This paper presents a novel approach for path-following control of a four-wheeled
autonomous vehicle. The rear wheels of the vehicle are driven independently, all four
wheels can be braked independently, and the front wheels are steered together. The
proposed cascade structure consists of two convex optimization-based parts: one for
path-following and another for the control allocation problem of the actuators. The
control algorithm presents cost functions for the allocation problem focusing on safety.
The proposed cost functions were examined and compared to former ones in a simulation
environment. After all, the controller was tested in real-time test on a Lotus Evora
test vehicle developed by ThyssenKrupp.