Issue 45 | Uncrewed Systems Technology Aug/Sept 2022 Tidewie USV Tupan | Performance monitoring | Bayonet 350 | UAVs insight | Xponential 2022 | ULPower UL350i and UL350iHPS | Elroy Air Chaparral | Gimbals | Clogworks Dark Matter

7 Platform one Uncrewed Systems Technology | August/September 2022 Microsoft has launched a software- in-the-loop tool to build, train and test autonomous aircraft through high-fidelity simulation (writes Nick Flaherty). The AirSim tool was originally developed as a cross-platform open source system that combines the Unreal and Unity 3D display engines with popular flight controllers such as PX4 and ArduPilot in hardware in the loop with PX4 for physically and visually realistic simulations. The open source version has now been integrated into a fully tested simulator running in the cloud rather than using the Unreal or Unity plug-ins on a desktop computer. “In 2017, Microsoft Research created AirSim as a simulation platform for AI research and experimentation,” said Paul Stubbs, principal program manager for Microsoft AI, focusing on autonomous systems. “That project has served its purpose as a common way to share research code and test new ideas around aerial AI development and simulation. “Users will still have access to the original AirSim code, but there will be no further updates; instead, we will focus our efforts on a new product, Microsoft Project AirSim, to meet the growing needs of the aerospace industry. It will provide an end-to-end platform for safely developing and testing aerial autonomy through simulation.” AirSim was a popular research tool, but it required a lot of expertise in coding and machine learning, said Microsoft. The end-to-end version, Project AirSim, allows machine learning models to run through millions of flights in seconds, learning how to react to a wide range of variables. This synthetic data can be used to look at how a UAV would fly in rain, sleet or snow, and how strong winds or high temperatures would affect battery life. “Everyone talks about AI, but very few companies are capable of building it at scale,” said Balinder Malhi, engineering lead for Project AirSim. “We created Project AirSim to simulate the real world accurately, capture and process massive amounts of data and encode autonomy without the need for deep expertise in AI.” Developers will be able to access pre- trained AI building blocks, including models for detecting and avoiding obstacles and executing precision landings. Microsoft is also extending the simulation to weather, physics and the sensors used on a UAV. It is working with physics-based simulation tool developer Ansys on models that can be used in AirSim, and with MathWorks so that developers can import their own physics models to the AirSim platform using Simulink. Helicopter and UAV developer Bell has been using an early version of the tool to improve the ability of its UAVs to land autonomously. AirSim enabled Bell to train its AI model on thousands of ‘what if’ scenarios in a matter of minutes. “AirSim allowed us to get a true understanding of what to expect before we flew in the real world,” said Matt Holvey, director of intelligent systems at Bell. “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, 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.” Simulation AI-based testing for UAVs Project AirSim is an end-to-end simulator to train AI systems for UAVs AirSim will be one of those tools that will accelerate the timeline for scaling aerial mobility. If we have to test everything by hand it will take decades

RkJQdWJsaXNoZXIy MjI2Mzk4