A Novel Vehicle Fleet Data-Assisted Map Matching Algorithm for Safety Ranking and Road Classification in Metropolitan Areas using Low-Sampled GPS Trajectories

Probe vehicle data permit accurate inference of a road-network condition and drivers’ behavior. However, even high-precision GPS devices deliver noisy trajectories. Therefore, a map matching algorithm is needed to infer the true trajectory. Available methods become inadequate in urban canyons at low sampling rates. Firstly, the authors demonstrate a novel data-assisted cost-based map matching algorithm. Their contributions include: (i) Robustness to sparse-sampled GPS trajectories due to the novel transition probability distribution used in the Markov model to solve the matching problem. (ii) Robustness to GPS noise. Its parameters adapt based on the number of satellites visible at each location sample, and the altitude value which assists the map matching. (iii) Improvement of accuracy using the sampled true maximum vehicle speed to narrow down the entire road-network to an accurate trust-region of the possibly-visited routes within a sample interval. And, (iv) using a connected sliding window when finding the most probable path. (v) Unlike other algorithms, the method is road-network and map independent. The algorithm is tested on data provided by the New York City Department of Transportation. Because this data lacked ground-truth, qualitative assessment against other open-source map matching algorithms, such as Google Snap to Roads API and BMW car IT Barefoot, reveals an improvement of accuracy at noisy and low-speed locations. Secondly, the authors create a Driver Behavior Index system (DBI) for New York City. It provides safety-related metrics of city roads and identifies speeding, congestion, hard braking, and acceleration hotspots, which will be correlated later with crash data.

• Supplemental Notes:
  • This paper was sponsored by TRB committee ABJ30 Standing Committee on Urban Transportation Data and Information Systems.
• Corporate Authors:

  Transportation Research Board

  ,    
• Authors:
  • Alrassy, Patrick
  • Jang, Jinwoo
  • Smyth, Andrew W
• Conference:
  • Transportation Research Board 98th Annual Meeting
  • Location: Washington DC, United States
  • Date: 2019-1-13 to 2019-1-17
• Date: 2019

Language

• English

Media Info

• Media Type: Digital/other
• Features: Maps; References;
• Pagination: 6p

Subject/Index Terms

• TRT Terms: Algorithms; Digital mapping; Global Positioning System; Metropolitan areas; Safety analysis; Traffic safety; Trajectory
• Uncontrolled Terms: Map matching
• Geographic Terms: New York (New York)
• Subject Areas: Data and Information Technology; Highways; Operations and Traffic Management;

Filing Info

• Accession Number: 01697523
• Record Type: Publication
• Report/Paper Numbers: 19-03327
• Files: TRIS, TRB, ATRI
• Created Date: Mar 1 2019 3:51PM