PASADENA — It was to have been the latest addition to the city's high-tech traffic-control arsenal, a computerized surveillance camera that 24 hours a day could snap pictures of cars zipping through red lights.
But after several months of testing at two busy intersections, enthusiasm for the device, called a Multafot, has stalled.
Police were dismayed to find that the red-light sensors--which along with the city's controversial photo radar system were firsts of their kind in California--were being triggered by legal driving maneuvers nearly 95% of the time.
'Lot of False Photographs'
"We were getting an awful lot of false photographs," said Lt. Robert Huff, adding that no citations were handed out during the tests. "I wouldn't say the problems are insurmountable . . . but there's a number of things that we have to take into account that we didn't know about."
At the two intersections--California Boulevard and Hill Avenue, and Fair Oaks Avenue and Union Street--every three or four days the magnetic sensors placed in the pavement indicated that about 200 cars were running red lights. But after developing the film, an average of only 10 to 15 of the 200 cars photographed were actually filmed in illegal moves, Huff said.
The device, which takes two rapid-fire photos every time it is triggered, showed that most of the cars had either stopped with their noses sticking a bit too far into the intersection or, in making a left turn, were forced to pass through a red light after waiting for oncoming traffic to come to a halt.
"Technically, those could be violations, but we would never issue a citation because they're not really endangering anyone," said Huff, head of the Police Department's traffic division. "The idea was to get the actual red-light runners."
However, the firm that sells the $45,000 Multafot, which has been used extensively in Switzerland and West Germany, said the problem in Pasadena was with the choice of intersections and the positioning of the sensors.
Had city officials been more willing to fine-tune the location of the sensors or to select streets with more regimented traffic flows, the Multafot would have detected violators more efficiently, said Robert P. Umbdenstock, president of Multanova/RPJ Inc., based in San Anselmo, the U.S. distributor of the devices.
"The equipment did what it was asked to do," he said. "Unfortunately, it was being asked the wrong questions."
City officials, however, defended the choice of intersections, saying they have a high incidence of right-angle collisions. Also, positioning the sensors in just the right spot of the road proved more difficult and expensive than city engineers had expected. Every time an adjustment was needed, a work crew had to saw into the street and reposition the devices in each lane.
An Incomplete Picture
Officials were further concerned that the camera, which was mounted on a nearby street post, wasn't capturing the whole picture. For each car detected, the two photos showed the signal lights, the vehicle, the license plates and, in some tests, the driver.
But neither of the two photos showed the car before it had crossed the stop line; they showed it only in the intersection. Without an officer there to observe the violation, police thought they would need an expert witness to interpret the pictures in court testimony.
Some of the same difficulties have been encountered in New York City, where officials also have been experimenting with the Multafot and two other brands of surveillance cameras.
But New York officials, who say the sensors were triggered by legal maneuvers about 50% to 90% of the time, are still going to solicit bids later this year for a firm to conduct a one-year trial run with four cameras.
"The main problem was the sensors being overly sensitive," said Emily Goodman, a spokeswoman for the New York City Transportation Department. "But we feel that they can be finessed and that the percentage of usable photos can be greatly improved."
Although Pasadena officials cautioned that a final report on the devices is still several weeks away, it is clear that the high hopes expressed in January, before the testing, have dimmed.
Unlike the photo radar--which was criticized by some drivers as an insidious "Big Brother" approach--the red-light sensor was expected to gain a more favorable reaction. Officials figured that, whereas nearly everyone speeds once in awhile, few condone running a red light.
But city engineers are now proceeding with caution.
"It has potential, but it needs further development," said Lawrence Tai, the city's chief traffic engineer. "Right now, I don't think it's for our city."
That came as a disappointment to some who live and work near the two test intersections.
"It would have been a good idea," said Nancy Wilder, a receptionist at Caltech's Athenaeum, near the intersection of California and Hill. "We hear crashes out there all the time."
Norman Hill, a parking lot attendant at the corner of Fair Oaks Avenue and Union Street, agreed. "This is a bad corner," he said. "It would have put a lot of food for thought in people's heads before they went through a red light."
Meanwhile, Pasadena's photo radar system, after surviving an initial legal challenge, has been working with merciless efficiency. Since it was installed last June, it has dished out more than 8,500 speeding tickets.