Simulation Model of Vehicle Inertial Sensor Based on Navigation Parameter Backtracking Algorithm

With the development of autonomous driving technology, vehicle navigation systems require much higher inertial sensor accuracy. In the traditional navigation test process, we need a three-axis accelerometer and gyroscope (or three single-axis accelerometers and gyroscopes) to form an inertial measurement unit (IMU) as a benchmark for experiments to test the accuracy of navigation calculations. It takes a lot of resources to implement the test, and the accuracy of physical sensors is hard to control during a long time. In order to facilitate the navigation test, the article proposes a novel simulation model of inertial sensors, which represents a physical IMU while maintaining a high-precision reference. Firstly, based on the output characteristics of the gyroscope sensor, use the navigation parameter backtracking algorithm to perform modeling, complete error compensation, and obtain high-precision angular incremental output. Secondly, use the existing angular incremental output and the output characteristics of the accelerometer sensor to calculate the simulation model of the accelerometer sensor. Finally, the two models are tested under extreme environments and compared with the accuracy of common sensors of the same type. The test results show that the error of this model is lower than the inertial sensors commonly used, the error of the gyroscope model is below the order of 10-4 deg/h, and the error of the accelerometer model is below 10-7g, which demonstrates the effectiveness of the simulation model. The model has important reference value for the accuracy analysis and test of the navigation system, it can be used as a sensor benchmark to test the accuracy of the navigation algorithm or overlay the sensor model with various error models to test the effect of different errors on the navigation results.