These drones can avoid midair collisions by listening for other aircraft
When a drone from a company called Zipline is zipping through the air at some 70 mph, the ideal scenario is that no other low-flying, but faster, aircraft smashes into it. Zipline drones deliver health-focused supplies (like blood) via parachute in countries such as Rwanda and Ghana. They also deliver commercially from a Walmart located in Arkansas . In Rwanda, they’ve even been delivering semen from bulls (and pigs), for the purposes of insemination with a focus on, in the case of the bull semen, genetic diversity and milk production.
Zipline, like all other aircraft operators, doesn’t want any midair collisions. Keenan Wyrobek is the company’s chief technology officer. He says it can be difficult to navigate the airspace in the US. Low-flying planes, such as those cruising in a Cessna for pleasure, or someone operating crop dusters or helicopters, may not have a transponder to announce their location. Wyrobek states that transponders are not mandatory for many of these planes. “There’s kind of a wild west spirit of aviation in this country.”
While a pilot of a small plane must look ahead to avoid hitting another aircraft, such as a hovering helicopter or a small plane, a drone needs to take the necessary actions to avoid being struck. Wyrobek states that it is the drone’s job to spot a Cessna approaching them and move out of their way. In this case, the cowboy must make way for the cowhand who is galloping faster.
This raises questions that aren’t limited to this company. How can a drone without a human on board use their eyes to see for traffic that might crash into it, and get out of the way a fast-flying aircraft? How does it identify the threatening traffic?
The solution to this larger problem, in Zipline’s case at least, does not involve radar, cameras, lidar, or other sensors, which tend to be approaches commonly used in the autonomous car space. Instead, the drone company decided to use microphones that can detect other aircraft and then allow the drone to get out of their way.
The setup goes like this: A total of eight microphones, each one placed on a probe protruding from the leading edge of the 11-foot wing, comprise the sensor array tasked with detecting other aircraft. The system must be able ignore ambient noise from the drone (the air around it, its propeller sounds) and just listen for other flying vehicles. Wyrobek states that the array is essential to help get enough signal-to noise to hear planes far away and also to find out where they are. It can then “triangulate” where the planes are actually coming from
To perform this trick, the drone uses a small amount onboard computing power. He explains that the drone uses a combination signal processing techniques, such as beam forming, as well machine learning and AI-based techniques to locate the aircraft. A small onboard GPU helps with this job, as those types of chips are good at handling AI-related tasks. Wyrobek states that the microphones don’t produce much data. He adds that the actual information pipe is so small [so] that it doesn’t have a large compute load.
To build the system, they collected training data that included some 15,000 planned interactions between a drone and a human-crewed aircraft like an airplane or helicopter. Although microphones won’t be very helpful for a hot air balloon or glider, Wyrobek states that this approach has a fortuitous advantage: faster-moving aircraft tend to be louder which means that the signal from a faster-moving threat will be stronger.
At the moment, the company is awaiting regulatory approval to allow the drone’s software to make decisions to avoid any aircraft that might strike it. This would include a maneuver like proactively turning off the runway and entering a holding position until the coast clears. The microphones are currently installed in some drones, even though the whole system hasn’t been turned on yet. He says that the microphones think they are in control, but they are not. The team reviews the data post-flight to ensure the drone did what it was supposed to.
Wyrobek expects to use the sound-based detection software in many of their operations. He says that he thinks it will be used in most areas. “As we scale up, we want to continue increasing our safety, so this is a way for us to do that.” This sounds good from an airspace management perspective.
I’m a journalist who specializes in investigative reporting and writing. I have written for the New York Times and other publications.