Microsoft presented the innovative AirSim project at the Farnborough International Air Show. The AirSim project uses high-fidelity simulation to safely build, train, and test autonomous aircraft on the Azure platform. It is essentially an artificial intelligence (AI) simulator for drones, flying taxis and other advanced air mobility (AAM) vehicles. A limited preview version of Project AirSim is available today.
Project AirSim customers will also be able to access a library of specific locations such as New York or London, or shared spaces such as an airport, and create millions of accurate 3D environments using data from Bing Maps and other providers.
What is the AirSim project? How it works ?
Automation is important in a rapidly changing technology world, and Microsoft wants to speed up autonomous flight by building an end-to-end platform with the AirSim project. Although AirSim is already available and a popular research tool, it required extensive knowledge of coding and machine learning.
Microsoft has now turned this open source tool into a comprehensive platform to make it easier for Advanced Aerial Mobility (AAM) customers to test and train AI-enabled aircraft in simulated 3D environments.
The power of Azure is used by the AirSim project to generate massive amounts of data to train AI models on exactly what to do at every stage of a flight, from takeoff to landing.
In addition to a suite of sophisticated and pre-trained AI models, it will also provide libraries of simulated 3D environments representing various urban and rural landscapes to accelerate last mile autonomous delivery, air infrastructure testing and urban air mobility.
Good news for the market
Commenting on the announcement, Gurdip Poll, Microsoft Vice President of Technology Business Incubation and Research, said:
Autonomous systems will transform many industries and enable a variety of aerial scenarios, from last-mile deliveries in congested cities to inspecting downed power lines from 1,000 kilometers away. But first, we need to safely train these systems in a realistic virtualized world. The AirSim project is an important tool that allows us to connect the world of bits and the world of atoms, and it demonstrates the power of the industrial metaverse – virtual worlds where companies create, test and improve solutions before applying them to the real world. World.
Through the AirSim project, manufacturers will be able to safely train future Advanced Air Mobility (AAM) systems in a virtualized world. Through the AirSim project, developers will have access to pre-trained AI modules, including sophisticated models for obstacle detection and avoidance, and precision landings.
By eliminating the need for in-depth knowledge of machine learning, these built-in capabilities expand the number of people who can start teaching autonomous aircraft. In addition, Microsoft said it is working closely with business partners to expand precision modeling to include weather, physics and, most importantly, the sensors that an autonomous machine will use to “see” the world.