14 DAY TRIAL // Microsoft has announced that it’s open-sourcing a simulator that its AI & Research teams created to train robots and gadgets that can move on their own, thus providing developers with an affordable way to perform testing in a super-realistic virtual environment.
Microsoft posted the code of the new project on GitHub and says that although it’s in beta, the software should work pretty well, simulating elements that exist in real life, including not only objects like trees, curbs, and doors, but also shadows, reflections and clouds.
“It’s all part of a research project the team dubs Aerial Informatics and Robotics Platform. It includes software that allows researchers to quickly write code to control aerial robots and other gadgets and a highly realistic simulator to collect data for training an AI system and testing it in the virtual world before deploying it in the real world,” the company explains.
Simulating a real-life environment
This solution comes in handy to researchers mostly because it serves as an affordable way to simulate the real world, so it allows for testing of robotic devices, including here drones, self-driving vehicles, and other gadgets, without involving the actual real life that could lead to high costs.
Microsoft says that researchers can thus crash their robots in this virtual environment without “burning through tens of thousands of dollars in equipment or damaging building or hurting someone.” Furthermore, it makes the whole process super-fast, as the collected data can be further analyzed to improve the robots to better respond to real-life environments.
The open-source simulator supports drones built on two platforms, namely DJI and MavLink, thanks to a library of software that comes built-in. “Normally, developers would have to spend time learning these separate APIs and write separate code for each platform,” Microsoft says.
The open-source software can work with any type of autonomous system, Microsoft explains, and it can substantially improve its navigation features in a real-life environment. More information on everything regarding this project can be found on GitHub, where Microsoft is also making the code available to all.