In recent years the drones have shown that they are here to stay. Today they are much more affordable, able to fly more stable longer and still have a host of other features that were not possible before. In the not too distant future, these aircraft will occupy the sky of the busiest cities on the planet. But what will they be?
Today's drones offer much more for a much lower price.
Fly a drone It is not a very simple task and requires some practice with the control. After all, they are not cheap and no one wants to miss one by dropping or throwing against the wall. They can be used for different tasks – recording tourist spots, a simple walk, journalistic coverage and even carrying objects. The problem has to do all this using a control.
Device cost about $ 20,000
Taking care of the drone so that it does its job a little too lightly and is not ideal for the future. And the one that comes in Skydioby investing in drone technology autonomous and intelligent.
Imagine transporting objects using drones: If carriers need to hire one person to control every ship that leaves to deliver, by the end of the month it may be expensive to keep so many employees. Thus, it is to be expected that investment in drones capable of doing the job alone – just following a schedule – is growing more and more.
As Skydio r1We can see that this is possible and not so far away, this is a quadcopter that surprises for its autonomy. In the video below you check out two Skydio R1 following an acrobat in your path. They are not being controlled by anyone and yet they can keep up with it perfectly. The result is a recorded clip with high stability and excellent camera movements.
The company started as a small startup in Redwood City, California and surprised when it presented the Skydio r1, a quadcopter with artificial intelligence able to record in 4k focusing on a moving target even in highly complex environments.
Although they have made a drone, they do not call themselves a company unique to this type of product. According to the note on their website, they started with drones for the freedom offered by them in 3D motion and the connection these machines allow between the physical and digital world.
We are building a future where they (the drones) are smart and elegant flying assistants. – Skydio through your blog
No doubt Skydio very ambitious and seems highly confident of his work. As your website states, the main goal is to constantly push the boundary of what is possible in the connected world we live in, exploring technologies and creating things we never imagined possible.
Even though their inclination is the technology in general, one would expect them not to stop making and improving their drones. The development of this first project took about five years and during this period they worked only with drones.
The Skydio R1
The R1 has stolen the attention of tech admirers and drones around the world. This is because even drones that have been on the market for a long time are not as good at flying alone as him. Other models use simpler programs, such as GPS to locate and sensors to prevent collisions. The vast majority who can follow the controller only allow this to happen in low-complex environments with a lot of space.
If you look at the demonstration videos of R1, you will see it drifting from trees and other object types and adapting in different directions. This is because it extremely complex.
The Skydio R1 does not come with control and flies alone completely autonomously.
Adam Bry, co-founder and Skydio CEOexplains how R1 works. The development of the software was fundamental to the operation of this machine, but all the programming done by the team is based on a powerful chip specifically designed for autonomous functionality: Nvidia Tegra X1.
Drone gained enhancements that bring more stability at the time of use
Developed primarily for cars, this superchip has a processing capacity greater than 1 teraflop. Without this card, it would not be possible to process data from 13 cameras positioned around the R1, take texture and depth readings and calculate speed and routes with very accurate precision.
The first step in smart device processing is to identify where the target is and where it is moving.
The NVIDIA chip provides all the processing necessary for it to do so well on its own.
Using the cameras, the software relies on high quality textures to mark the ground on incoming images and then tracks what is moving. This is the basis of all the following steps. Then the image received by the 13 cameras are compared between them and merged to draw the 3D world. These readings are constantly being updated by the software in conjunction with the chip. To learn more about how it works, I recommend watching this video broadcast by the company on its YouTube channel.
Skydio R1 can reach up to 40km / h, have 64GB of memory for filming on 4k and your battery guarantees flights of at 16 minutes. Also, remember that it comes with the chip Nvidia Tegra X1 It is a revolutionary drone if we consider its artificial intelligence. This is not a product for any consumer and is considered one of the best drones in the world today, bringing the latest in autonomous technology.
By now you might be wondering how much this joke costs. As I write this article, it is costing $ 2,499 – about R $ 8 thousand. And that's in an international scenario, around here it will be around 15k.
We see a future of independent cars, robots and drones that see and learn with intelligence beyond imagination – Jen-Hsun Huang, co-founder and CEO of Nvidia
Skydio R1 is not made for filmmakers or photographers, but that doesn't mean it can't be used by them.
In the Skydio R1 publishing and demonstration videos, I saw some comments showing dissatisfaction with the camera's recording quality or the limitation of the aircraft not coming with a control. The problem that Skydio was not made for those who want to control it or specifically say where it should go, was not meant for filmmakers, which also does not prevent it from being used for this function.
Since the graphics chip is unchanged and already very powerful, the updates are up to Skydio, which from time to time provides an update to improve the drone's intelligent responses. One of the biggest achievements so far has added the ability to follow cars, as you can see from the video below.
I believe the important thing is to have this drone, the technology it carries. We already have highly accurate drones with a controller and we also have amazing cameras available on the market, just a matter of time before this combination comes up with a highly autonomous drone like the R1. However, the quality is very satisfactory and fits perfectly what was done.
The Skydio 2
After a year and a half of the R1's launch, Skydio announces another ship with plenty of potential to change the drone market – the Skydio 2. This time, the company bet on a different design and high quality technologies that allow even a child to pilot it. And still claims that the drone "Easier to ride and harder to hit in the world."
Even a child can fly Skydio 2. If she can hit, the repair is free.
Compared to its predecessor, the drone 50% smaller, lighter and quieter, plus 43% more battery life with its 23 minutes. Managing your autonomous resource is up to the Tegra X2 processor from Nvidia, that creates the 3D model of the refresh rate environment greater than 1 million points per second and helps the drone identify up to 10 objects at once, deflecting brilliantly at speeds at 60km / h.
Another great update is on algorithm. While R1 uses the stereo depth perception more traditional in 12 VGA cameras, Skydio 2 is backed by an algorithm based on Deep learning only six 4K cameras for the navigation system.
Skydio claims he is the "easiest to use and hardest to hit planet drone"
What people do by combining understanding of the contextual scene with photometric information – when we look at a scene, we know what things are and generally how the world is structured around us. And the combination of this information means that people and animals essentially have a perfect depth perception based only on sight. We created a deep learning algorithm that mimics that (…). " – Adam Bry.
The change brings a lot of advantage because, as already mentioned in this article and explains CEO Adam Bry, the perception of stereo depth goes well in simple scenes, but begins to fail when we talk about repeated or texture-free structures. However, it was not so easy for Skydio to achieve this feat with half the number of R1 cameras.
"(…) we have moved from these conventional global shutter cameras to the rotary shutter fisheye cameras, which are very challenging from an algoremic point of view. But the big advantage is that at a lower price we can get very good performance." best of the sensor. " – Adam Bry.
Still talking about camera, the main system includes a sensor Sony IMX577 which promises great quality images (and I'm not just talking about resolution) in 10bit HDR, 4K 60fps and still 12.3MP photos – all managed by a Soc Qualcomm Snapdragon 605.
And for those who were dissatisfied with the lack of a controller on R1, Skydio 2 has up to 3 options. One is still via smartphone (both Android and IOS), but now the (co) pilot can also opt for a joystick style controller or the new Skydio beacon. This last one looks like a television control, but it does bring some very interesting features like, for example, a GPS locator for Skydio 2 to find you even if you are not "seeing" it.
And as could not be missing, we will end here with values. The drone hits the market for $ 999, which is a significant reduction from the $ 2,500 of R1.
With this combo of news, it's a surprise that Skydio 2 comes as the big bet to leave the Mavic 2 Pro and the DJI With wide open eyes. A little competition is always good for the end consumer, and we will certainly see more news coming up. What you can do from now on is to watch how the drone market revolution moves.
Source: Nvidia, Skydio. (TagsToTranslate) drones (t) skydio (t) skydio r1 (t) drone 4k (t) standalone drone (t) nvidia (t) skydio 2 (t) sony (t) qualcomm (t) skydio beacon (t) tegra x1 (t) tegra x2 (t) camera drone (t) drone price (t) professional drone