photo: SIGMA AI DRONE, prepared by Muhammet Emin Yanik
There’s no denying that those who were 10 year old boys 10 to 20 years ago, those who loved playing with RC cars they were gifted for birthdays, are largely the same people that drool over the new drones. Both are remotely controlled, both are battery powered, but the similarities end here. While RC cars can definitely be expensive, for drones the hefty price tag is just a feature.
Owning one is a risky investment, because your monthly check can disappear just like your newly-bought toy does in the bushes when its propellers shatter after unexpected impact with an oak tree. It doesn’t matter how good of a pilot you are, we all make mistakes, but your seldom unpredictable clumsiness can become a non-issue thanks to intelligent drones.
Drones, however, are much more capable devices, photos and filmmaking aren’t their only purposes. Security and education are other fields that highly benefit from the existence of those. Let’s discuss the benefits now so you can see for yourself how intelligent drones can shape up the future of consumer and professional unmanned aerial vehicles.
Alongside its impressive, portable size, the Hover Camera Passport is one of those intelligent drones that is capable of fully autonomous flight and following the user. It uses a separate sensor dedicated to its tracking abilities and another one, dedicated photo and video shooting sensor. The whole device folds to help transport it in a backpack, but when you go about unfolding it, the software is so advanced it can autonomously follow you indoors. Operating in confined spaces is quite a challenge, but the team behind this ai drone did a very thorough job crafting and perfecting its self-piloting abilities.
Let’s now take a step back from consumer electronics and check out a drone that, with a single camera, is capable of self-navigating, obstacle avoiding, and accurate image recognition. The device was developed as a submission to NVIDIA® Jetson™ Developer Challenge. Its main purpose is wildfire detection, which, admittedly, is a pretty amazing idea considering how much safer it makes firefighters’ jobs. That is, however, not the reason why it’s being mentioned here. If this intelligent drone is capable of doing all those things with just a single camera, and with such high precision, imagine the potential of introducing its capabilities to consumer and professional electronics. Some professional cinematic drones incorporate second camera, main one being for capturing the shots, controlled by one operator, and the additional being the way for a second person, a pilot, to navigate the drove safely. This all raises the costs and increases the weight of the device, that, in this case, is so crucial in this case.
Yet again, a drone whose main purpose is to be a platform for developing self-navigating capabilities, this one is called NVIDIA Redtrail. This project had a goal of making a system capable of traversing through a forest trail with no human guidance. Those paths are unmarked in most cases, and the only reference for the drone is differentiating between a beaten trail and leaves or branches. As those are very chaotic as compared to, say, tight city alleys, it takes quite a feat of engineering to execute successfully. Once again, such technology, if brought to consumer electronics, would prove extremely helpful. Just imagine, not a single damaged drone ever again, no matter how good or bad of a pilot you are.
Do you enjoy swimming in the ocean? Is taking aerial shots when on a surfing board your passion? Say no more, a drone that detects sharks is what you need. Australian team Scout Aerial created their intelligent drone to be a tool in Shark Detection system. We first saw this drone as a submission to NVIDIA® Jetson™ Developer Challenge. Their project , however, is no longer just in work-in-progress stage, their creation is fully operational and in use. Their efforts just go to show that the first connotation when hearing the word “drone” is just barely scratching the surface of what these devices are capable of. Sure, filmmakers are the main market, but it’s important to recognize other uses as well.
There’s no need to introduce DJI brand to any drone enthusiast, as they pretty much dominate the consumer UAV market. Having such an impressive brand recognition as they do, the AI features introduced in Spark mean consumers can get the glimpse into what deep learning algorithms can offer to everyone. The gesture controls system uses deep learning to recognize different shapes and movements a human hand can perform in order to translate those into instructions like record, stop, follow, etc. In addition to those, once you tell this intelligent drone to follow you, or reach a given point, it will automatically do so avoiding any obstacles it might encounter. It is worth nothing though, that obstacle recognition might not be as advanced as in something like previously mentioned NVIDIA Redtrail.
This autonomous drone is taking the approach of the more the better. It uses a 3D stereoscopic camera, lidar, radar, GPS, and a 9-axis sensor to accurately calculate its position and path it has to travel without crashing. Image recognition and obstacle avoidance are of an unmatched quality. In addition to all of that, the intelligent drone is capable of appealing, smooth cinematic movements for the best videos. SIGMA AI Drone was submitted as a project to NVIDIA® Jetson™ Developer Challenge. Their Sigma2 autopilot system is open-sourced and can be modified, improved, and used by everyone.
When the time comes for you to buy a new drone maybe consider waiting a bit for the technologies mentioned above to be introduced into consumer products. Your last drone’s landing wasn’t planned, it crashed miserably into a lamppost, so save yourself the trouble (and money) for yet another similar device. Give AI the chance to make your wallet happier!
For all things AI, and much, much more, visit ChallengeRocket.com and sign up for a hackathon!