1. Steep Streets Speed Humps on Gabriel Ho, P.E.
San Francisco MTA
Traffic calming program – lead engineer
In 2017, we updated our traffic calming policy for speed hump placement based on a study of how speed humps perform on steeper streets. Today, I’m here to tell you more about that study.
Thank co-authors
Kate Beck, researcher at Berkeley SafeTREC
Casey Hildreth, principal planner at MTA
Gabriel Ho, P.E.
Livable Streets Subdivision
2019 ITE Western District Meeting
Co-Authors:
Kate Beck, SafeTREC
Casey Hildreth, SFMTA

2. Traffic Calming Program by the numbers
100+ applications per year
100+ speed humps or variants installed per year
Speed humps reduce 85th %ile speeds by 6 mph
Streamlined our TC program, transitioned away from area-wide projects to single-block projects that require a signed petition from residents of the block.
We now receive over 100 block applications from residents in districts throughout the city.
Install 100 plus humps, cushions, tables, raised crosswalks; asphalt vertical deflection devices
2018 evaluation -> speed humps are single most effective measure for reducing speeds -
Average 6mph reduction of 85th percentile speeds on both flat and steep streets
Statistically significant results

3. Traffic Calming Program Acceptance Criteria
Proximity to schools, parks, senior centers 
Speed & Volume thresholds 
Restricted on arterials, multi-lane roads, grades >12%
8-12% case-by-case only
Proximity to community centers
Measure 85th percentile speeds and compare to the speed limit

4. Roadway Slopes 9-12% >12%
9-12% >12%
Grade criteria has been limiting us, as San Francisco is truly a hilly city.
Over a quarter of blocks have a slope above 8%, most of which are residential areas eligible for traffic calming.
Every year, we receive applications from residential blocks in this steeper range of slopes.

5. Question: 
Do speed humps on steeper roadways present traffic safety or driver/passenger comfort issues? 
So we asked ourselves this fundamental question:
Do speed humps on steeper streets actually pose safety or comfort issues?

6. Initial Research Literature Review
ITE Guidelines for the Design and Application of Speed Humps (1993)
Interviews with other jurisdictions
Portland
New Jersey DOT
Delaware DOT
Based on our research, we found that most jurisdictions around the country use 8% or less as a maximum slope;
Current ITE and NACTO guidance on the subject also refer to an 8%, with all sources leading back to the 1993 ITE Speed Hump guidelines which cite 8% as a recommended maximum.
However, we couldn’t find any specific research, testing, or evidence to substantiate this 8% threshold.
We did find a few jurisdictions with thresholds above 8% . Portland and New Jersey allow tables up to 12%. Delaware developed a speed hump with an asymmetric profile specifically for steeper streets but it was never built due to the complexity of forming asphalt pr

7. Field Test Locations: Four blocks with existing speed humps
Blocks with slopes of 1%, 4%, 8%,13%
Crash analysis
Resident feedback
So we decided to conduct our own study and field test speed humps already installed on San Francisco streets;
We selected speed hump locations to represent the widest range of grades available. On closer examination, we actually located a speed hump on a 13% grade in the Bernal heights neighborhood.
On these blocks, we reviewed crash histories and didn’t find anything interesting. We also didn’t receive any public feedback on speed humps on these blocks that mentioned the slope of the street posing an issue.

8. Field Testing Procedure:
Speed 15, 20, 25, 30mph
uphill and downhill over speed hump
Driver rating of vehicle control 
Driver/Passenger rating of comfort 
Measure acceleration during deflection
We developed a field testing procedure to evaluate vehicle control, passenger comfort, and the forces experienced in the car while traversing the speed humps at various speeds.
MTA staff drove a city vehicle over the speed humps on each of the four blocks at varying speeds, both the uphill and downhill directions.
After each run, the driver and passenger each rated their comfort., and the driver also rated whether he was able to maintain control of the vehicle during and after deflection at the speed hump. Both of these ratings diminished as speeds increased, but at all speeds and roadway grades, the driver was able to maintain control of the vehicle.
Lastly, we wanted to collect objective data of these vehicle dynamics.

9. Field Testing Procedure:
Physics Toolbox Accelerometer
g-Force in x, y, z axes
Recording 200Hz
Output to .CSV
Averaged into 0.04sec bins
During these runs, the passenger used the accelerometer and an app on her smartphone to record g-forces in the cabin.
Data was recorded at 200 Hertz and output into Excel, where we averaged the data into bins of 1/25th seconds to smooth out the curves and reduce the noise in the data.
And for simplicity, we only analyzed the total combined force in the x y z axes, ignoring the directionality of the force for simplicity.

10. Accelerometer Reading
g-force measured in passenger seat
Here’s an example of the accelerometer data graphed out. Again this is a measurement of g-forces inside the cabin of the car.
The first spike is the peak acceleration as the front axle of the car is deflected vertically.
The following dip is the moment of weightlessness as the front axle is at its peak height and beginning to fall back down. A measurement of 0 at this point would mean the car is airborne.
For each run, we measured the difference between this first peak and valley as an approximation of how jarring that particular speed hump impact was.

11. Field Test Findings: g-force increases with speed, not slope
This graph summarizes our accelerometer data for all roadway grades and speeds tested.

12. Conclusions: Speed humps on various grades perform similarly:
Crash history
Comfort
Deflection forces
Our small study demonstrated that speed humps on various roadway grades perform similarly in terms of crash safety, comfort, and deflection forces.

13. Policy Change
Increase standard max grade to 13%. >13% to be considered case-by-case
Emphasis on advanced warning markings and advisory speed signs
We have about a half dozen locations now with speed humps on 12/13% grades.
Since forces at the speed humps are dependent on vehicle approach speed, we’re putting extra emphasis on giving drivers enough advance warning in order to reduce their speed.
Marking our speed humps with an extra set of chevrons.

14. Research Next Steps Effects on Bicycles, Heavy Vehicles Cold climates
Slopes >13%
Larger sample size for more statistically rigorous analysis
More precise measurement tools
Larger vehicles with different suspension designs and handling characteristics
We have only a moderate climate in the Bay Area, so this study may result in different results in climates with ice or snow in limited traction scenarios.

15. Thank You
Gabriel Ho SFMTA Livable Streets sfmta.com/traffic-calming

16. SF Standard Speed Hump