{"id":359810,"date":"2017-02-15T06:00:48","date_gmt":"2017-02-15T14:00:48","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-project&p=359810"},"modified":"2022-10-12T15:51:28","modified_gmt":"2022-10-12T22:51:28","slug":"aerial-informatics-robotics-platform","status":"publish","type":"msr-project","link":"https:\/\/www.microsoft.com\/en-us\/research\/project\/aerial-informatics-robotics-platform\/","title":{"rendered":"Aerial Informatics and Robotics Platform"},"content":{"rendered":"
\n\t
\n\t\t
\n\t\t\t\"aerial\t\t<\/div>\n\t\t\n\t\t
\n\t\t\t\n\t\t\t
\n\t\t\t\t\n\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\n\n

Aerial Informatics and Robotics Platform<\/h1>\n\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t<\/div>\n\t<\/div>\n<\/section>\n\n\n\n\n\n

In 2017 Microsoft Research created AirSim (Aerial Informatics and Robotics Simulation) \u2013 an open-source robotics simulation platform. From ground vehicles, wheeled robotics, aerial drones, and even static IoT devices, AirSim enabled data capture data for models without costly field operations.\u200b<\/p>\n\n\n\n

Over the span of five years, the open-source AirSim research project served its purpose and is now archived in anticipation of a new aerial autonomy simulation platform. Users can still access the original AirSim code, but no further updates will be made. For more information about migrating to the new platform, please visit the GitHub (opens in new tab)<\/span><\/a> repo.\u200b<\/p>\n\n\n\n

Read on to learn more about the AirSim research project.<\/p>\n\n\n\n

Bridging the sm-to-real gap with AirSim<\/h2>\n\n\n\n

Microsoft AirSim (Aerial Informatics and Robotics Simulation) is an open-source robotics simulation platform. From ground vehicles, wheeled robotics, aerial drones, and even static IoT devices, AirSim can capture data for models without costly field operations.<\/p>\n\n\n\n

AirSim works as a plug-in to Epic Games\u2019 Unreal Engine 4 editor, providing control over building environments and simulating difficult-to-reproduce, real-world events to capture meaningful data for AI models.<\/p>\n\n\n\n

Machine learning has become an increasingly important artificial intelligence approach in building autonomous and robotic systems. One of the key challenges with machine learning is the need for massive data sets\u2014and the amount of data needed to learn useful behaviors can be prohibitively high. Since a new robotic system is often non-operational during the training phase, the development and debugging phases with real-world experiments face an unpredictable robot.<\/p>\n\n\n\n

AirSim solves these two problems: the need for large data sets for training and the ability to debug in a simulator. It provides a realistic simulation tool for designers and developers for seamless generation of the amount of training data they require. In addition, AirSim leverages current game engine rendering, physics, and perception computation to create accurate, real-world simulations. Together, this realism, based on efficiently generated ground-truth data, enables the study and execution of complex missions that are time-consuming and\/or risky in the real-world. For example, collisions in a simulator cost virtually nothing, yet provide actionable information for improving the design of the system.<\/p>\n\n\n\n

\n