How do you teach an autonomous drone to fly itself? Practice, practice, practice. Now Microsoft is offering a way to put a drone’s control software through its paces millions of times before the first takeoff.
The cloud-based simulation platform, Project AirSim, is being made available in limited preview starting today, in conjunction with this week’s Farnborough International Airshow in Britain.
“Project AirSim is a critical tool that lets us bridge the world of bits and the world of atoms, and it shows the power of the industrial metaverse — the virtual worlds where businesses will build, test and hone solutions, and then bring them into the real world,” Gurdeep Pall, Microsoft corporate vice president for business incubations in technology and research, said in a blog posting.
The platform can take advantage of data from Bing Maps and other providers, plus processing power from Microsoft Azure, to create millions of detailed virtual 3-D environments. Customers can build their own spaces, or draw from a library of specific locations around the globe.
Microsoft’s Project AirSim builds on an earlier open-source tool from Microsoft Research that required deep expertise in coding and machine learning. The newly unveiled commercial platform makes it easier to test and train AI-powered aircraft. Partners in the project include Ansys and MathWorks, two companies that provide software for designing AI-based simulations.
One of Project AirSim’s early-access users is Airtonomy, a North Dakota-based company that received a $100,000 TechSpark grant from Microsoft and has been training drones to inspect wind farms and oil tanks across the Midwest.
“You don’t want to fly drones into wind turbines, power lines — or really anything, for that matter,” Airtonomy CEO Josh Riedy said. “Coupled with the fact that winter can literally last seven months in North Dakota, we realized we needed something other than the physical world to design our solutions for customers.”
Project AirSim helped Airtonomy train drones to cope with environmental conditions ranging from snow and rain to strong winds and high temperatures.
During a recent test flight for a NASA project aimed at accelerating the integration of drones into the national airspace, Bell proved that its Autonomous Pod Transport could maintain contact with ground-based radar monitoring systems while flying through a corridor in the Dallas-Fort Worth area.
“AirSim allowed us to get a true understanding of what to expect before we flew in the real world,” said Matt Holvey, Bell’s director of intelligent systems. “It’s going to be one of the tools that will accelerate the timeline for scaling aerial mobility. If we have to test and validate everything by hand, or in a physical lab, or on a flying aircraft, we’re talking about decades, and it’s going to cost billions. But Project AirSim pulls that forward through high-fidelity simulation.”
Microsoft’s Pall said that the company is working with civil aviation regulators to figure out how Project AirSim could help with the certification of autonomous flight systems. For example, the flight software could be put through a series of simulations to check on how it handles bad weather or loss of GPS connectivity.
“A ton of data gets generated when an aircraft flies through space in Project AirSim, said Ashish Kapoor, who created the original AirSim at Microsoft Research and is now general manager of Microsoft’s autonomous systems research group.
“Our ability to capture that data and translate it into autonomy is going to significantly changes the landscape of aviation,” Kapoor said. “And because of that, we are going to see many more vehicles in the sky, helping to monitor farms, inspect critical infrastructure and transport goods and people to the remotest of places.”