Drones that can fly autonomously with the help of AI is currently in massive demand for businesses, entertainment, and personal use. However, most drones fall into the category of remote-controlled and therefore have to be manned every time they are in use. Now, two startups are announcing that they are presently developing drones that are capable of flying all by themselves.
Skydio and Exyn are two of the startups that are competing to build and market drones that can autonomously figure out its own flight path without crashing or dropping from the sky. Both young companies want to integrate the use of artificial intelligence in operating their drones. Additionally, both also see a powerful potential for drones to be put to use in places where it’s not safe or easy for humans to work.
Skydio's entrepreneur Adam Bry believes that AI-powered drones are what's missing in today's market. Using deep-learning technique, not unlike Google's use of the system, the startup can power image search feature of drones to help identify and track people against their background. According to the Technology Review, this capable machine vision will help drones keep track of their targets while avoiding potential obstacles.
The automated flying software can give the flying robots flawless flying capability, increasing its usefulness and dependability. It can also use cues from a person’s appearance and movement to avoid confusing individuals near one another. However, Bry admitted that their first product isn’t yet perfect, citing that machines can also have trouble telling two people apart.
Similarly, Exyn founder Vijay Kumar said that the system for AI-powered drones uses an even more complex circuitry than for self-driving cars or ground-based robots. According to TechCrunch, the startup has shown in their demo video a drone that flies autonomously, even in dark, obstacle-filled environments or beyond the reaches of GPS. It aims to further develop the use of sensor fusion to give drones situational awareness much like a human’s.
