Using Environmental Information for Enhancing AEB System

Sammanfattning

The Advanced Emergency Brake (AEB) system has had a significant effect on road safety since its introduction. Particularly, safety regarding semi-trailer trucks seem to greatly benefit from implementation of the AEB system due to their longer stopping distance. However, AEB systems can be improved upon in various areas. The effectiveness of AEB systems decreases when driving in adverse weather conditions. Many studies assume that information about the road and environment are known constant certainties. However, the road status is a typical uncertainty that can significantly affect the braking distance during emergency braking of a truck. While multiple methods for estimation or measurement of various road parameters have been researched, most have been developed with passenger vehicles in mind. As truck dynamics differ significantly from passenger car dynamics, these estimation and measurement methods require further analysis for adaptation to articulated vehicles. The objective of this thesis was to develop a model to enhance the AEB systems of Scania trucks to adapt to external conditions. This was done by first performing a literature study to analyze what external conditions can affect braking and how these factors impact the braking performance. Afterwards a model, that estimates or predicts these factors, was developed. From the literature research, it was concluded that the most important environmental factors to consider during emergency braking are the road slope and tire-road friction coefficient. Within previous research, these factors were often determined using online estimation methods as this reduces problems with correlation relationships and expensive sensor use. The estimation model for road slope and tire-road friction was developed in MATLAB Simulink and validated using vehicle dynamics data from IPG Truckmaker. The model makes use of the longitudinal equation of motion of a semi-trailer truck combined with recursive least squares filtering to estimate the road slope while driving. To determine the tire-road friction coefficient under braking, the longitudinal equation of motion was combined with a single wheel dynamics model to first estimate the longitudinal tire forces through Kalman Filtering. These estimated tire forces are then used as an input to estimate the tire-road friction coefficient using a longitudinal linearized Dugoff tire model and recursive least square estimation method. The validated model showed that the estimator model produces results within an accurate range with reasonable consistency and reliability. Overall the work in this thesis present a proof of concept for estimating the external conditions that are of primary importance for braking performance and the AEB system. While further development is necessary before real-time implementation in safety systems, the work presented allows for a starting point of enhancing AEB systems in semi-trailer trucks further.