Decisions in Highly Automated Vehicles for Passing Urban Intersections with Support Vector Machines

Intersections and road junctions are with higher traffic accidents due to the wrong decisions of human drivers. In this paper, we consider an artificial intelligence method to mimic decisions of human drivers for highly automated vehicles at passing an urban intersection. We applied the Principal Component Analysis and uniformly scaling for the Support Vector Machine learning is applied to model time series features. The effect due to misspecification by ignoring time series issue is investigated through the comparison of predicted action accurate rate is investigated by a simulation study on the software PreScan.