Deep in the Otay Mountain Wilderness, there is no wall.
The only boundary between the U.S. and Mexico is a section of barbed-wire fence in a pastoral valley. And miles and miles of treacherous terrain.
It’s a territory crisscrossed with steep trails that disappear into tunnels of thick brush, a place looped by violently rutted roads that Border Patrol agents drive daily.
Land like this is not a likely candidate for President Donald Trump’s “big, beautiful wall.”
But it is fertile for an invisible kind of fence, one built of artificial intelligence, radar, drones, sensors, motion-activated cameras and even lidar, the technology used in self-driving vehicles.
“Virtual walls” or “smart walls” along the southwest border are increasingly being billed as an alternative to the proposed concrete and steel barriers that have so sharply divided public opinion.
An electronic fence is not about preventing intrusion as much as it is about detecting intrusions and then intercepting them.
While even the strongest proponents for such a technological solution admit that physical barriers are probably best in urban areas like San Diego and El Paso, Texas, they see a virtual wall as a cheaper and more effective way to police much of the rest of the 2,000-mile southwest border.
U.S. Customs and Border Protection has asked for $223 million of its fiscal 2019 budget to focus on technology improvements, funding that could have bipartisan support.
But the increased focus on smart walls is deepening concern about a growing Big-Brother-is-watching network, and civil liberties organizations have asked lawmakers to proceed with caution.
“Warrantless use of these technologies comes at an unacceptably high cost,” Neema Singh Guliani and Michelle Fraling, officials with the American Civil Liberties Union, said last month. “They allow the government to track, surveil and monitor individuals indiscriminately and with precise detail. Individuals in the border zone should not be subject to near-constant surveillance that intrudes on the most intimate aspects of their lives.”
An electronic fence is not a new idea. It is one that the federal government has failed to implement on a wide scale at least three times in the past few decades.
From 1997 to 2005, the government spent about $429 million on two border technology programs deemed unsuccessful, according to congressional reports.
The third, most recent iteration was known as SBInet, a project started in 2006. Boeing was contracted to build a network of surveillance towers and ground sensors that would detect intrusions and relay the information back to command posts where agents would decide how to react.
It was introduced first along the Arizona border, but expansion was stopped in 2010 because of serious concerns with its feasibility.
The Government Accountability Office said SBInet was poorly managed and beset by cost overruns and missed deadlines. The technology was also troubled, pulled off the shelves rather than custom-designed for the border environment and job at hand.
The program was killed in 2011, at a cost of more than $1 billion.
“I terminated it when I was commissioner. It was a failure,” Alan Bersin said in an interview. Bersin, a former U.S. attorney in San Diego, worked as a “border czar” under two presidents and then as CBP commissioner from 2010 to 2011.
Part of the problem, Bersin said, was pursuing a one-size-fits-all approach to the border.
“Instead of one technology system borderwide, we now follow a system customized to the particular conditions on the border,” Bersin said. “The setting up of an electronic fence in the Sonoran Arizona desert is far different than doing so around San Diego’s urban area.”
Even before SBInet was terminated, doubts about the program led authorities to redirect millions of dollars in funding to deploy other technology across the border. That included mobile and fixed surveillance towers, cameras, drones and thermal imaging devices.
Many of those systems are in use today, although their concept is dated as technology leaps forward.
Border Patrol sectors have largely focused on incremental improvements in a piecemeal strategy.
Evidence of that effort can be seen in the Border Patrol’s Brown Field Station east of San Diego, a 568-square-mile region stretching from Otay Mountain to Tecate, with 11 miles of international border.
Agents there typically spend much of their shifts investigating alerts from seismic sensors tripped by potential intruders. The problem is that the sensors are also tripped by rabbits, deer, rain and other agents.
The recent addition of several motion-activated Buckeye cameras has helped tremendously, agents said. The camera takes a snapshot immediately after being triggered, sending the information back to the Tactical Operations Center at the station for quick analysis.
The command post, known as TOC, monitors other surveillance, too. Radar returns predict the presence of suspicious activity and send images that can distinguish vehicles and people. An agent sitting at the computer can then direct a field agent to intercept.
The TOC also takes a high-tech approach to one of the Border Patrol’s most low-tech tactics — tracking “sign.”
Sign refers to the traces left behind by a person on the move, like footprints, litter, broken branches and overturned rocks. Reading sign is how agents have done the job for decades, something unchanged by even the best technology, they say.
What the TOC has essentially done is crowdsource sign intelligence. For example, an agent who finds fresh footprints of a group of people headed up a specific trail will report it to the TOC. The spot will be marked with GPS coordinates and updated as more sign is discovered. The information can be passed onto the next shift for continued tracking.
“The TOC can track sign for two, three days,” said Border Patrol spokesman Fabian Carbajal, who has spent 13 years of his career patrolling Brown Field Station.
The capabilities are a start, but innovators at many smaller companies have their sights on technology that seemed futuristic not too long ago.
Quanergy, a Silicon Valley company that developed lidar for driverless vehicles, has been testing the technology in the Texas border town of Del Rio with the cooperation of a local sheriff, The New York Times reported. The laser sensors give a 3D view of an area, building on the technology already being used.
Anduril Industries, based in Orange County, Calif., is working to apply the power of artificial intelligence to national security settings. One of its founders is Palmer Luckey, the entrepreneur behind the virtual reality company Oculus.
“One of the big advantages to the use of AI-enabled technology is that you can deploy it without having to increase manpower,” said Matt Steckman, Anduril’s head of corporate and government affairs. “You get computers to do a lot of the work and let the humans be the decision-makers.”
Anduril’s system, called Lattice, uses hundreds or even thousands of sensors, and then lets artificial intelligence scan the environment five times a second and interpret the results. Only meaningful results will be transmitted to an agent in the field, on a smartphone or tablet, in the form of a cropped image of the form. The agent can then decide what to do about them.
The system takes out the middleman — the agent who has to sit at a bank of computer monitors in a darkened room and interpret the sensor results in real time.
The network uses different types of sensors to create a more complete image of the environment, and each sensor is able to interpret its own data — brand new technology known as edge computation that is still in its infancy.
The system completed a pilot phase last summer in part of Texas and became operational. It was also tested east of San Diego, leading to 10 interceptions during the first 12 days, Wired magazine reported last June.
Tech companies are also working to make devices more mobile, and therefore more nimble.
Benchmark, based in Tempe, Ariz., has built on the idea of surveillance towers and applied it to Border Patrol pickup trucks. At the top of a 30-foot retractable mast are powerful day and night vision cameras and lasers, with a range of about 6 miles.
The first order of 14 Ford F-150s were deployed to Texas about a year ago, but five were shifted to the San Diego Sector as a large migrant caravan neared Tijuana late last year, said Jan Janik, Benchmark’s chief technology officer.
The company is working on delivering 15 more trucks, each with technology worth $300,000 to $400,000.
Benchmark hopes to use the trucks as a launch pad for other capabilities, including giving agents the ability to have control on tablets if they need to leave the trucks.
Tying a drone into the system has also been suggested by Border Patrol.
“Today I can identify an intruder at 6 miles, but I can’t tell you if he’s wearing a sidearm or if that’s a lunch box,” Janik said.
But what if an agent could push a button and the unmanned aerial vehicle could fly to the GPS coordinates, get a closer look, track the person and return to the truck?
“You can’t overload the agent with data or task them with things that keep them from their primary mission,” Janik said. “You don’t want to train every agent on a drone and have them focused on a joystick. You want technology to have that capability.”
Former acting CBP Commissioner David Aguilar said technology will unquestionably be a large part of securing the border — used in combination with physical barriers where appropriate and manpower.
He gave the example of SAR radar, being looked at to constantly map and remap a portion of land in remote areas to detect terrain change, such as a new trail or road implicating fresh trafficking routes.
“That’s good intelligence,” Aguilar said in an interview. “You can go in there, drop ground sensors in the area, fly a drone on a persistent basis or post tethered aerostats to watch that area.”
But he warned against viewing any one solution, including technology, as a silver bullet.
“There is no such thing. You’re never going to have a border that is going to be impenetrable. That’s just not going to happen. Does that mean we stop working toward it? No.”
Keeping a virtual wall alive and running has its challenges.
Devices require power sources, which can be difficult in remote places. They need robust communications infrastructure to transmit information. They have to stand up to the rigors of a rugged work environment. And they can be destroyed or vandalized.
Agents who work the Otay Mountain Wilderness say the emerging technology helps. But nothing can replace boots on the ground.
“All the technology is not going to do what we do,” said veteran Agent Jose Damian. “When a scope spots something, you still need agents to go retrieve it.”
With all the tech talk, it can be jarring just how rudimentary some of Border Patrol’s tools are. On a recent Wednesday, Damian stood atop a crude helicopter landing pad deep in the mountains, with cheap, old binoculars to his eyes, scanning the valley below that separates the U.S. from Mexico.
Agents here don’t have vehicle laptops. They communicate with radios, powered by batteries that have died at times when agents needed them most, and will likely die again.
Farther east, near the Tecate Port of Entry, mobile floodlights stand sentry along the old landing mat fence that separates dirt parking lots in the U.S. from colorful, dense housing a stone’s throw away in Mexico. The lights require an agent each shift to fill them up by hand with gas.
© 2019 The San Diego Union-Tribune
Distributed by Tribune Content Agency, LLC.