Prediction of Pedestrian Crashes at Midblock Crossing Areas using Site and Behavioral Characteristics Preliminary Findings Timothy J. Gates, Ph.D., P.E. June 22, 2016
Pedestrian-Involved Crashes Pedestrian-Involved Fatalities Background Pedestrian Crash Facts ~4,800 pedestrian fatalities in US annually (steady) ~2,275 pedestrian crashes in Michigan annually (steady) 130 150 pedestrian fatalities annually (increasing) 2300 2280 2260 2240 2220 2200 2180 2160 2140 2120 Annual Pedestrian Crashes and Fatalities in Michigan 2010 2011 2012 2013 2014 Pedestrian-involved crashes Year 150 148 146 144 142 140 138 136 134 132 130 Pedestrian-involved fatalities 2
Background The Highway Safety Manual (HSM) was published by AASHTO in 2010 and sets the standard for SPF/CMF development Challenges: HSM only provides for pedestrian SPFs at intersections based on land-use characteristics Research is limited in terms of disaggregate-level studies considering volume, geometry, etc Ped crashes are extremely rare events Are proxy measures available? 3
Background Why do we need better SPFs for pedestrians? In Michigan, from 2010 to 2014: Total traffic fatalities decreased by 7.14% Pedestrian fatalities increased by 12.12% Recent policy initiatives encourage more pedestrian infrastructure (and subsequent ped activity/exposure) Safe Routes to School Complete Streets TIP Trails and Greenways 4
Methods: Site Selection More than 30 sites were selected in Detroit, East Lansing and Kalamazoo County # Pedestrian Crashes 2010-2014 County population (2013) Pedestrian Crashes per 10,000 people Ranking Wayne 3531 1,775,273 19.9 1 Ingham 430 282,234 15.2 2 Kent 884 621,700 14.2 3 Washtenaw 500 354,240 14.1 4 Kalamazoo 345 256,725 13.4 5 5
Methods: Site Selection Sites were selected to provide a broad range of: Vehicle and pedestrian volumes Geometric characteristics Crossing facility types Traffic control devices 6
Methods: Node vs. Segment Analyses Data were procured for analysis in two methods: Segment (purple) Node (red) Segment analysis encompasses the duration of the segment, between signals or stop signs Node analysis covers a 150 radius from the center of the crosswalk 7
Methods: Site Characteristics For both segment and node analysis: Speed Limit Presence of Street Lighting Type of Crosswalk Crosswalk Length Number and Type of Lanes Bike lanes, parking lanes, shoulders, turning lanes For node analysis: Driveway or Minor St. Presence Distance to Nearest Bus Stop For segment analysis: Crosswalk Density Driveway Density Bus Stop Density 8
Methods: Video Recording Elevated cameras were used to record volume and behavioral data In most cases, video was recorded and data extracted on weekdays during mid-day (9 AM to 4 PM) 9
Methods: Video Review/Assessment Pedestrian and vehicular counts and behavioral characteristics were assessed by trained technicians 1 hour increments Behavioral characteristics Staged crossing behavioral data used to measure driver yielding compliance Naturalistic behavioral data used for evasive maneuvers (conflicts) and jaywalking 10
Methods: Staged Crossing (1) 11
Methods: Staged Crossing (2) 12
Methods: Evasive Maneuver (1) 13
Methods: Evasive Maneuver (2) 14
Methods: Crash Data 10 years of ped crash data collected Collected for node analysis and segment analysis For node analysis, crash data collected within 150 radius of the crosswalk Crash reports downloaded and reviewed/screened 15
Pedestrian Crashes per Marked Crosswalk Cluster 5 17 27 103 107 110 Cluster 3 Cluster 1 126 213 217 Cluster 7 214 Vehicle Hourly Volume and Pedestrian Crossings per Hour Yielding Compliance Rate Pedestrian Evasive Maneuver Rate Preliminary Results: Segment Analysis Pedestrian Crashes per Marked Crosswalk Yielding Compliance and Evasive Maneuvers 4.00 2500 100% 10% 3.50 3.00 2000 90% 80% 9% 8% 2.50 2.00 1.50 1500 1000 70% 60% 50% 7% 6% 5% 1.00 0.50 500 40% 30% 4% 3% 0.00 0 20% 2% 10% 1% Site Number 0% 0% Pedestrian Crash per Marked Crosswalk Vehicle Hourly Volume Pedestrian Crossings per Hour Yielding Compliance Rate Site Number Pedestrian Evasive Maneuver Rate Slight relationship between hourly vehicle volume and total pedestrian crashes Slight inverse relationship between pedestrian evasive maneuvers and yielding compliance 16
RRFB-1 In Street-1 In Street-4 Marked-3 Marked-6 Marked-9 Marked-12 Marked-15 Marked-18 Marked-21 Marked-24 Marked-27 Not Marked-1 Not Marked-4 Pedestrian Crashes Vehicle Hourly Volume and Pedestrian Crossings per Hour RRFB-1 In Street-1 In Street-4 Marked-3 Marked-6 Marked-9 Marked-12 Marked-15 Marked-18 Marked-21 Marked-24 Marked-27 Not Marked-1 Not Marked-4 Yielding Compliance Rate Pedestrian Evasive Maneuver Rate Preliminary Results: Node Analysis Pedestrian Crashes per Site Yielding Compliance and Evasive Maneuvers 5.00 4.50 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 2500 2000 1500 1000 500 0 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% 0% Pedestrian Crashes Site Number Pedestrian Crossings per Hour Vehicle Hourly Volume Yielding Compliance Rate Site Number Pedestrian Evasive Maneuver Rate Some relationship between ped crashes and volumes (vehicle and ped) Strong relationship between type/features of crosswalk and compliance/evasive maneuver rates 17
Yielding Compliance Rate Pedestrian Evasive Maneuver Rate Preliminary Results: Node Analysis 100% 18% 90% 80% 70% 60% 50% 40% 30% 20% 10% 16% 14% 12% 10% 8% 6% 4% 2% 0% PHB RRFB In-Street Marked Unmarked Treatment Type 0% Yielding Compliance Rate Pedestrian Evasive Maneuver Rate Sites with PHB and in-street pedestrian crossing signage exhibit higher yielding rates 18
Preliminary Conclusions: Factors Associated with Driver Yielding and Evasive Maneuvers Increased Driver Yielding Compliance Lower Vehicle and Bicycle Volumes Marked Crosswalks Pedestrian Crossing Signs Two-Way Left Turn Lanes Not Present Decreased Rates of Evasive Maneuvers Pedestrian Crossing Signs Crosswalk Width Limited to 50 ft. Maximum Two-Way Left Turn Lanes Not Present No Shoulders No Medians 19
Preliminary Conclusions: Driver Yielding Compliance by Lane Proximity and Median Presence Laneage Driver Yielding Compliance with Median Presence Driver Yielding Compliance without Median Presence Near Lane 58% 60% Center Lane 75% 79% Far Lane 82% 79% There is no significant difference between driver yielding rates with or without a median present 20
Upcoming Work Midblock crossings SPF development Volume (ped and vehicle), site characteristics, and behavioral information Node vs. segment Negative binomial regression and other techniques may be investigated Signal and stop controlled intersection Evasive maneuvers, ped entry on red, ped completion on red, RTOR conflicts, driver yield rate on permissive turn SPFs (node only): volume (ped and vehicle), site characteristics, and behavioral information 21
Evasive Maneuver for Signal Controlled Intersection 22
Questions? Timothy J. Gates, Ph.D., P.E., PTOE Associate Professor Michigan State University Department of Civil and Environmental Engineering voice: 517-353-7224 gatestim@egr.msu.edu 23