The Skywall 100 brings the drone down intact, safely. (Image: OpenWorks)
The FAA is expected to announce regulations for the operation of small non-recreational UAV flights sometime later this year. For the most part, this should be welcome information for those individuals and companies that want to use drones for commercial operations.
Unfortunately, these new regulations, because they only affect commercial UAV flights, won’t do much to control irresponsible drone flights conducted by recreational pilots. In an interview with The Sydney Morning Herald, Chris Anderson, president of 3D Robotics, put it like this: "One of the ironies of this drone age is that because we've made drones so easy to fly and the process of learning to fly and all the safety and responsibility lessons that come with it are now no longer required….as a result, all that wisdom about safe and responsible flight doesn't come automatically.”
To deal with the threat of inexperienced recreational pilots, the FAA recently imposed a mandatory registration requirement for all small UAVs weighting between 0.55 and 55 pounds. The process exposes new operators to the FAA rules and presumably makes each pilot accountable for their drone flight, since the registration number on the UAV identifies the owner in the event of a crash.
Knowing the rules will certainly reduce the potential for abuse, but the simple fact is regulations don’t dictate compliance. If drones are going to be safely integrated into the U.S. airspace, it can only be done by the industry. The UAV industry is responding to this challenge in the following ways:
No-fly maps and apps
A number of independent developers have created no-fly maps. These applications render prohibited flying areas like airports, stadiums, utility facilities, courthouses etc. A list of these maps can be found at Directions Magazine.
Both DJI and 3D Robotics have added no-fly maps to their mission planning firmware. Owners of 3DR’s Solo see a pop-up if the drone is too close to an airport and, depending on airport proximity, DJI’s Phantom 3 pilots will either get a warning or will be forced to land their drone.
DJI also has another nice feature that let’s Phantom 3 beginners define a three dimensional geofense. Once set, pilots are able to safely fly within the area without worrying about losing control of their drone.
In addition to no-fly maps there are a number of apps, from both UAV manufacturers and third party developers, that provide enhanced navigation and photographic capabilities. Parrot’s FreeFlight 3, DJI Assist, and Drone Control US are just a few examples. The FAA has also gotten in the game with the B4Ufly iOS and Android app.
Autonomous flight is the ability to program or automate the drone’s flight path. Basically these are mission-planning modes that instruct the drone to follow the operator, navigate a specific set of waypoints or circle a point of interest. Programmed flight also provides an increased degree of safety because it takes control out of the hands of inexperienced operators while at the same time enabling them to get shots they may not have the skill to get by flying manually. DJI, 3DR, Parrot, Yuneec, Lily and Airdog are just a few of the drone manufacturers offering this tool. The screen shot below illustrates an automated flight path created from 3DR’s Tower app.
Object avoidance technology
Although autopilot is a valuable tool, it has a downside because it may instill a false sense of security in novice UAV operators. The technology works great in open spots like ski slopes or surfing locations, but not so well in areas with obstacles like trees or buildings. Just because the flight course is automated, it doesn’t mean your drone can’t crash into a course obstruction.
DJI, the undisputed market leader, has once again set the bar high with the introduction of its Phantom 4 quadcopter. Although Phantom 4 shoots 4 K video and 1080p at 120 frames/second, its most compelling feature is object avoidance technology. Two cameras on the front and two on the bottom allow it to detect and avoid obstacles while flying. DJI is one of the first companies to offer object avoidance features, but it will soon become the new normal for all prosumer drones, because it is a critical element in making drones safer. The market potential of this technology has not escaped some very big companies.
The DJI Phantom 4 is a master of object avoidance. (Image: DJI)
Intel Real Sense
Intel’s RealSense sense and avoid technology is a camera and software system that was originally designed to enable users to control their computers without physically touching their keyboard or mouse. Ascending Technologies, a German UAV manufacture, was the first to use RealSense in its Firefly drone. The partnership with Ascending Technologies has spawned new Intel investments in Airware, Precisionhawk, and Yuneec.
Qualcomm plans to offer competing object avoidance and geofencing technology in the form of its Snapdragon 820 chip. They demonstrated Snapdragon’s capabilities at the 2016 CES in January by flying the drone (lower right of the picture) through and around the arch (center of picture). In addition to making flights safer this technology reduces manufacturing costs by enabling multiple flight components to be placed on the same circuit board.
Qualcomm demonstrated Snapdragon’s capabilities at CES 2016 in January by flying a drone through and around an arch without mishap. (Image: courtesy of William McNeil)
Panoptesuav is a bit of a DIY project because it is designed to add object avoidance to the DJI Phantom 2 and the 3DR’s Iris by replacing the drone’s top cover with Panoptesuav’s eBumper4 product. According to CrunchBase, it looks like the start-up may not have gotten the funding they were seeking and it’s not certain if they are still taking orders. Either way, they are a pioneer and worth mentioning for attempting to make the skies safer.
Massachusetts Institute of Technology
In addition to commercial ventures, there are ongoing research projects. MIT
has developed technology that will operate when a drone is flying at 30 MPH. “Everyone is building drones these days, but nobody knows how to get them to stop running into things,” says Andrew Barry, PhD at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL). He, together with ProfessorRuss Tedrake, have built an obstacle-detection system that allows drones to avoid obstacles while flying autonomously.
Barry’s stereo-vision algorithm extracts depth information at a speed of 8.3 milliseconds per frame. This system enables flying drones to detect objects and build a map of the surrounding area at 120 frames/second. The software is open source.
Self-flying drone navigates between and around trees at 30 MPH. (Image: CSAIL MIT.)
Unusual and extreme solutions
OpenWorks Engineering is a UK start-up that protects restructed airspace from unauthorizated UAV flights by capturing the offending drone in a net fired from a shoulder-held gun they've developed. Skywall 100 uses compressed air to launch, track, and precisely intercept the UAV. OpenWorks expects to start shipping Skywall 100 later this year.
Birds of prey may be used to capture drones, but animal safety issues are still under review. (Image: euronews, YouTube.)
In the Netherlands, police have been training birds of prey to catch drones mid-flight. While it's fun to watch, one can't help but question the practicality of controlling drone flights with birds: how many eagles and trainers would be required to protect just the airspace around national airports, for example?
Innovations like these will ultimately bring the industry into alignment with regulatory bodies and increase flight safety regardless of pilot training, knowledge or experience. Irresponsible or willful misuse of drones will become increasingly difficult as manufacturers expand their abilities to assert control over drone flight.