Reporting from Washington — An ambitious, multibillion-dollar project to hot-wire the new Southwest border fence with high-tech radar, cameras and satellite signals has been plagued with serious system failures and repeated delays and will probably not be completed for another seven years -- if it is finished at all.
The system, originally intended to be completed next year, languishes in the testing phase in two remote spots of the border in Arizona.
There, the supposedly state-of-the-art system combining sensor towers, communication relay systems and unattended ground sensors has been bogged down with radar clutter, blurred imagery on computer screens and satellite time lapses that often permit drug smugglers and undocumented workers to slip past U.S. law enforcement agents, government officials candidly admit.
"It was a great idea, but it didn't work," said Mark Borkowski, executive director of the electronic fence program at the Homeland Security Department.
"One of the kickers was that these radars had too many problems with clutter," Borkowski said. "Wind moving a tree shows up on the radar. And if you have too much of that, how do you find the person in the clutter? Same with cameras. The image is blurry."
The problems have prompted Homeland Security Secretary Janet Napolitano to order a department-wide assessment of the technology project once billed as the capstone to the controversial 2,000-mile combined physical and electronic border fence.
Borkowski acknowledged in an interview that the government and its main contractor, Boeing Co., had made a series of mistakes since announcing in 2005 the plan to build sensor towers and radar scans alongside the new border fence.
Although they remain hopeful the problems can be fixed, he cautioned that the technology ultimately might not cover the entire border.
"It turned out to be a harder technological problem than we ever anticipated," Borkowski said. "We thought it would be very easy, and it wasn't."
He said the government was primarily to blame for not being more specific in its contract with Boeing. But, he added, "we have a border we've got to secure, and technology has to be a key part of the plan. It's not there. So what do we do in the meantime?"
Tim Peters, vice president of Boeing Global Security Systems, which is handling the project, said his company remained dedicated to correcting the problems, acknowledging that in the Arizona testing sites "our customer's and our expectations were not initially met."
Although the testing has taken longer than planned, costing about $20 million so far, Peters said a much-improved high-tech system would evolve.
"We have every reason to be confident," he said, that "future deployments will do exactly what they are designed to do -- provide the Border Patrol with an unprecedented level of situational awareness to enhance border security and improve agent safety."
But the delays and breakdowns have prompted critics on both sides of the border debate to call for fresh ideas to improve security along the frontier with Mexico.
Ira Mehlman, a spokesman for the nonprofit Federation for American Immigration Reform, said his group thinks that more and higher conventional fences and old-fashioned border agent surveillance are more reliable than the technology.
"Instead of spending a lot of time reassessing," Mehlman said, "they should get out there and do the sorts of things we know work effectively to get control of the border, such as double fencing and more manpower."
That thought is not lost on Napolitano, who in ordering the reassessment said that "a comprehensive border security strategy must include an effective combination of physical infrastructure, manpower and technology."
The Secure Border Initiative was once heralded as a sweeping plan to throw up a physical barricade and high-tech equipment to keep drugs, guns and undocumented workers away from what has for years been criticized as a far too porous border.
Borkowski said nearly all of the planned physical fencing was in place along about 620 miles of terrain "where we think we need it." He said an additional 30 miles still must be fenced in the Rio Grande Valley in Texas. The cost for the physical fence was $3.4 billion.
The high-tech phase, known as SBInet, carries a price tag upon completion of about $8 billion. It was initially envisioned for the entire 2,000 miles of the border.
The technological wizardry was designed to send signals and video images to Border Patrol command centers, much like an aircraft control center, so agents could be quickly dispatched to investigate border breaches.
But it still has not gotten out of the testing phase in the two Arizona sites, which cover just 50 miles of border. Borkowski said satellite communications, even sending signals at two-second intervals, were too slow because by the time cameras could fix on trouble spots, the people or vehicles passing the border were often gone.
"As it turned out, a few seconds is enough to lose the image," he said.
Borkowski and Peters, the Boeing vice president, said the company was trying to work out what Peters called the "bugs or issues." The government has not given up, Borkowski said.
"I have to know what's going on. I have to have good, accurate, timely intelligence. I have to have the ability to act on that knowledge," Borkowski said.
And either way, "technology usually helps with the surveillance part. A couple of towers frees up to 50 agents to do more response work. They really complement each other."