Informative Path Planning of Autonomous Vehicle for Parking Occupancy Estimation

Parking occupancy estimation holds significant potential in facilitating parking resource management and alleviating traffic congestion. While recent strides have been made to integrate robotic systems for occupancy detection, these efforts have predominantly focused on improving the accuracy of spotting individual parking space occupancy through enhanced perception pipelines. Regrettably, the significance of robot path planning often gets overlooked in these methodologies, resulting in operational inflexibility and imprecise estimations for the entire parking area. In light of this gap, we introduce the concept of informative path planning for parking occupancy estimation using an autonomous vehicle. We frame this problem as a Partially Observable Markov Decision Process task, addressing the intricacies of uncertainty inherent in the environment. We proceed by constructing an occupancy state transition model and developing a Bayes filter to estimate occupancy based on noisy sensor measurements. We subsequently propose the Monte Carlo Bayes Filter Tree, a computationally efficient algorithm that leverages progressive widening to generate informative paths. We demonstrate that the proposed approach outperforms benchmark methods in various simulation environments, effectively striking a balance between optimality and computational efficiency.