clone deletion, the clone variation and the antibody circulation supplement and so on. The clonal selection expanding enables the excellent antibodies with high affinity obtain bigger clone scale, and the clone variation and the clone deletion enable the algorithm search more excellent antibodies according to a higher probability in the foundation of the excellent antibodies, which makes the algorithm have the strong partial optimization ability. Meanwhile, the immune body circulation supplement mechanism enables the algorithm maintain the multiplicity of the antibody and strengthen the whole optimization ability of the algorithm.Finally,through the simulation to 6 kinds of typical vehicle classification problem, the parameters of the improved clonal selection algorithm which influenced the algorithm performance are analyzed and the validity of the improved algorithm is validated. In the end, the simulation result of the algorithm in this paper is compared with the clonal selection algorithm and the genetic algorithm which further explains that the improved clonal selection algorithm has the performances of the good convergent, the rapid classification speed and the better classification accuracy rate in the application of vehicle classification.