Weigh-in-motion sensors: an underutilized source for inferring occupancy trend

Weigh-in-motion (WIM) sensors are widely deployed traffic monitoring systems that capture vehicle characteristics, including vehicle weight and axle spacing. While they have been used for a variety of applications, such as pavement management, their potential to support the analysis of passenger vehicle occupancy trends has not been investigated. This study explores this potential by developing a WIM-based framework that uses vehicle payload information to infer temporal changes in passenger vehicle occupancy. In passenger vehicles, where payload variations are expected to be largely associated with changes in passenger load, this approach provides an indirect means of investigating occupancy trends. Using an open dataset from New York City, the proposed framework is evaluated using data from the COVID-19 pandemic, a period characterized by significant disruptions in travel behavior. The methodology comprises three components. First, a Welch t -test confirms significant differences in gross vehicle weight between the pre-pandemic and pandemic periods, suggesting sensitivity to payload changes. Second, a change point detection approach, formulated as a penalized cost minimization problem and implemented using the Pruned Exact Linear Time algorithm with a kernel-based cost function, identifies key transitions coincide with major pandemic-related events. Third, we develop a generalized clustering-based framework to construct a relative payload index that more closely represents occupancy variation than gross vehicle weight alone. The resulting patterns are compared with external benchmarks from the National Household Travel Survey and show consistency with the observed decline and daily temporal variation in vehicle occupancy. The results highlight the potential of WIM data to provide automated, cost-efficient monitoring of aggregate occupancy trends, particularly in contexts where traditional data collection methods are limited. Therefore, WIM sensors can serve as a valuable tool to support aggregate-level transportation planning and policy evaluation, and long-term monitoring of occupancy trends.