Drone collision avoidance for beginners

The lightweight nature of drone airframes and flight components generally makes them intolerant to impacts, which typically require costly repairs – a risk that is reflected by the cost of insurance. One rugged solution to the risk of impact-related damage is to cage the entire vehicle in a protective structure that absorbs the force of collisions: otherwise known as collision-tolerance. While this has proven useful in working environments with enclosed spaces that are generally unsafe for workers, it cannot provide the level of precision offered by smart drone collision avoidance systems.

Intuitive drone collision avoidance first requires an accurate picture of the drone’s position relative to static and dynamic obstacles. Conventional positional technologies for outdoor UAVs include global positioning satellite (GPS), but GPS signals can be inaccurate. Some small obstacles like buildings and trees - even individual branches - can impede GPS signals, which means systems are not accurate enough to avoid small obstacles. Typical indoor applications like warehouses and factories can be even more challenging, as they may be GPS-denied.

The second thing to note about drone collision avoidance technologies is that they are a crucial step on the path towards true Intelligent Flying Machines (IFMs), but they are not a fully-autonomous flight solution. They encompass an array of high-resolution sensors with fast refresh rates, which feed positional data to the flight control system so that it can autonomously adjust vertical thrust or roll to avoid an object that has entered the sensor’s field of view (FoV). In effect, it assists the pilot to varying degrees, depending upon the technology’s specifications and limitations.

Several sensing technologies are available for drone collision avoidance: Infrared (IR); LiDAR; Monocular vision; Stereo vision; Time-of-flight (ToF), and more.

The operating principle of each of the above technologies varies significantly, but one of the preferred methods is to combine LiDAR with time-of-flight sensing. In a typical arrangement, two distinct sensing systems are integrated into the flight control electronics for robust drone collision avoidance:

A single time-of-flight sensor attached to the bottom of the drone functions as a precision altimeter to maintain a constant height; while a multidirectional LiDAR time-of-flight sensor array provides up to 360° positional monitoring.

These integrated systems enable the drone to minutely adjust its lateral position without ascending or descending due to pitch and roll. This is because the time-of-flight sensor acting as an altimeter holds the UAV at a fixed height with the assistance of autopiloting software. Vertical thrust is adjusted autonomously to maintain a desired height, even when the UAV pitches to change direction.

Meanwhile, the drone’s lateral position is adjusted by the multidirectional LiDAR array which simultaneously acquires proximity data from up to eight different axes at high speeds. Continuous proximity sensing feeds back to internal accelerometers that decelerate the UAV to avoid static or dynamic objects that enter this comprehensive field of view.

Terabee is a world-leader in the field of time-of-flight proximity sensing technologies and drone collision avoidance. We offer a broad range of drone obstacle avoidance systems for commercial and industrial UAVs alike. If you would like to learn more, read our technology insight article.