Doug Finlay

All of us hate being stuck in traffic - and wonder if there was a better way to go.  SpeedInfo has spent the past two years developing a system to provide drivers with radically better traffic information. We give the driver a color coded map in the car that shows how fast cars are going on all of the major highways in the area. At a glance, the driver can see which roads have the slowest traffic, and make an informed decision about when and where to drive.

We supplement public highway speed information with data from low-cost Doppler speed sensors that we have developed. Our sensor is solar powered, and uses cellular data connections to backhaul its data. It measures speeds every 15 seconds, and reports the average speed once a minute. After the data is validated by our servers, it is transmitted to the car over FM radio. The design allows us to integrate the data with navigation systems and newer car radios for less than $10.

Besides saving drivers time, money, and gasoline, the system also helps us use our highways more efficiently. When cars avoid the most congested areas during peak times and choose alternative routes, we do a better job of balancing the load across the highway network. Fewer cars stuck in traffic jams means less air pollution. Drivers can also respond better in an emergency by taking the best route to avoid problem areas.

Doug Finlay, the founder and CEO of SpeedInfo, will give an overview of the company's technology and product rollout timetable.

Doug Finlay began his talk by explaining that $70 Billion is wasted every day in the USA on traffic congestion, one way or another. SpeedInfo was founded on the idea that if drivers could get more reliable and timely information on where traffic bottlenecks were, they could use alternative routes to go around them. He pointed out that just reducing the number of people trying to squeeze past an accident by 10% could make a big difference in net delay times.

He then explained a bit about the current traffic information system. It is based on street sensors that have been there for 20 years, helicopter traffic reports, and things like volunteers calling in problems as they become aware of them. There are many gaps in the street sensor network. Some of them were caused by a lack of available phone and power cabling to connect I/O to (for example I-280 between San Francisco and San Jose), and others are caused by maintenance type issues. One of the problems with expanding it is the expense of cutting holes in the concrete to put sensors in, which works out to be very disruptive, expensive, and time consuming.

The SpeedInfo solution for data collection is an extruded metal tube about four inches across and two feet long with a solar panel above it that can be easily strapped to a light or telephone pole with metal bands. Inside the tube is a set of sealed batteries that can drive the electronics for ten years, a sensor that uses technology similar to that in police radar guns to sense cars going by, and a proprietary circuit board. Finlay explained that the software in the box "glopts" the radar signal into a blob for each car, and then measures how far it moves from one frame to the next to figure out how fast the traffic is moving. This information is then uploaded via a standard cell phone modem to SpeedInfo's server, where it is combined with information from others to create a gestalt for the commuter.

The SpeedInfo solution for information distribution is a box about an inch thick, maybe 5" by 7". Most of the front is taken up by a flat panel display, a map that can be scrolled to where the user wants to see the traffic situation. The plan is to get a station with a high powered FM transmitter like KQED's to broadcast the information, so that it can be picked up throughout the area. Finlay explained that many of the companies doing geosynchronous mapping equipment are interested in combining SpeedInfo into their systems so that car computers can simply pick the best way to get somewhere. He passed around the display, and everybody liked the concept.

At the moment SpeedInfo is beta testing their traffic sensors. He pulled up the information coming off of ones on I-405 in Orange County, 16th Street in Washington, DC, and another in New York City. We could see that I-405 speeds varied from about 55 to 80 MPH, with most of the readings in the 60-70MPH range. 16th Street moved much slower, about what you would expect. The test units have been on location since last spring. Finlay expects to be ready to instrument the Bay Area next spring at the earliest. Partly this depends on whom he can make a deal with to get it done. Both CalTrans and 511 have expressed an interest in the project, but nobody has committed funding to it yet.

During Q&A a lot of interesting points were discussed.

The radar sensor has a diode that broadcasts at 24.125 GHz.

Each sensor will cost $600 installed, including the labor of the guy in the cherry picker strapped it to the pole.

At this point in time he has patents pending on several aspects of the sensor.

One of the reasons this hasn't been done before is that cell phone technology wasn't widely available until recently. This is important because that is where they save a lot of cost by eliminating dedicated wiring from the system.

SpeedInfo would much rather be an information wholesaler than a retailer. Because of this they are talking to other companies that already have the shelf space and brand recognition to sell to the driving public about integrating this into their product line. So far nobody has gone for it yet though.

The main change required for northern latitudes would be a bigger solar panel.

Rain is not be a problem for the system, although hail might be. One difficulty SpeedInfo is having is that hailstorms are not common enough that any of their sensors have gone through one to test the theory.

Tian Harter