Modeling Driver Intent in Potential Right Turning Scenarios

The objective of this work was to build predictive models of driver intent at T-intersections based on vehicle speed, acceleration, brake pedal state, and speed limit data from the Second Strategic Highway Research Program. The generated models predict if the driver will go straight or turn right at the intersection. The models show promise in predictive ability and could be incorporated into an automated driving system to help inform the system when more conservative driving might be beneficial. However, the derived models should not be used as the only source for safety decision-making because the risk of a crash would be too high.