Most of today's vehicles are over-actuated. Active researches are carried out to find satisfactory chassis systems coordination strategies. Regarding autonomous vehicles, qualitative objectives such as comfort and confidence feelings should also be ensured. This paper shows how to take into account different constraints in a multi-objective control allocation framework. A weighted least squares algorithm is used to favour an objective over another depending on the situation. Results show that different coordination solutions could be generated leading to different feelings. This can be adapted to fit driver's expectations when it comes to autonomous driving.