Browsing by Author "Das, Subasish"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Autonomous Delivery Vehicle as a Disruptive Technology: How to Shape the Future with a Focus on Safety?
    Das, Subasish; Tsapakis, Ioannis; Wei, Zihang; Elgart, Zachary; Kutela, Boniphace; Vierkant, Valerie; Li, Eric (SAFE-D: Safety Through Disruption National University Transportation Center, 2022-09)
    The National Highway Traffic Safety Administration recently granted permission to deploy low-speed autonomous delivery vehicles (ADVs) on roadways. Although the mobility of ADVs is limited to low-speed roads and these vehicles are occupantless, frequent stops and mobility among residential neighborhoods cause safety- related concerns. There is consequently a need for a comprehensive safety impact analysis of ADVs. This study examined the safety implications and safety impacts of ADVs by using novel approaches. This research prepared several datasets such as fatal crash data, aggregated ADV trips and trajectories, and real-world crash data from the scenario design for an ADV-related operational design domain. Association rules mining was applied to the datasets to identify significant patterns. This study generated a total of 80 association rules that provide risk patterns associated with ADVs. The rules can be used as prospective benchmarks to examine how rule-based risk patterns can be reduced by ADVs that replace human-driven trips.
  • Thumbnail Image
    Developing AI-Driven Safe Navigation Tool
    Das, Subasish; Sohrabi, Soheil; Tsapakis, Ioannis; Ye, Xinyue; Weng, Yanmo; Li, Shoujia; Torbic, Darren (Safe-D University Transportation Center, 2023-09)
    Popular navigation applications such as Google Maps and Apple Maps provide distance-based or travel timebased alternative routes with no real-time risk scoring. There is a need for a real-time navigation system that can provide the data-driven decision on the safest path or route. By leveraging data from a diverse range of historical and real-time sources, this study successfully developed a user interface for a navigation tool or application that offers informed and data-driven decisions regarding the safest navigation options. The interface considers multiple scoring factors, including safety, distance, travel time, and an overall scoring metric. This study made a distinctive and valuable contribution by designing and implementing a robust safe navigation tool driven by artificial intelligence. Unlike existing navigation tools that offer multiple uninformed route options, this tool provides users with an informed decision on the safest route. By leveraging advanced AI algorithms and integrating various data sources, this navigation tool enhances the accuracy and reliability of route selection, thereby improving overall road safety and ensuring users can make informed decisions for their journeys.
  • Thumbnail Image
    Exploring Crowdsourced Monitoring Data for Safety
    Turner, Shawn M.; Martin, Michael W.; Griffin, Greg P.; Le, Minh; Das, Subasish; Wang, Ruihong; Dadashova, Bahar; Li, Xiao (SAFE-D: Safety Through Disruption National University Transportation Center, 2020-03)
    This project included four distinct but related exploratory studies of data sources that could improve roadway safety analysis. The first effort evaluated passively gathered crowdsourced bicyclist activity data from StreetLight Data and found promising correlations (R2 of 62% and 69% for monthly weekday and weekend daily averages) when the StreetLight data were compared to bicyclist counts from 32 locations in eight Texas cities, and even better correlation (R2 of 94%) when compared with countywide Strava data expanded to represent total bicycling activity. The second effort evaluated the pedestrian counting accuracy of the Miovision system and found 15% error for daytime and 24% error for nighttime conditions. The third effort used INRIX trip trace data to determine origin-destination patterns and developed 40 decision rules to define the origin-destination patterns. The fourth effort analyzed crowdsourced Waze data (i.e., traffic incidents) and found it to be a reliable alternative to observed and predicted crashes, with the ability to identify high-risk locations: 77% of high-risk locations identified from police-reported crashes were also identified as high-risk in Waze data. The researchers propose a method to treat the redundant Waze reports and to match the unique Waze incidents with police crash reports.
  • Thumbnail Image
    Use of Disruptive Technologies to Support Safety Analysis and Meet New Federal Requirements
    Tsapakis, Ioannis; Das, Subasish; Khodadadi, Ali; Lord, Dominique; Morris, Jessica; Li, Eric (SAFE-D: Safety Through Disruption National University Transportation Center, 2021-03)
    States are required to have access to annual average daily traffic (AADT) for all public paved roads, including non-federal aid system (NFAS) roadways. The expectation is to use AADT estimates in data-driven safety analysis. Because collecting data on NFAS roads is financially difficult, agencies are interested in exploring affordable ways to estimate AADT. The goal of this project was to determine the accuracy of AADT estimates developed from alternative data sources and quantify the impact of AADT on safety analysis. The researchers compared 2017 AADT data provided by the Texas and Virginia Departments of Transportation against probebased AADT estimates supplied by StreetLight Data Inc. Further, the research team developed safety performance functions (SPFs) for Texas and Virginia and performed a sensitivity analysis to determine the effects of AADT on the results obtained from the empirical Bayes method that uses SPFs. The results showed that the errors stemming from the probe AADT estimates were lower than those reported in a similar study that used 2015 AADT estimates. The sensitivity analysis revealed that the impact of AADT on safety analysis mainly depends on the size of the network, the AADT coefficients, and the overdispersion parameter of the SPFs.
  • Thumbnail Image
    Vehicle Operating Speed on Urban Arterial Roadways
    Fitzpatrick, Kay; Das, Subasish (SAFE-D: Safety Through Disruption National University Transportation Center, 2019-01)
    This research explored (1) the relationship between suburban vehicle operating speed and roadway characteristics,especially the presence of bicyclists and (2) whether crowdsourced speed data could be used to estimate theunconstrained speed for a location. Both vehicle volume per lane and bicycle volume were found to be influential inaffecting average speed on lower speed urban arterial roadways. For 40.3 km/hr (25 mph) sites, an increase of 19vehicles per 15-min period would decrease average speed by 1.6 km/hr (1 mph), and an increase of more than 39bicyclists per 15-min period would decrease average vehicle speed by a similar amount. Because of the limited numberof 15-min periods with bicycle counts greater than 1, the research team also developed a model using all available 15-min periods with on-road speed data. Speed and volume data in 15-min increments for 2 weeks at nine sites wereobtained using on-road tubes and via a vendor of crowdsourced speed data. The difference between the tube data andthe crowdsourced data was calculated and called TMCS as a representation of tube (T) minus (M) crowdsourced (CS).The geometric variables that had the greatest influence on TMCS were the number of signals and the number ofdriveways within a corridor. When only including non-congested periods, weekends (Saturday or Sunday) wereassociated with the smallest TMCS.