Autonomous drones are one of the most exciting technologies in the drone industry, and they have so much potential that they will soon create a whole new generation of new businesses.
Imagine that you have an autonomous drone and you give it an order: cross that forest through the trees.
Then the autonomous drone will fly directly towards the trees, and thanks to its sensors it will be able to understand its surroundings and thus avoid each trunk of each tree, until it reaches the other end of the forest.
And without crashing.
What is an autonomous drone?
By the way, an autonomous drone is a drone that has the ability to fly making its own decisions on how to move. It may sound simple, but it is not.
We don’t have to confuse it with scheduled flight mode, which is when we give a flight plan to the drone, and then the drone executes the route automatically. This happens for example when we do digital photogrammetry.
The autonomous flight mode is different, and it is that the drone has the ability to understand its environment, in order to move efficiently to meet an objective.
To achieve this, an autonomous drone must have sensors to capture data from its environment.
You must also have computer vision algorithms to understand what is around you. And finally there must be an objective, which is normally delivered by a human.
At least for now. We’ll see what happens next year.
The 4 elements of an autonomous drone
Therefore, when we talk about an autonomous drone, we talk about 4 elements:
- sensors to capture data from the environment
- computer vision algorithms to analyze that data in real time
- artificial intelligence algorithms to make decisions
- a target given by a human (at least for now)
With these 4 elements, we already have an autonomous drone.
Let’s look at a fictional example (at least for now).
In 2012, a science fiction film by Ridley Scott called Prometheus was released.
There is a scene where a group of humans enter a maze of caves. They are in that place for the first time and since they don’t know those caves, the drone operator opens his backpack and extracts some autonomous drones.
The drones take off and automatically start exploring the caves, while scanning their surroundings.
In this scene, the drones are making their own decisions where they are going to move, since they understand their environment and their objective is clear.
While the drones are flying, they are sending the data of the topography of the caves to a central, where there are humans who are witnessing on a 3D screen, how the entire labyrinth of caves is being drawn in real time.
That is the magic of autonomous drones.
What does the autonomous drone market say?
Just to get an idea of the new business potential, the market study published in FactMR, by the year 2033, the autonomous drone market will be worth $91 trillion dollars.
That’s a lot of money.
If you start an autonomous drone business now, by the year 2033 you could be at the beach enjoying the sun and sand, while a drone brings you a pina colada.
MIT Liquid Neural Networks
Most likely, you have already realized that there is a close relationship between autonomous drones and artificial intelligence.
And since artificial intelligence is all the rage, autonomous drones will soon be all the rage.
Here is an example.
Last April, the MIT Laboratory for Computing and Artificial Intelligence announced that it was working on creating autonomous drones. And for this they are using liquid neural networks.
These liquid neural networks are capable of quickly analyzing the pixels that the drone’s sensors are seeing, analyzing what is relevant and discarding what is not.
This allows them to quickly adapt to unfamiliar surroundings.
The future of autonomous drones
The impact in the future is still hard to imagine. It will impact industries such as agriculture, security or logistics, among many others. They may even save lives.
Let’s imagine that a couple of people have been lost in the woods.
Then the search and rescue party arrives, and they order their autonomous drones to start searching the forest for humans.
Within minutes the data of the coordinates of the location of the lost humans will arrive.
We are talking about saving lives.
If you want to learn more about how liquid neural networks work for autonomous drones, I leave you this link from MIT: https://www.science.org/doi/10.1126/scirobotics.adc8892
And if you want to know how drones are being used in the industry, I invite you to listen to the most famous Spanish-speaking drone podcast: https://dronity.com/podcast/