Improving Law Enforcement and Emergency Response to Disabled Vehicle Crashes Using Waze Crowdsourced Data

Abstract

Disabled and abandoned vehicles (DAVs) on limited access freeways are exposure risks that can result in congestion and crashes. Improving law enforcement response to DAVs can reduce their impact and potentially prevent some crashes from occurring. Crowdsourced data, such as Waze user alerts, can provide near real-time information on highway conditions, including the presence of DAVs. The challenge is effectively filtering the data so law enforcement can more efficiently respond to actual DAV incidents. In this paper, over 3.8 million Waze alerts and 329 DAV crashes on Florida limited access roadways from July 1, 2019, to December 31, 2020, were analyzed to determine appropriate matching parameters, characteristics, and potential benefits that could be provided by the Waze alerts. Evaluation of various spatiotemporal buffers to match Waze and crash data showed that buffers of 0.5 km (0.31 mi) and 30 min would provide the best matching. Using these buffers, 41 crashes were found to have matched Waze alerts. Analyses of these matched data showed that DAV Waze alerts would likely provide the most benefits during morning peak hours and on urban interstates. The earlier detection because of Waze alerts could have potentially allowed responders to reach 12 DAVs before a crash occurred. This earlier response could have potentially prevented over US$23 million in comprehensive crash costs. The results from this paper can be used to develop more efficient DAV response strategies using crowdsourced data, allowing law enforcement to better detect DAVs while keeping the number of alerts at a manageable level.