Florida-based Prioria Robotics formally launched the Maveric unmanned aerial vehicle on 13 May and with it introduced a unique philosophy for answering the package of evolving regulatory requirements known as "sense and avoid".
The five-year-old start-up did not begin as an aircraft designer, but as a company focused on designing embedded systems for autonomous decision-making and sensing applications for robotics systems in general.
From this beginning came a different approach to the collision avoidance problem for man-portable sized UAVs.
"What we saw was that the [UAV] industry was doing a lot of work trying to send all of the information to be processed on the ground," says Bryan da Frota, chief executive and founder of Prioria. "Given the existing datalinks, this obviously wasn't the best choice."
The company is guided by an incremental philosophy on addressing the requirements for the US Federal Aviation Administration's still-evolving standard called sense and avoid. The challenge is to make a UAV smart enough to autonomously detect and steer clear of other objects in their flightpath.
"If you try to take the whole, entire see and avoid problem, it is a massive problem," da Frota says. "We said: 'Let's pick one segment of the problem that we can do very well'."
Prioria's focus has been on demonstrating a basic, collision avoidance capability for a micro-UAV. The goal is enable the aircraft to avoid trees, buildings and helicopters.
STREAMING VISUAL DATA
For small UAVs equipped with a forward-looking camera, Prioria's approach is to integrate an on-board processor able to interpret the streaming visual data for collision avoidance cues.
That requires a combination of processor technology small enough and powerful enough, as well as algorithms sophisticated enough, to interpret the data and command the aircraft to take appropriate action.
Using a proprietary set of algorithms - and "a lot of trial and error", he adds - the UAV can anonymously sense and avoid certain fixed objects in its flight path.
Prioria first developed a 100g (0.04oz) on-board avionics suite dubbed Merlin. The system interprets the UAV's sensor data using hardware that accelerates processing rates. The key to refining the system's ability is a combination of improving the algorithms and streamlining the hardware required for processing.
The capability will improve as the embedded processor matures. Prioria's technology, for example, is not ready yet to allow a UAV to fly autonomously through an urban canyon, but the goal is to reach that level through further refining the basic techniques.
"We are not trying to do everything," da Frota says. "We are taking one problem and making sure we have solved it very well."
The design of the Merlin processor eventually prompted Prioria to develop its own aircraft, but the latter started almost as an afterthought. "We built our plane originally as just a demonstration system for our hardware and software because no one believed us," da Frota says. Despite its humble beginnings, Prioria developed its demonstration aircraft into a marketable airframe for both military and civil surveillance roles.
The new product is actually two different variants of a man-portable aircraft known as the Maveric 100 and Maveric 150. Both 1.13kg (2.5lb) airframes feature bendable, 73.6cm (29in) wings for tube stowage. The Maveric 100 is the "bare bones" aircraft with only a side-looking camera. The Maveric 150 includes a nose-mounted camera for guidance, a gimballed camera in the belly for surveillance and the embedded Merlin processor.
The bendable wing is a unique feature that allows the aerofoil to simply wrap around the fuselage for tube-stowage. As an unintended benefit, the flexible wing also improves the Maveric 150's flight characteristics in certain conditions, allowing the aircraft to remain controllable in wind gusts up to 30kt (55.5km/h). Prioria has delivered the Maveric 150 to its first customer - the US Army - which is interested in understanding the capabilities of the aircraft and the anti-collision system.
Meanwhile, a civil agency customer in the UK has also purchased the system, while two more orders by Asian customers are in process, da Frota says.
The company's next step is to find an industry and government partner to support a new phase of demonstration flights scheduled to begin next year. Using a new set of algorithms licensed from University of Florida researchers, Prioria will attempt to integrate a target-tracking capability into the Merlin processor, potentially offering an autonomous capability to navigate around moving objects.
Source: Flight International
