AD1

Monday, November 1, 2010

Car motion state of online measurement and forecasting technology

Motor Sport to carry out the state of online measurement and forecasting technology, early warning of future time travel vehicle state of motion, pre-warning the driver, or even direct drive vehicle control system, take the initiative for improving traffic safety and reduce road traffic accidents will play a very important effect. Motorsport state line measurement and forecasting technology is the process of moving cars on the road, through the arrangement of inertial sensors in the car on-line monitoring of vehicles driving on the basis of the motion state, the use of certain techniques and methods to improve the accuracy reliably predict the future car short time (1-3 seconds) traveling within the state of motion. To achieve the state parameters of motor sport line measurement purposes, self-designed micro inertial measurement unit MIMU (Micro Inertial Measurement Unit), its error model is analyzed, and designed a MIMU calibration systems and sensor calibration work carried out. Further research the car's attitude and speed integration algorithm, taking into account the MEMS (Micro-Electro-Mechanical Systems) sensors are relatively low accuracy of the factors, the design of the Kalman filter to complete the integration of the sensor signal processing parameters to obtain the best motor sport status excellent estimate. And further work carried out Matlab simulation, the simulation results show that the vehicle attitude and velocity integration method and the Kalman filter design accuracy and reliability. Access to sports car driving on the basis of state parameters, based on autoregressive (Auto-Regressive, AR) model, the Ladder (Multi-Level Recursive, MLR) model to carry out the car motion state parameters of the prediction. A detailed analysis of the specific model prediction of both the process and focus on traffic safety is closely related with the 6 main parameters of the motion state of the work carried out Matlab simulation, the simulation results showed that the two forecasting methods are forecast to reach a certain effect, and more layer hierarchical approach to motor sport because when you take into account variability and nonlinear characteristics, and its forecasting accuracy is higher than the AR prediction method. Based VB2005, NI Measurement studio and Matlab.NET generator developed a car motion state line measurement and prediction software. VB and Matlab through the introduction of mixed programming ideas, give full play to the advantages of VB and Matlab interface advantages of mathematical operations, reduce programming effort and improve efficiency of software development.The developed software not only enables the car motion state parameters of line measurement and real-time forecasting purposes, but also has a good interactive interface. Finally, vehicle test platform built and launched a vehicle road test. The results show that the method of AR model forecast state of the car a strong correlation between the parameters from the yaw rate, yaw angle, roll angle, vertical speed of 4 who achieved very good prediction results, the average forecast forward 1 second relative error rate of 5% to 6%, 2 seconds ahead of 12% to 13%, 3 seconds ahead of 20% to 22%. And test and simulation results have good agreement, the full description of the car motion state carried out online research measurement and prediction accuracy and feasibility. 
 
Key words: motor sports state parameters, on-line measurement and prediction, micro inertial measurement unit, since the regression model, multi-layer hierarchical model

No comments:

Post a Comment