This paper summarizes the results of research on target parking position designation for automatic parking systems and proposes development directions. First, as parking-related systems become increasingly common in the intelligent vehicle field, many terms are used confusingly, so automatic parking systems are defined in comparison with other systems. After that, various methods of the target parking position designation in automatic parking systems are classified into user interfaced-based, free space-based, parking slot marking-based, and intelligent parking management system (IPMS)-based, and the various approaches are summarized. In this paper, the fusion of image sensor-based parking slot marking recognition and range sensor-based free space recognition, which have complementary characteristics, is expected to be the most important approach in the near future. In particular, this paper proposes that in order to effectively apply the deep learning-based method with high recognition performance, open competition through datasets for various challenging situations is necessary. If vehicle to infrastructure (V2I) communication is standardized in the distant future, the importance of IPMS-based is expected to increase. In order to cope with this, this paper proposes that the commercialization and standardization of high-precision indoor positioning along with the standardization of vehicle to everything (V2X) communication are necessary.
展开▼
