Understanding the Truck Parking Behavior Using a Data-Driven Approach

Commercial vehicle parking is a significant issue in the United States. However, it is unclear why truck drivers choose to park in unauthorized areas, like ramps, rather than authorized areas. This study used INRIX big truck probe data in Maryland to explore the factors that affect truck drivers’ parking decisions. Two interpretable machine learning frameworks—SHapley Additive exPlanations (SHAP) and Stable and Interpretable RUle Set (SIRUS)—were adopted to aid the research. The study finds that whether or not a designated parking lot is at capacity has a limited impact on truck drivers’ decision to park in authorized or unauthorized areas. Instead, the number of trucks parked in unauthorized areas upon a truck’s arrival is a dominant factor that triggers a decision to park in an unauthorized area, regardless of whether the parking lot is full or not. A possible explanation is the presence of trucks in unauthorized areas gives truck drivers the impression that the designated parking lot is full. Trucks with relatively longer next trip (longer than 60 miles) are more likely to park at authorized areas. Trucks with a shorter next trip are more likely to park in unauthorized areas. These trucks are probably staging, waiting for delivery windows, or avoiding peak hour traffic. The results also show that medium-weight class trucks are more likely to park in unauthorized areas than heavy trucks. Truck drivers are more likely to park in authorized areas during the mid-day and afternoon. These findings could help transportation agencies understand parking behavior and implement appropriate strategies and policies to improve truck parking.