Study On Automatic Recognition Of Vehicle Test Scenarios And Conditions

In the traditional site test process,the test data will be derived through a certain technical means and transferred to the relevant computing terminal equipment for data processing and analysis and according to the data of vehicle performance evaluation.In the process,on the one hand due to the continuous migration of test data may cause data loss,on the other hand,after the end of test whether the test in line with the test standards are not determined in time,once the data is failed found in the data processing phase,a new round of car test has to be done again,which wastes lots of car test resources.Therefore,a set of test support system is needed to realize the judgment of the effectiveness of the current test and to calculate the relevant vehicle performance evaluation index in tome to realize the automatic supervision of the test process.The purpose of the project is to design a set of test system for automobile test.The system is able to test validity of the current test scenario and the working condition according to the collected test signal curve.For the invalid test,promptly prompted the pilot to conduct a new round of car test.For effective testing,timely calculation of vehicle evaluation indicators for the pilot reference.The entire test system can be automated to assist the pilot to carry out automotive testing,to further improve the test efficiency.The pilot test system is divided into two sub-topics,of which the subject is responsible for the experimental program and the method of automatic identification of the method of study;the sister project is responsible for the method and automotive test aids in the engineering application and calculate the relevant test evaluation indicators,to achieve the test of the auxiliary test system of the entire process of automation,intelligent.In this paper,we mainly carry out a pattern recognition method which can be classified as a qualified test mode.The test scenario and the condition can be successfully classified as qualified test mode.This means that the experiment is effective and the classification is successful.In order to achieve the goal of this research,the paper mainly carried out the following three aspects:(1)Based on the characteristics of the continuous time series with the discrete points of the experimental signal curve,the implicit Markov(HMM)pattern recognition method which is more mature in the field of Chinese character recognition and speech recognition is introduced.The HMM method and the characteristics of the automobile test signal curve Combined with the development of automotive test program and automatic identification of engineering applications.(2)Due to the limitations of the HMM model revaluation,the improved basic clonal selection algorithm—immune network clone optimization algorithm and HMM parameter reassignment algorithm,greatly avoids the problem that the parameter revaluation falls into the local extremum problem,and achieved a more satisfactory application effect.(3)At the same time,the filtering effect of several typical filters is analyzed and the Tukey window FIR filter is used to realize the experimental data of the test signal.And a result display interface for the automotive test assistance test system was designed.The experimental results show that the HMM recognition method can achieve the success rate of pattern recognition and the success rate of the vehicle test scheme and the condition of the vehicle,and successfully apply it to the advanced driving simulator of the automobile.In this paper,To achieve the automatic supervision of the automotive test process.