Recently, Microsoft has made the beta version available of an advanced virtual world for training autonomous drones in addition to other gadgets which would move on their own. The software, AirSim Simulator is now
available on GitHub, and it goes on to recreate conditions such as shadows, reflections as well as other potentially confusing real-world conditions in better details, with higher realistic virtual environment.
Image Source: blogs.microsoft.com
Ashish Kapoor, a Microsoft researcher who is leading the project, said,
“The aspirational goal is really to build systems that can operate in the real world.”
Microsoft informs that it hopes to help the “democratization of robotics” with this particular move, and it will go on to assist individuals, researchers as well as companies with the testing of systems that would otherwise be completely impossible, ot too resource-intensive for them to conduct on its own.
The main reason for testing drones and other self-navigating devices in a virtual world instead of the real is that it can stimulate testing in a wider and more affordable environment especially when part of your process is being encouraged to self-navigating software, in order to distinguish between things, such as shadows as well as solid, dark colored walls. In the real world, if you go on to ask a drone with an advanced onboard brain to fly into something which might be solid or which might be immaterial; possibility is that you might end up with a smashed and extremely expensive failure. However, if you conduct similar action in the virtual setting, all you will lose are few moments and some electricity powering of your PC.
Simulation also allows you to increase the volume as well as the speed of the testing and learning scenario more efficiently. However, if you want this to be an effective method of educating autonomous flight software, it is essential for it to be really accurate.
Microsoft, in its official blog, states,
“The Aerial Informatics and Robotics Platform could help developers make advances in planning capabilities, which aim to help the gadgets anticipate what will happen next and how they should respond, much like humans know to anticipate that cars will drive by when we cross a street. That kind of artificial intelligence – which would closely mimic how people navigate the real world – is key to building practical systems for safe everyday use.”