[Quietcars] NHTSA report: DOT HS 811 204 September 2009

Thu Oct 1 15:38:56 UTC 2009

The National Highway Traffic Safety Administration (NHTSA) released "Incident of Pedestrian and Bicyclists Crashes by Hybrid Electric Passenger Vehicles" a technical report that can be downloaded. We've looked at this report and I would recommend anyone with an interest in this subject to get a copy from:

http://www-nrd.nhtsa.dot.gov/Pubs/811204.PDF

This study is based upon looking at all accident data, not just fatalities, in the years 2000-2007 from 12 states that to a greater or lessor extent record accident. It is a fine report and includes the necessary disclaimers about the low numbers of hybrid electric vehicle (HEV) versus Internal Combustion Engine (ICE) vehicle accident data. But this is how it looks to us.

The critical data are located in tables 6a "Vehicle maneuver prior to pedestrian crashes HEVs and ICE vehicles" and 3d "Vehicle maneuver prior to crash." It wasn't clear from the language but the number of accidents in table 6a are divided by the number of maneuvers in table 3d. This generated the "Incident rate of pedestrian crashes" for both HEV and ICE vehicles.

Now table 6a shows two areas where the distribution within either HEV or ICE vehicles differ:

a.) Making a turn - 19 events or 1.8% HEV vs. 698 or 1.0% ICE
b.) Backing - 7 events or 5.3% HEV vs. 261 or 2.9% ICE

What these numbers discuss is the distribution of accident maneuvers within each group, HEV or ICE vehicles, and there is a difference. In the case of the 7 backing events, they did not flag that as statistically significant. That they tagged the 19 turning events as significant is a little too close call for my taste but the 1,061 records in table 3d is large enough. They only had 132 backing accident records and having 7 events makes calling this statistically import. For example, consider what happens if the number of backing events changes by +/- 1:

6 events = 4.5%
7 events = 5.3%
8 events = 6.0%

A single backup event makes a substantial change in the percentage, which is why more data, as is called out in the report, is important. But table 3d "Vehicle maneuver prior to crash" carries and importance not really addressed in the report.

The "Vehicle maneuver prior to crash" data breaks down the percentages of maneuvers for both HEV and ICE vehicles and the two of interest, "Making a turn" and "Backing" are 15% of all HEV maneuvers:

"Making a turn" - 13%
"Backing" - 2%

In fact "Going straight", the single largest source of accident data had all but identical percentages for HEV, 9%, and ICE vehicles, 8%. Yet "Going straight" is 44% for HEV and 47% for ICE vehicles ... in absolute terms, the HEVs are less likely than an ICE vehicle to have a "Going straight" accident in this 12 state, 2000-2007 data.

What this report tells us is only 15% of all HEV maneuvers have any difference between HEV and ICE vehicles and the largest, 13%, occur while turning. Upon reflection, this makes sense especially if we can get left and right turn data rates.

In the USA, we drive on the right hand side of the road and right-turn on red after stop is all but universal. So when a vehicle approaches a stop light and plans to make a right turn, the driver looks to the left for on-coming traffic and then switches to looking to the curb and then turning right to make their turn. 

A pedestrian on the curb may also be looking to the left at on-coming traffic and then decide to step out (aka. jay walk) but change their attention to getting off the curb and the far side of the road. The pedestrian can move faster than the car can react especially if crowds or the car some 6-8 ft. behind the crosswalk, is not seen by the pedestrian. Even with the turn signal on, only the driver has the audio 'click click' in the cabin, there is no external 'click click' to alert anyone that the car intends to make a left turn and should be watched by the pedestrians. This problem is compounded for those with limited or no sight.

The paper calls out for more data and it occurs to me there is a natural lab already out there ... hybrid electric taxi cabs in New York and California cities. These mix of Prius and Ford Escape Hybrids are a natural lab and it would relatively easy to install digitial video recorders that save the window around every left and right hand turn. Then equip half of the vehicles with audio alerting, external turn signals. There are after-market backup lights that have a standard socket that should be able to replace the turn signal lights and generate an external noise. Then run the experiment for 6 months.

This type of experiment would provide turning incident, video records so both close calls and accidents can be tracked and analyzed. The Japanese and some USA universities have used similar video recording to gather data like this before.

I do have a couple of concerns about this report and believe it needs to be revised. Having gotten the accident data from 12 states, the HEV and ICE vehicle registrations were not tabulated to give us an absolute risk factor either by vehicle type or miles per vehicle type. This is important because if HEV vehicles have say half the rate of accidents as ICE vehicles, then it means the other 85% of accident types are half the rate for HEV vehicles compared to ICE vehicles and the "Making a turn" and "Backing" accident rates are identical between the HEV and ICE vehicles. Also, I would like to see 'right turn' versus 'left turn' numbers, if it is available in the data. My speculation is the 'right turn' case is likely to be the one needing attention and perhaps only right-turn signal noise alerts are needed for both HEV and ICE vehicles.

I do recommend this report and reading the data tables. There is quality data here and it provides a roadmap to where investigations need to go next.

Thanks,
Bob Wilson

_________________________________________________________________
Microsoft brings you a new way to search the web.  Try  Bing™ now
http://www.bing.com?form=MFEHPG&publ=WLHMTAG&crea=TEXT_MFEHPG_Core_tagline_try bing_1x1