Today’s Microsoft Corp. publication Launch Project AirSimA powerful physics simulator that replicates the real world in a 3D digitized environment to quickly and safely train autonomous aerial drones.
The best way to train self-driving cars is in the field where you can experience and learn from the same conditions that artificial intelligence encounters every day. But the real world is also a dangerous place where you can damage or destroy your equipment before you learn how to handle it.
How do you teach artificial intelligence to operate drones to deal with flock of birds, power lines, severe gusts and then throw them into the real world? You put it in a simulator that is as loyal to the world as possible, then run as many different scenarios as possible and let them understand how to handle them.
Project AirSim, Microsoft’s solution to this problem by training Drone Pilot AI in its own realistic metaverse, is now available to interested customers in a limited preview.
“Autonomous systems are transforming many industries, enabling many aerial scenarios, from delivering last one mile goods in crowded cities to inspecting downed power lines from 1,000 miles away,” Microsoft said. Gurdeep Pall, Corporate Vice President of Technology Business Incubation, said. And research. “But first, we need to safely train these systems in a realistic virtualized world.”
The platform is cloud-based and uses the processing power of Microsoft Azure along with data retrieved from BingMaps to create vast global maps of known environments with millions of data points. You can also get data from a huge external pool and quickly create realistic models based on 3D models and mapping data.
To get the job done, Microsoft is working closely with industry partners on global scale such as weather, dust, smoke, lighting conditions, and other factors that can affect drones as they move through the air. Simulated physics. Sensor simulation software company Ansys has worked with Microsoft to bring expertise to a high-fidelity physics simulation platform. The MathWorks helps customers deploy their own physical models in AirSim through the graphical modeling programming environment Simulink.
AirSim was born from 2017 project Built as an open-source cross-platform simulator designed for drones, cars, etc. using Unreal Engine, you can use a flight controller to support AI and humans. It was a powerful research tool that required coding knowledge and machine learning training to use.
The new Project AirSim is designed as an end-to-end platform with pre-trained models for AI drones, requiring no advanced coding knowledge or expertise. By already incorporating the components of AI operations such as mid-air collision avoidance, ground tracking, and emergency landing into the simulator, you can land quickly.
Balinder Malhi, Engineering Lead for ProjectAirSim, said: “That is, the ability to accurately simulate the real world, capture and process large amounts of data, and encode autonomy without the need for deep AI expertise.”
Airtonomy, a participant in Project AirSim’s Early Access program, is using AI-controlled drones in remote areas where technician safety is a concern by allowing its customers to inspect critical infrastructure. , I used the platform like this. With AirSim, you can deploy your drone cheaply, even if you don’t have much machine learning expertise.
Josh Riedy, Chief Executive Officer of Airtonomy, said:
Another advantage of AirSim is that it can simulate AI drones at incredible speeds of seconds on millions of flights. As a result, it is a powerful research tool because it can generate and capture vast amounts of data that sensors cannot capture in real-world situations.
This project is well-positioned to connect with the new landscape of the so-called “Industrial Metaverse”, the growth of interconnected industries of simulation technology that recreates the real world in the 3D digital world. This can be done efficiently by recreating complex physics objects and simulating them before they are created in the real world, or by using digital twins such as factory floors to simulate the real world environment. Includes features to improve. Nvidia and Siemens.. It can also be used to simulate an entire city, road network, or other large environment.
“Project AirSim is an important tool that bridges the world of Bit and Atom. It demonstrates the power of the Industrial Metaverse, a virtual world where companies build, test, polish and make solutions a reality. The world, “Paul said.
With the platform’s technology and its surrealistic simulation capabilities, Microsoft plans to work with civil aviation regulators around the world on how to use the platform to provide self-driving car certification.