| Road friction coefficient is one of the most important parameters in theroad traffic accident and identification, which is very important that how tomeasure accurately and predict. The paper gets the friction coefficient at allsorts’ road and under different condition through road friction coefficient test,and then establish database of pavement characteristics parameters to providesthe basic data for analyzing traffic accident. It also predicts the frictioncoefficient based on General Regression Neural network.First, road friction coefficient test of highway and urban road at thecondition of different levels, different surface, different pavement behavior anddifferent service life, is tested by JN-type1road friction coefficient tester, andthe result of which is analyzed.Based on test data and using Microsoft Access2000to establish a database,the database system of road characteristic parameters has been developedadopting Visual C++6.0software. This system can inquiry road frictioncoefficient and manage Software system parameters Then, based on General Regression Neural network (GRNN), the author fpointed out that the feasibility of road friction coefficient prediction model. Thepaper determines the impact fact of the friction coefficient prediction model andreaches the smoothing parameter optimization method. Finally the papervalidates the prediction model acco-rding to the road test data.Finally, the paper using the JN-1road friction coefficient tester and theBM-Ⅱ pendulum friction tester measure the same road friction coefficient.Using the analysis of the two results, the paper research the relationship offriction coefficient measured the JN-1road friction coefficient tester and BPNmeasured the BM-Ⅱ pendulum friction tester. And the paper usingthe SG-630portable brake performance tester does automotivebrake performance test, researching the conversion relationship of road frictioncoefficient between using in the automotive brake and measured the JN-1roadfriction coefficient tester. |
PDF Full Text Request