Research on Data-Driven Methodologies for Expressway Emergency Rescue Point Location

The rapid expansion of expressway networks has highlighted the irregular distribution of traffic flow in both spatial and temporal dimensions; there is an escalating demand for more detailed positioning of expressway emergency rescue points. This research delves into the spatiotemporal distribution traits of expressway vehicle models, based on expressway toll data employing community-detection algorithms to partition the operating origin and destination of four basic models, namely, minibuses, buses, minivans, and large trucks. Separate weights are assigned to expressway crash probability and crash intensity for the base model. Then the weighted shape centers are identified by integrating the shape centers of each model community using the K-nearest-neighbor algorithm. Following this, K-dimensional tree algorithms are engaged to match the weighted shape centers with toll stations, using tollbooths as site selection for rescue points. Using vehicle toll data from a Chinese city expressway as a case study, we implement the aforementioned method. With a layout of eight first-level emergency rescue points and 23 second-level emergency rescue points for the region, when juxtaposed with the P-center siting model, our method reduces the average rescue time for first- and second-level crashes by approximately 22.02%. Similarly, for third- and fourth-level incidents, there is a 21.33% reduction in response time. The variability in emergency response times across both siting models also decreases by 37.37% and 16.14%, respectively. These metrics underscore the suitability of our method for addressing the distinct needs of expressway emergency response, enhancing the effectiveness of the rescue-center placement.