New speed cameras won't eliminate errors, radar experts say

Baltimore transportation officials have set high expectations for the city's new speed cameras, telling state lawmakers the devices won't be susceptible to errors that plagued the system over the past three years.

"We won't have this problem moving forward," said Barbara Zektick, acting deputy transportation director, at a recent briefing for the city's legislative delegation.


"The new cameras have tracking radar," said Frank Murphy, the agency's acting director.

But radar experts say tracking radar isn't necessarily the cure-all it might seem. They predict the new cameras will reduce — but not necessarily eliminate — the kinds of errors that have cast a shadow over the city's automated enforcement system. Those errors, many of which came to light in a Baltimore Sun investigation, have included generating excessively high speed readings and tagging the wrong vehicle as the speeder.


Speed cameras will be on the agenda in Annapolis this week. A wide-ranging reform bill sponsored by Sen. James Brochin, a Towson Democrat, is scheduled to be heard Wednesday at 1 p.m. by the Senate Judicial Proceedings Committee. One provision would more explicitly bar the so-called "bounty system" that pays contractors based on the volume of citations.

While city officials believe tracking radar will produce more accurate results, experts say a rigorous human review process and frequent calibration of the cameras will remain essential if the city wants to move toward Mayor Stephanie Rawlings-Blake's goal of achieving "a zero-error program."

"In the laboratory, in the hands of a skilled scientist, these things could be tested and proved reliable," said Christopher Davis, an electrical and computer engineering professor at the University of Maryland. "But put out on the street and left there for months at a time, I wouldn't have confidence they'd give reliable readings 100 percent of the time."

City officials have emphasized that, while they trust the new cameras, they will also beef up the process of reviewing and verifying the automated cameras' citations before they are mailed to motorists for payment of the $40 fine. A key change will be at the Police Department, where officers had been checking up to six tickets per minute. The department says it will double the number of officers available to review citations and have supervisors spot-check their work.


After initially planning to continue using its existing 83 cameras, the city decided to replace them as it switched to a new vendor, Brekford Corp. of Anne Arundel County, Jan. 1. The new cameras cost the city $25,000 apiece or about $2.1 million in all. Some are in operation, and all are scheduled to be in place by late March.

In documents filed with the city last year, the company said Baltimore "will be the first municipality to be offered this technology by Brekford."

As its name suggests, tracking radar tracks objects — in this case, cars and trucks — as they pass a camera. Brekford says it will help ensure accurate readings and avoid citing the wrong vehicle.

The new system is an improvement because the previous models took more of a radar snapshot and were more prone to various errors, said Fernando Berra III, a photonic engineer in private industry with 20 years of experience in laser and radar systems. With multiple radar readings, it will be apparent if there's an error because the speeds won't be consistent, he and other experts say.

"If done properly, it would absolutely minimize the issues," Berra said of the tracking radar. Still, he said, while the technology is reliable, trained technicians will be needed to examine the results.

One challenge that will remain even with tracking radar is correctly determining who is speeding when multiple vehicles are on the road, as often happens in the city, said Jin Kang, chairman of the Electrical and Computer Engineering Department at the Johns Hopkins University.

"The problem is that they will still have a difficult time analyzing which [radar] signal corresponds to which vehicle," Kang said. "It requires somebody sitting there analyzing the results."

The Police Department's practice of having a single officer review several tickets per minute "makes no sense," Kang said, adding: "That person has to spend at least a minute or five minutes per photograph to make an error-free analysis."

Davis, the Maryland professor, said the only surefire way to check the tracking radar's accuracy will be for a reviewer to have a way of comparing the speed reading with physical evidence of how far the vehicle traveled. Because speed equals distance over time, he explained, it's possible to double-check the radar using time-stamped photos and pavement markings.

"If it was me caught speeding, I'd like to see two photographs showing me moving a certain distance," he said. "Otherwise there's still a possibility they captured the wrong vehicle or got a spurious effect" such as a "mirror effect," in which radar bounces off some other object before hitting a vehicle, producing a speed reading far higher than the vehicle's actual speed.

To facilitate such verification, Davis said, the city and all Maryland jurisdictions that use speed cameras should paint white lines on the road at set intervals — every 10 feet, say — near each camera. Brochin's bill would mandate such line painting.

In the same vein, he believes governments should ensure that the two photographs required by state law show the precise time each was taken — down to the fraction of a second. While the city already does this, Baltimore County, Howard County and the State Highway Administration generally give motorists time stamps that are rounded just to the second, making it impossible to fact-check the speed. Brochin's bill would also require detailed time stamps.

The Sun used that method to document erroneous speed readings from six city speed cameras. Because the city does not have lines painted at set intervals, markings such as lane dividers and cracks were used to measure how far a vehicle went over the split-second between the two photos.

The Sun also reported that a Mazda had been cited for speeding while stopped at a red light. Xerox State and Local Solutions, the city's previous contractor, acknowledged that that camera and four others had error rates of 5 percent, prompting the city to take them offline. Xerox said high-walled trucks played a role in causing "radar effects" that produced inaccurate readings.

Rawlings-Blake has defended the safety benefits of the city's speed camera program, which has generated more than $48 million in paid fines since 2009, amid the revelations that errors got through a multi-layer review meant to catch mistakes. The vendor does the initial review before sending citations to the police for authentication. Only if police approve a ticket can it be mailed to a motorist.

Speaking to reporters last month, Rawlings-Blake said "our focus, as always, is on safety," adding, "This is about making sure that our young people, who are going to and from school, have safe passage." Speed cameras are allowed in school zones and highway work zones; $40 tickets can be issued to vehicles clocked going at least 12 mph over the limit.

Maurice R. Nelson, managing director of Brekford, did not respond to requests for comment. But he has said his firm's cameras would bring marked improvement to speed enforcement in Baltimore.


"The old radar cameras have not progressed with technology," he said in December, adding that tracking technology can focus on and follow a specific car, reducing machine-created errors. "If you're using the old radar cameras and it's picking up something that's not the car in the photograph, you leave yourself open to errors."


While the city has described Brekford's cameras as "state of the art," Kang, of Hopkins, said tracking radar technology is more than 70 years old. "It's been around since before World War II," he said.

In basic terms, radar uses radio waves to determine an object's speed by calculating changes in frequency as the object moves toward or away from a transmitter. This change is known as the Doppler shift.

Other speed cameras, including those used in state highway work zones and in Baltimore County, use laser technology called "lidar." Instead of radio waves, lidar uses light pulses to gauge speed.

Though more accurate than radar, lidar can be thwarted by fog because it relies on light particles, said Kang, who has researched lidar. "That's the problem with lidar: You don't get 100 percent detection. But you will reduce the error. With radar, you will always get the signal, but the signal may be wrong."

"The question," he said, "is which poison do you want to take?"

Berra, the photonic engineer, said: "If everything is going well, both radar and lidar are extremely accurate. If you have ideal conditions, one car on the road, and the system calibrated, the error rate should be zero."

But by their very purpose, speed cameras are often put on major roads with lots of traffic. Davis said being outside for weeks or months on end can cause misalignment that can, in turn, impair accuracy.

While state law requires every speed camera to be calibrated to the manufacturer's specifications once a year, Davis says that's "way too long" between calibrations. And Kang said the only real way to test accuracy is to drive a vehicle past it at a known velocity, difficult in a city with 83 cameras.

Speed camera proponent David Kelly said in hindsight it might have been wise for the city to use tracking radar when it launched its speed program in 2009. But he's glad the switch is occurring now with a new vendor onboard.

"Now they do have a chance to start over and get it right from the beginning," said Kelly, executive director of the National Coalition for Safer Roads. "This is a real natural time to say, 'Let's make sure we've got this right.'"

Recommended on Baltimore Sun