[ad_1]
Like the rest of the world, Canadian drone maker Draganfly has been anxiously watching the spread of the new coronavirus. And when COVID-19 cases began to surface in Washington nursing homes in mid-February, the team began exchanging ideas. In March, Draganfly had licensed the artificial vision and artificial intelligence technology necessary to offer social distance and health monitoring services from the air. The demand to test the technology was “insatiable”, not only from the government and law enforcement, but also from healthcare, airlines, cruise ships, hospitality, theme parks and other commercial industries. In mid-April, the Westport, Connecticut police department had a pilot running, the first of its kind in the US. USA In addition, Draganfly had three to seven more American pilots planned. By April 23, the Westport pilot was dead.
But the story does not end there. We spoke to Draganfly CEO Cameron Chell before and after the abrupt termination of the Westport pilot. The drone company has two more pilots scheduled to start in less than two weeks. Chell says Draganfly has been “inundated” with requests from other jurisdictions, while the numbers on the private side “are even more prolific.” In fact, the next two American pilots will be in the private sector. One is drone based and the other is facility based. Additional pilots from the US public sector USA They will start “relatively soon”. As for Canada, Chell said “a couple of institutions” are also interested, particularly in the transportation industry.
As federal and local governments grapple with the coronavirus pandemic, from tracking the spread of COVID-19 to measuring when to lift the restrictions, everyone is closely watching autonomous technologies like drones and robots. The public and private sectors are desperate for a technology that can help limit human contact and provide early detection data on the implementation and effectiveness of measures such as social distancing. Any business that relies on human interaction, whether with customers or between employees, will be hungry for data to understand health trends. And drones could play a critical role in outbreak detection and tracking, safeguarding both public health and business operations.
Deploying drones since ’98
Unlike most drone companies, Draganfly has decades of experience. It was founded in 1998, and Chell is proud to lead “the oldest commercial drone manufacturer in the world”. The Canadian company has about 25 employees and is based in Saskatoon, Saskatchewan, with offices in Vancouver; The Angels; and Raleigh, North Carolina. Until this month, Draganfly was possibly best known for developing the first drone credited with saving a human life, in 2013.
Cheap consumer drones have become popular in recent years, thanks to market leaders DJI and Parrot. But none of Draganfly’s four sources of income is consumer related. The first is contract engineering (primarily for top tier US-based military contractors). The second is the manufacture of original systems, which means building drones with fixed wings that can perform vertical takeoff and horizontal landing, along with ground robots and other specialized products. That line of business encompasses the design and development of drone software for other companies and includes the health measurement system that has been in the news. The third line of business is managed services, when Draganfly essentially becomes the business data collection and drone arm of a business. The fourth line, which is still emerging, covers data analysis and management.
While Draganfly doesn’t build consumer drones per se, it has engineered various payloads, gimbal accessories, and software integrations for its customers that do. The company prefers highly specialized work for United States agents and border patrol. Forget cheap drones: think of batteries that work in colder climates, specialized sensors and a North American supply chain. In the past five years or so, Draganfly’s work has begun to lean toward the realm of public safety.
Draganfly customers “generally have a higher performance requirement,” Chell told VentureBeat. “They have some specialized needs that our engineering is in tune with. Also, they tend not to necessarily buy as much foreign product. Shoppers in that area are a little bit more aware, certainly on a military level, of security concerns and possible foreign parts and the like. But even when that leaks into public safety and law enforcement, they tend to have a bit more bias toward a North American or NATO-based solution. So naturally we end up migrating that way. “
Vital intelligence project
About 12 weeks ago, when “things were scary in Washington with nursing homes,” the Draganfly team was trying to figure out how drones could help. But they wanted to do more than just use drones to yell at people from the skies.
“We thought:” Put a speaker on a drone? “Big deal. Really? That’s not innovative,” Chell said. “We were thinking that if this hits, we need to be able to provide more value than very typical use cases like that.”
The company realized that it did not have the artificial intelligence skills to accomplish what it really wanted to do. Then he started talking to his partners.
“We went looking for him,” said Chell. “We were thinking” Oh, thermal cameras! “Probably everyone in the world has thought about thermal cameras. And we discredit the use of those very quickly. Thermal temperature does not measure core temperature, which is what is required to understand if a possible fever is present. We were like , ‘U.S [have] I have to find something very different. “And behold, and fortunately [for] us, it was the University of South Australia (UniSA), which is more of a coincidence than anything. But given the fact that they bought their first drone from us in 1999, the relationship of trust [was there, and] we could move very fast. “
The Canadian company investigated, built use cases, conducted tests and within a few weeks had signed an agreement. Draganfly paid $ 1.5 million to license a respiratory and health monitoring platform, the Vital Intelligence Project, developed in collaboration between UniSA and the Australian Department of Defense Science and Technology Group. Draganfly would commercialize and deploy the computer vision technology.
The Vital Intelligence Project can help estimate the distances between people, but it can also monitor the temperature, heart rate, and respiratory rate of people in crowds and in the workforce. Draganfly envisioned the technology that airlines and cruises would deploy; for potential groups at risk, such as older people in care centers; and in convention centers; at border crossings; and within critical infrastructure facilities.
“We license it for camera networks and drones,” Chell explained. UniSA developed the core technology: “specific machine vision and AI in a non-productive way.” Draganfly simply had the public research university in its Rolodex. The company then developed the production form, which includes camera networks and drones.
“That includes everything from going out and doing the policy development work to what the GUI should look like,” Chell said. “Both software and mechanical engineering provide stabilization for drones so they are optimized to collect this data. That piece of IP, go to the market, and the actual marketing piece is all internal [at] Draganfly. But the research and intellectual property behind machine vision and AI up to this point have been [at] University of South Australia. Lots of test and learning data we’re bringing in: We’re developing that part of the IP code now, with it. However, it is the staunch doctors who are doing the AI work. “
Still, a reuse was required, as the Vital Intelligence Project was not exactly being used to monitor groups of people.
“They were using it to be able to fly helicopters over … disaster relief areas and detect survivors’ vital signs on the ground,” Chell said. “They were able to determine what resources they needed to apply where or the severity of the survivors’ current situation, and whether they needed to be right at the time.” They also ended up using it to monitor wildlife. You have a migration and there may be fires or droughts. Wildlife officials should see: “What is the health of the herd? Do we need to take any action?” They also used it for prenatal babies, where they didn’t want many people to go in and out of the room. due to possible introductions of infections, and also in that situation where the probes and monitors that are taping babies are usually not remaining or uncomfortable in some way. Those are the registered and peer-reviewed use cases that are available. “
Draganfly and the university took the technology and adapted it for social distancing and health monitoring. To be clear, the Vital Intelligence Project had never been tethered to a drone and drew a crowd before the Westport pilot.
“The previous use case that would be more similar to this one was designed to be used in disaster relief areas to obtain vital signs from survivors on the ground,” Chell said. “Those are the same vital signs that we can now collect anonymously from a crowd to determine if there are infectious or respiratory problems.”
Westport: flattening the cornering program
Draganfly began test flights in Westport, Connecticut to identify social estrangement and detect symptoms. The city is in Fairfield County, adjacent to New York City and considered the epicenter of the coronavirus spread in Connecticut.
The three-phase pilot was supposed to validate the use of technology and that officials develop a public safety policy around it. The next step would have been to test those policies. The entire process was supposed to take about 60 days, during which Draganfly hoped to start additional pilots.
Phase one: social distancing
The first phase was to test whether the technology could be an effective multiplier of resources, allowing officials to cover more ground to see if social distancing is being effectively accomplished, for example. Instead of sending a few cruises and having officers walk, they could put a camera in the sky and assess where to apply the resources.
“The social distancing, which we have shown in the videos, is real visual data,” Chell said. “It gives the camera operator, typically the officer, real-time data. They must make operational decisions at that point if they need to separate a crowd. Or all is well [and] they don’t need to go in and waste time there. “
Westport Prime Selectman Jim Marpe called it the “Flatten the Curve Pilot Program.” It was supposed to help the community “practice safe social distancing, while identifying potential coronaviruses and other life-threatening symptoms.” Police Chief Foti Koskinas said at the time: “Using drones remains an access technology to reach remote areas with little or no manpower required. Because of this technology, our officers will have the quality information and data they need to make the best decision in any situation. “The hope was to deploy to the city’s beaches, train stations, parks and recreational areas and shopping malls, and “It will not be used in individual private courtyards, nor will it employ facial recognition technology,” the police department said.
There was no health monitoring in phase one. Before the Westport rider finished, Chell was already calling phase one a “hit,” so we asked what exactly that meant. “The technology worked in a real-world environment,” said Chell. “So that was successful. We were able to obtain very good operational data on social distancing. And the working relationship between public safety officials and us was also a success. “
Until it wasn’t.
In announcing the end of the pilot just a couple of days later, Marpe said: “In our good faith effort to get ahead of the virus and the potential need to safely manage and monitor crowds and social estrangement in this environment, our announcement perhaps it was misinterpreted, it was not well received and it raised many additional questions. We listen to and respect their concerns and therefore we go back and reconsider the full impact of the technology and its use in the enforcement protocol. “Koskinas added: “We thank Draganfly for offering the pilot program to Westport and we sincerely hope to be included in future innovations once we are convinced that the program is appropriate for Westport.”
When we spoke to Chell a few days later, he seemed to understand why the pilot had to end.
“The official setback was about health monitoring and the misunderstanding about how it works, what it does and what it is for,” he said. “And at this point, Westport feels politically that they just don’t want to go ahead with the project.” They had been a great help. They provided us with a great vision, a great policy framework and everything in between, but they are not going to move forward to move forward with us, at least not right now, which is totally fine. There are many people with whom we can move forward, and [the people in Westport have] been totally professional and great to work with. “
Draganfly has consistently said that no part of the Vital Intelligence Project employs facial recognition technology. Still, we wonder if the resulting output from phase one could be reused to do it. Could someone take the video and run a facial recognition algorithm on it?
“No more than you could do today on a security camera system,” Chell explained. “The requirements to run social distancing, in terms of resolution and stabilization on the video platform, are minimal compared to what you must have to run the health measurement platform. So while today we can take advantage of existing safety nets and do social distancing, I couldn’t take the same video feed and do heart rate, respiratory rate, or even often facial recognition stuff that we don’t use. “
Phase two: anonymous health measurement
The pilot never reached phase two, which was the anonymous health measure. The plan was to test the crowds as sample sets: how many are coughing, sneezing, feverish (considering heart rate, respiratory rate, high blood pressure, and biometric measurements based on skin tones). We asked if health monitoring technology had been tested on a variety of skin tones, given AI issues in the past.
“It has. There are some challenges with that sometimes,” Chell admitted. “Darker skin tones and different types of lights and the rest can create some problems. If you have someone walking to a kiosk and using this type of technology, it is a different scenario because it is a controlled environment, you control the lighting and you can go from there. Unlike, if you’re trying to make it through a large room that has 300 or 2,000 people, you won’t get all the people, always. But you will get a very significant population sample, especially as you get more and more over time. You will get 85% of people, 15% of people, because of lighting, because [have] a hoodie or [a] a certain type of skin tone, it just won’t catch. But then again, it’s not meant to catch an individual. “
Chell insists that it is not a problem, because the technology is not intended to identify people. The objective is to measure the health of a population.
“When you combine things like fever, cough, elevated heart rate, particular respiratory rates, then you have a picture of health,” Chell said. “You are not diagnosing whether someone has COVID-19 or not, but you are taking a health measurement and getting a pretty good idea of the rate of potentially infectious or respiratory diseases in an area.” If it is below 0.02%, we are in good shape. If yesterday it was 0.02%, and tomorrow in a sample of similar size in the same area, it is 1%, and the next day it is 3%, it will be above. Now you have information that can be correlated with [the question of whether] social distancing [is] is it going to be required OR you certainly can have developed policies. We are not caught in a situation where there is something spreading in a pandemic and we don’t know yet. “
A big difference between phase one and phase two that led to great confusion is that the latter does not broadcast video. Draganfly posted videos to show how the technology worked, but that was misleading. The health measurement system does not register subjects in a location that the drone “saw”.
“He just goes back and says that in this particular geographic location, where he made the health measurement data, there were 22 people in the field view. Here are the heart rates, here are the respiratory rates, here are the fevers. Here is the probability and percentage of infections and / or respiratory diseases. “
The system takes “15 stable seconds” to acquire the data.
“At least that data time is needed to understand the respiratory rate,” Chell explained. “In that time period, you also have to beat your heart rate so you can collect that data simultaneously. You can also get biometric measurements of your skin tone data pretty quickly in that time period. Within reason, if you have good exposure to skin tone, you get core temperature, along with these other things for anyone in the field of view. So if there are 20 people in that field of view, and you have a good angle over those 20 subjects, in those 15 seconds you can collect 20 data sets. “
The drone sends the data it collects to the cloud (Draganfly uses AWS and Fortinet) for processing. “All of this happens in the cloud through encrypted lines. So that’s a secure cloud environment where all that AI happens. If the drone goes down, there is no SD card I can take out with a ton of cool data. “
Interestingly, Westport was not using Draganfly drones: the plan was to deploy its Commander series in phase three. For the sake of time, the first two phases were relying on third-party drones, which Westport already had. The police department launched its drone program in early 2016 to support the operations of its diving team in locating submerged objects or victims. Later it expanded to accident investigation, scene documentation, search and rescue, public works projects, and pre-event planning.
Draganfly claims its technology works up to 190 feet from subjects. “At 190 feet, you’re talking about a $ 35,000 drone, just for the cameras and the additional sensors and stabilization. On a $ 600 drone, that doesn’t have really optimized stabilization software, and as such it’s more like 20 feet. Using a drone that has optimized stabilization and a zoom lens, any distance is theoretically possible. We are currently working with different ranges, and 190 feet have worked well. ”
But again, that was for phase three. The Draportfly system piloted by Westport could have worked almost 10 times as far. For now, it appears that Westport will never verify those claims. Other towns and companies, however, want to try it.
