Predictive manoeuvre generation for automated driving

This paper focuses on the problem of trajectory planning in an autonomous guidance application for one-way, two-lane roads. The problem is formulated in a receding horizon framework, as the minimization of the deviation from a desired velocity subject to a set of constraints introduced to avoid collision with surrounding vehicles, and to stay within the lane boundaries. As well known, this formulation can result in planning algorithms with prohibitive computational complexity, thus preventing real-time implementation. To avoid this limitation, the paper shows how the structured environment of one-way roads, can be exploited in order to formulate a low complexity receding horizon problem that can be efficiently solved in real-time. The proposed algorithm is demonstrated in simulations considering overtake manoeuvres.