Real-time Recognition Algorithm of Longitudinal Road Slope Based on Vehicle Dynamics and Acceleration Sensor Information

Longitudinal slope is a usual parameter in the electronic control system of vehicle chassis and transmission. The work veracity and drive comfort of control strategy are directly affected by the accuracy and smoothness of the slope recognition. The problem of recognition is that the signal quality of acceleration sensor is affected by the vehicle dynamic behavior with high frequency and the complexities of drive cycle, which make the estimated road slope with this method have a low accuracy. And the observational method based on a complicated nonlinearity model can't realize a real-time operation. Meanwhile if a kind of recognition method can be used by multi-system application, the basic and usual input variables are required. In order to solve these problems, an integration method with the acceleration slope method and vehicle dynamic slope method is presented in this paper. According to their high precision area sufficiently, the confidence factor is calculated. And then the raw slope can be obtained. Eventually the raw slope is processed by a generalized hysteresis filter to get the accurate and smooth slope. The test verifies that this algorithm is effective. This recognition algorithm with small calculating amount, low cost and easy realization can be adopted in multiple real-time electronic control systems.