Using Crowdsourced Data to Reduce Traffic Congestion by Improving Detection of and Response to Disabled or Abandoned Vehicles on Florida Limited-Access Facilities

Abstract

Disabled or abandoned vehicles (DAVs) cause a significant proportion of non-recurring traffic congestion. Waze data has the potential to help traffic management center (TMC) operators detect and respond to DAVs more quickly, reducing congestion. Previous studies have examined Waze data on a small scale, but not at a statewide level. This paper analyzes over 2 years of DAV Waze data on Florida Department of Transportation (FDOT) limited-access roadways (with more than 10 million alerts) and compares them with reported DAV events (crashes and non-crash TMC reports). Over 46% of the DAV events had an associated Waze alert before the DAV event occurred, indicating significant potential for earlier detection. These Waze alerts typically happened about 16 min before the DAV event and were most common during daytime hours and in urban areas. Two roadway segments, I-4 in FDOT District 5 and State Road 91 (SR-91) in FDOT District 4, were examined in more detail. A methodology was developed to estimate the delay reduction and congestion savings achieved by the earlier detection provided by the Waze alerts for lane-blocking and shoulder-blocking disabled vehicle events. The estimated congestion savings for the second half of 2019 were $4.3 million (I-4) and $2.5 million (SR-91), with benefit-cost ratios of 61 and 27, respectively. An additional $4.3 million could potentially have been saved as a result of Waze alerts allowing responders to reach DAVs before a crash occurred. These results can help agencies understand how crowdsourced data can be effectively utilized to best improve freeway operations.