In this paper, it is investigated how the planning of physically feasible trajectories for automated vehicles can be ensured in a motion control architecture with separated planning and control stages. Within such an architecture, the planning stage shall not be adapted to a specific vehicle during development to assure reusability of developed algorithms. It is shown, that the planning stage does not need any information about downstream modules such as the controller and the actuators, as long as certain physical execution limits are considered during trajectory planning. For these constraints, a feedback loop from the control to the planning stage is proposed that ensures feasible trajectories, as long as these limits are accurately estimated. An additional escalation mechanism is developed to prevent safety critical behavior of the vehicle if errors are made during estimation of the relevant limits. The proposed approach hence encapsulates information within the vehicle and enables a greater separation of planning and execution stages as well as the use of different planning algorithms with a common controller. It can therefore raise the reusability of developed functions and reduce the necessary customization of algorithms to a specific vehicle architecture.