Estimation of the acceleration of a car under performance tests by using an optimal observer

In this paper, the acceleration of a car under performance tests is estimated by using a Kalman filter. Here, the observation vector consists of the observation of both the velocity and the longitudinal acceleration of the car. This is the process vector and is the input of the filter. The output is the filtered estimate of the state vector, which consist of the velocity and longitudinal acceleration of the car. The accelerometer is modeled as a linear dynamical system in which the acceleration is a Wiener process, the state vector is corrupted by process noise and the observation vector by measurement noise. The process noise and the measurement noise are modeled as zero-mean, white-noise processes. The error-performance surface of the filter is obtained by taking into consideration several values of correlation matrix of process and measurement noise, and the experimental results show a satisfactory improvement in the signal-to-noise ratio of the system.