In a future that could likely be dominated by automation, drones have a lot to offer. They have greater mobility than human workers, and can reach a variety of places that humans and stationary machines have trouble accessing. Once they reach those areas though they need to be able to recognize, interact with, and/or productively manipulate whatever it is they were sent there to investigate. This is where machine learning comes into play.
Machine learning is exactly what it sounds like, even though humans have been taught to think machines can’t learn. It’s a method of training machines with artificial intelligence capabilities to learn in a similar way to how organic creatures such as humans learn. The machine learning algorithm is usually presented with a series of images or scenarios. With machine learning, eventually it learns to recognize and understand what it needs to do.
This has obvious advantages in the way drones will eventually be used in the real world. Drones can be sent into inaccessible areas to check for defects in structures and objects. They can be used to collect items in a warehouse. They can be sent to deliver the items they collected. They can even be used for bomb disposal. They can be used to monitor crowds and public areas and can then report any safety concerns. The potential for commercial and other applications for autonomous and semi-autonomous drones makes for a huge potential market for both drone manufacturers and businesses making use of drones.
Two companies that recognize this potential are drone manufacturer DJI, and software and hardware company Microsoft. DJI and Microsoft are currently working on developing drones with artificial intelligence that are capable of recognizing objects. Recently the two companies organized the AI x Drones Joint University Competition, which involved students from three universities in Hong Kong. The competition had the students making AI machine learning models to help a DJI Phantom 4 Pro drone recognize different types of fruit that it encountered.
The drone used in the competition was equipped with a 4K ultra HD capable camera which recorded video of the fruit. That video was then processed by the algorithms created by the students to recognize which fruit the drone was looking at.
Drones are unlikely to replace human anytime in the near future. More likely they will become a useful tool to help humans, allowing us to reach and interact in places and ways that are normally inaccessible or too dangerous. Projects like the one undertaken by DJI and Microsoft help move us towards the future and may eventually unlock the potential of commercial drones, but there’s still a great deal of work to be done to make this happen. But thankfully, many different companies besides DJI and Microsoft are also working on it and we can expect new technology to be developed for companies to take full advantage of drones.