Feasibility of Using Advanced EDRs for Assessing Active Safety Systems H. Clay Gabler Kristofer D. Kusano Virginia Tech 7 November 2013 Center for Injury Biomechanics COLLEGE of ENGINEERING C I B
Challenge: Determining Driver Actions, Pre-Crash Common Data Sources: 1. Crash Databases 2. FOT, Naturalistic Alternative Approach: Event Data Recorders
Event Data Recorders (EDRs): The Black Box Store vehicle data after crash Pre-Crash Data: Speed Accelerator Brake ON/OFF Advanced EDRs: Steering To PC Bosch Download Tool EDR
Objective Determine feasibility of using advanced EDR data to evaluate active safety systems? Lane Departure Warning Systems Forward Collision Avoidance Systems
Data Source: Crashworthiness Data System NHTSA s NASS/CDS 5,000 in-depth investigations per year Tow-away crashes Download EDR, when available
EDR Downloads by NASS/CDS Year 1000 6,467 Total Downloads 800 EDR Downloads 600 400 Bosch CDR Tool GM Ford Chrysler Toyota 200 0 Case Year
Example EDR Data NASS/CDS Case 2011-11-148 2009 Ford Escape Vehicle Speed (mph) 40 35 30 25 20 15 10 5 0-5 -4-3 -2-1 0 100 80 60 40 20 0 Speed (mph) Accelerator Pedal (%) Brake ON/OFF Time Before Crash (s)
Advanced EDR Data NASS/CDS Case 2011-11-148 Ford Escape Steering Input (deg) 120 80 40 0-40 Steering Yaw Rate (deg/s) 15 10 5 0-5 -10-15 Yaw Rate -80-5 -4-3 -2-1 0-20 -5-4 -3-2 -1 0 Time Before Crash (s) Time Before Crash (s)
Reconstruction Approach Develop Vehicle Model Create Simulation using EDR data Validate using EDR Dynamics Steering Public Specifications NHTSA Rollover Tests -5-4 -3-2 -1 0 Time EDR Steering EDR Speed Yaw-rate Lat./Lon. Acceleration Simulate crash with System (e.g. LDW)
Simulation without LDW Impact Location Road geometry from NASS/CDS scene diagram Yaw Rate (deg/s) 20 15 10 5 0-5 -10-15 -20 EDR Simulation -5-4 -3-2 -1 0 Time Before Crash (s)
Simulation with LDW Impact Location Steering Input (deg) 120 80 40 0-40 -80 Lane Touch EDR With LDW -5-4 -3-2 -1 0 Time Before Crash (s) Reaction Time
Forward Collision Avoidance Systems
Rear-end Collisions Single most common collision type in U.S. Forward Collision Avoidance Systems mitigate or prevent rear-end collisions 1 1 http://www.caranddriver.com/photos-08q3/217767/fords-forward-collision-warning-technology-illustration-photo-217777
Design of Pre-Collision System Activation Timings Time Collision Forward Collision Warning (FCW) Questions: When do drivers brake? Driver brake force? How many don t brake? Dynamic Brake Assist (DBA) Increase Driver Brake Effort Crash Imminent Braking (CIB) Automatic without driver braking
Objective Quantify driver avoidance maneuvers in rearend collisions using Event Data Recorders Data Source: NASS/CDS crash database Focus: Tow-away Rear-end Collisions
Example Crash with EDR Data NASS/CDS 2006-74-098 Vehicle Speed (kph) 90 80 70 60 50 40 Speed Throttle Brake 1 0.8 0.6 0.4 0.2 Brake ON/OFF and Throttle % 30-5 -4-3 -2 0-1 Time (s)
Computing Time to Collision (TTC) at Brake Initiation v 1 V 1,0 a v 2 x 0 Δt 1. TTC at Braking: TTC brake = t s 1 2V 1,0 at s 2 2. Average Deceleration from Vehicle Speed
Selected Cases NASS/CDS 2000-2011 EDRs from GM, Ford, and Chrysler vehicles with valid pre-crash data Group N All Crashes 57,223 Rear-end Crashes 6,748 Rear-end, with pre-crash EDR 143
Pre-Crash Brake Application n = 143 Crashes 28% 72% No Brake Brake Injured (MAIS2+) 65% 35%
Average Brake Deceleration vs. Brake Initiation Time 1 n = 88 Mean: 0.4 g Brake Deceleration (g) 0.8 0.6 0.4 0.2 0-5 -4-3 -2-1 0 Time of Brake Initiation (s)
Distribution of TTC at Brake Initiation % 0.35 0.3 0.25 0.2 0.15 0.1 0.05 n = 84 Histogram Lognormal Fit Mean: 1.18 s 0 0 1 2 3 4 5 TTC at Brake Initiation (s) Log-normal approximation (suggested in literature) is appropriate
Limitations Coarse sample rate of EDR pre-crash data GM 1 Hz most common in sample Part 563 specifies 2 Hz Some automakers up to 10 Hz Limited sample size
Outlook: EDRs for Active Safety System Evaluation EDR Pros: Low-cost Increasing availability in fleet Large sample of serious collisions EDR Cons: No video, difficult to determine distraction Low sample rate (Part 563-2 Hz) EDR data - a promising evaluation tool
Acknowledgement This project is sponsored by Toyota s Collaborative Safety Research Center www.toyota.com/csrc
Feasibility of Using Advanced EDRs for Assessing Active Safety Systems H. Clay Gabler Kristofer D. Kusano Virginia Tech 7 November 2013 Center for Injury Biomechanics COLLEGE of ENGINEERING C I B
Findings 143 EDRs from striking vehicle 28% had no braking V 12 L 0 TTC of 1.2 s for those that apply brake