This article focuses on the use of aquatic robots for cleaning water bodies. Brooklyn Atlantis, a 40-pound unmanned aquatic vehicle, has been developed and used to collect gather water quality data and other info, such as pictures of oil-smeared animals or dying fish, vital for assessing the impact of environmental disasters. Brooklyn Atlantis has already made important discoveries from its data collection. With its underwater photographs of water turbidity, Brooklyn Atlantis also pinpointed sediment deposits that naturally accumulate with the canal’s water flow. The findings can inform engineers where the marsh grasses are more likely to grow and where the fish are more likely to spawn. As a result, these areas are being earmarked for a wetland reconstructions prototype. With a grant to commercialize the technology, Laut and Porfiri are actively looking to convert their project into a company that assesses environmental situations, helping to clean up and rebuild.
Launching Brooklyn Atlantis, a 40-pound unmanned aquatic vehicle, is a two-person job. Jeffrey W. Laut, a researcher at the New York University Tandon School of Engineering, snapped a GoPro camera on the robot and picked it up, balancing on a small plastic dock above the slippery rocks. Meanwhile, Laut's co-worker Tommaso Ruberto carefully descended from the dock, sinking ankle-deep into the brown muck in his oversized rubber boots. Ruberto grabbed the robot from Laut, and slowly pulled it down, avoiding splashing himself.
No one wanted to get wet as Laut and Ruberto were launching Brooklyn Atlantis into Brooklyn's murky, notoriously filthy Gowanus Canal, one of the most polluted bodies of water in North America.
“I haven’t fallen into the canal yet, but one of these days it's going to happen,” Laut joked as the robot finally was afloat next to a dead fish and a small elongated brown object of questionable origin. In the distance loomed floating plastic bags—a potential problem.
“A couple of times a plastic bag got wrapped around the robot's thruster and I had to go get a canoe and rescue it,” Laut said as he manipulated the controls, sending Brooklyn Atlantis underneath a nearby highway overpass.
Laut worked with Maurizio Porfiri, professor of aerospace and mechanical engineering at NYU, to help design Brooklyn Atlantis in the aftermath of the 2011 Deepwater Horizon oil spill. They thought it would be easier, safer, and cheaper to send robots to map the extent to which the oil and other pollutants had spread before an accident, Porfiri explained. The robot would gather water quality data and other info, such as pictures of oil-smeared animals or dying fish, vital for assessing the impact of environmental disasters.
Once the team built the first prototype, it needed a fairly polluted waterway to test it. The Gulf of Mexico was far. Gowanus Canal, which cuts through a mixed residential and industrial neighborhood in Brooklyn, was only a short drive from campus. And while Deepwater Horizon made headlines and magazine covers around the world, the Gowanus is a no less impactful, long-standing ecological concern. The canal's bottom is laden with carcinogenic byproducts of the gas and coal industries that lined its shores in the 19th century. In 2010, the Environmental Protection Agency named the waterway and some surrounding lots a Superfund cleanup site.
Today, its water is still contaminated by the runoffs from the overflowing sewage systems.
“When it rains, we get about a million gallons a day of raw sewerage spilling into the canal,” said Eymund Diegel, an environmental and urban planner with the New York City Department of Transportation, who keeps watch on the canal and its surroundings in his free time, mapping pollution levels and underground streams.
The ongoing legacy of pollution made the canal as a good place as any to send the robot for a test. “We realized that it was as bad as the Gulf of Mexico,” Porfiri said. “So we decided to stick with the Gowanus Canal.”
Over the next few years, Brooklyn Atlantis ventured across Gowanus numerous times. It took thousands of pictures, which the team turned into a citizen scientist project by posting them online and inviting people to help identify objects in them. It mapped the canal, recording area-specific water quality readings such as pH, salinity, temperature, and levels of dissolved oxygen, all vital for marine life.
If all the robot did was operate as intended, that would have been a victory for the design team. But Brooklyn Atlantis discovered something in its measurements—unexpected fluctuations in the water readings that helped Diegel localize hidden spots of pollution—that will enable engineers to better accomplish a restoration of the waterway.
To solve the pressing pollution problems, one needs to better understand the complicated engineering history of the canal and its surroundings. Brooklyn Atlantis managed to shed enough light on Gowanus's darkness to accomplish just that.
The Gowanus Canal is actually an engineering marvel in and of itself. But before there was a canal, there was Gowanus Creek, a tidal estuary with multiple freshwater creeks feeding into it. In the mid-19th century, the surrounding marshes had been filled in and the creek was dredged and built into a canal extending the city's water-based transportation infrastructure—New York Harbor's docks served everything from transatlantic ships to Erie Canal barges. Industry soon followed and the canal was lined with gas plants, coal yards, and paint factories, all dumping their waste into the water. Part of Brooklyn's sewer system terminated there, too.
The canal was essentially a gigantic toilet, but one where the water stagnated.
The Gowanus stunk so badly that in 1911 the city built a flushing tunnel system, which was meant to purge the canal. The system used a 7-foot-diameter propeller to pump water from the canal's “head” through a 12-foot-diameter tunnel running 1.15 miles underneath Brooklyn's neighborhoods, eventually dumping the canal water into the Buttermilk Channel separating Brooklyn from Governors Island in the harbor.
At the time, the tunnel itself was a remarkable engineering feat, according to Kevin Clarke, an environmental engineer and portfolio manager at the New York City Department of Environmental Protection. The under-street tunnels of the city's earliest subways had been dug using the cut-and-cover method, in which an excavated trench is eventually roofed with an overhead support system. Engineers designing the flushing tunnel opted instead to construct it using compressed air to help support the structure from collapsing as its new sections were dug and fortified.
“The guys would have to go in through an air lock and they worked under pressurized conditions,” Clarke said. “It was pretty phenomenal that they were able to construct a tunnel like this over 100 years ago.”
The engineers of the time also considered whether the system should pump the water in or out of the canal. They decided that syphoning the water out at the canal's head would benefit the boat traffic because the resulting current would pull the barges up further inland. “At the time the thinking was that if boats were coming in laden with cargo, it would be easier for them to go with the current induced by the flushing tunnel, and it would be easier for them to fight the current on the way out when they were empty,” Clarke said.
The system failed in the 1960s and sat broken for three decades. Eventually, the neighborhoods near the canal began to gentrify and pressure from the new residents compelled the city to repair the flushing system. In 1999, engineers installed a new impellor system that ran in reverse of the original, drawing in water from the relatively cleaner Buttermilk Channel and pushing it into the canal. The goal was to improve the canal's dissolved oxygen levels, which at the time couldn’t be lower—they were at 0 mg per liter. The East River boasted very healthy oxygen levels—8 mg per liter. “By reversing the flow through the tunnel, you are basically taking the highly oxygenated water from the East River and you’re bringing it right into the head of the canal,” Clarke said.
The new impellor helped, but in 2010, the Department of Environmental Protection opted for a more ambitious upgrade. In 2014, the DEP completed the $160 million project that included three vertical turbine pumps of 470 horse power each. On an average day, all three pumps push about 215 million gallons of water—more water as the tide comes in, less when it goes out. Having three pumps provides resiliency, since if one breaks, the other two would keep working while repairs go on.
The new setup has significantly improved the water quality—there's less stink and more life in the canal, said Andrea Parker, executive director at the Gowanus Canal Conservancy, a community-based environmental nonprofit. There are more cleanup efforts looming, including a plan to seal off the canal's carcinogenic bottom and rebuild the previously destroyed wetlands around it.
With the cleanup efforts under way, the canal's once trash-strewn shores have suddenly become desirable. Instead of industrial wasteland, developers can now describe nearby lots as waterfront property. Old workshops have been replaced by condominiums.
Ironically, that made the robot's testing harder. Easily accessible banks were fenced off, and high-rises claimed parking places, leaving no space for Laut's van. Despite the obstacles, the NYU engineering team persevered, upgrading and improving the robot over several iterations. The robot's current model is half of its original weight, and is faster and more maneuverable. But even before the latest improvements, which include advancement in autonomy and obstacle detections, the robot already proved useful—by digging up some sewage secrets.
As Brooklyn Atlantis buzzed along, Laut put away the control box and tapped on his computer to switch the robot to GPS. “This is a newer model,” Laut said, as the robot sped away, water bursting from under its two Blue Robotics T200 thrusters. “I’m still trying to work out the autonomy.”
Powered by two 14.8 V lithium polymer batteries, rated for 16000 mAh each, Brooklyn Atlantis can travel at almost 8 km per hour with more than enough battery charge to cover twice the length of the canal. The current model can carry two LIDAR optical distant measurement sensors and an additional camera for obstacle detection, such as the walls of the canal or other objects. Equipped with a Gumstix embedded computer, two motor drivers, GPS, and an ArduPilot software that runs on a Pixhawk module, the robot uses a frontal camera to snap images and a multi-parameter sonde to measure pH, salinity, temperature, and dissolved oxygen.
Although today Laut left out water quality sensors to focus on testing the robot's autonomy, Brooklyn Atlantis has already made important discoveries from its data collection.
As the robot was gathering water readings during its first few years, Diegel and the NYU team expected to see lower salinity and oxygen levels at the head of the canal, even with the new flushing tunnel inflows. According to maps, that's where the largest underground freshwater stream was in the area, and typically underground streams are low in oxygen because they flow underground and low on salinity because they’re freshwater streams.
When the readings didn’t show that, it tipped off Diegel that the stream must have been diverted into a sewer decades ago.
With the stream flowing into the sewer, that meant that the sewer system had less capacity to handle runoff from the streets on rainy days, which leads to overflows and raw sewage flowing into the canal. Untangling the streams from sewage would improve the canal's water quality, but finding where the streams enter the sewage system is not always an easy task.
“[The robot's findings] helped us understand where the stream got diverted into a sewer,” Diegel said, “and what can be done to turn it back into a freshwater stream—to untangle the clean water from the dirty water.”
With its underwater photographs of water turbidity, Brooklyn Atlantis also pinpointed sediment deposits that naturally accumulate with the canal's water flow. That fact is important for a potential wetland restoration project, which would benefit from using the already existing soil deposits. The findings can inform engineers where the marsh grasses are more likely to grow and where the fish are more likely to spawn. As a result, these areas are being earmarked for a wetland reconstructions prototype.
With a grant to commercialize the technology, Laut and Porfiri are actively looking to convert their project into a company that assesses environmental situations, helping to clean up and rebuild. After all, Brooklyn Atlantis had already made a valuable contribution to the environmental restoration business, Diegel said. “It helped us find what our grandparents overlooked.”