Autonomous valet parking optimization with two-step reservation and pricing strategy

With the development of autonomous vehicle, the autonomous valet parking has been attracting extensive attention, by relieving users' inconvenience from parking. Compared with shortrange autonomous valet parking, long-range autonomous valet parking extends parking features with bridge of long-range travelling and parking in a user-friendly cycle. However, due to the uncertainty of trip routes, the performance of single reservation mechanism on parking lot status prediction is limited. Besides, unified parking pricing intensifies parking competition and causes local congestion. Therefore, we propose a long-range autonomous valet parking framework based on Two-step Reservation Mechanism. The proposed scheme provides the first reservation service before arriving at Drop-off spots (regular step), so as to anticipate parking demand of parking lot. Then, it provides the second reservation service when arriving at Drop-off spots (extra step) for updating the optimality of parking decision. The future parking lot status can be predicted with the impact of concurrent requests and reservation execution probability. Meanwhile, we propose a heterogeneous pricing strategy, in which the parking lots with higher parking demands would be adjusted with higher prices. Simulation results show that our scheme outperforms literature works on reducing parking cost and balancing parking demand.