This paper presents an evaluation procedure for an infrastructure based automated valet parking system. Because parking is one of the difficult and complicated tasks for drivers, park assistance system(PAS) has been already developed and commercialized by several auto-manufacturers. As a further step for PAS, researchers are focusing on automated valet parking system which is a fully automated system of PAS and periodically demonstrated their automated parking system. However, from an institutional point of view, evaluation standard and scenarios for automated parking system are progressing slowly compare to the automated driving. Because of specialty that in valet parking system, driver has to get out of the vehicle, most of the developers can not mass produce the system by legal issues. Thus, the necessity of evaluation procedure for parking system rise. Considering many automated valet parking systems are designed with digital map or infrastructure. Thus, in this paper, automated valet parking system cooperating with infrastructure is focused.
To design the evaluation process for automated parking, we divided the parking situation into two sides. 1) Nominal parking process, which is a static obstacle avoiding case related to static factors. 2) Complicated parking process, which is avoiding not only static obstacles but also moving obstacles such as pedestrians. As a valet parking is a very sequential service, to design evaluation items for nominal parking, we considered a procedure of manual parking and divided valet parking process into three different stages. 1) Driving in the parking lot while moving near to the parking space. 2) Parking to the aimed space. 3) Parking out from the space. Finally, the component of nominal parking scenario classified as static factors to test basic parking performance of the target vehicle and addable evaluation scenarios listed up as dynamic factor which can reproduce the complicated and frequent situation that can happen in the parking lot by combining with the nominal scenario. Both the nominal scenario and additional scenarios are organized as an evaluation matrix.