Real-time Wind Calculations based on AI and CFD

Un-crewed Air Vehicles (UAVs) have many applications in different sectors, such as security, surveying, and logistics. However, the flight autonomy of the UAVs is a key factor that can limit their mission potential. Shearwater Aerospace is a Canadian company based in Montreal that develops autonomous operating software based on artificial intelligence, Smart Flight, for professional and commercial drones. Smart Flight improves drone capabilities, allowing them to achieve longer flight durations, higher speeds, and more frequent operations through the integration of advanced artificial intelligence and wind-powered autonomy. The proposed research project aims to expand the capability of Smart Flight by using a Neural Network (NN) that can estimate the wind speed over any terrain in real-time. The NN will produce the three-dimensional wind velocity field over a specific area that is given as an input. The NN would be much faster than the Computational Fluid Dynamics (CFD) simulation where a velocity field will take about one second to generate. A Convolutional Neural Network (CNN) that is trained on CFD data based on a high-fidelity turbulence LES model. By enhancing Shearwater’s ability to measure the wind speed and direction before and during flight operations, they will be able to give drone operators accurate estimates of the flight performance. This is vital for successful mission planning, as well as improving flight safety, operational efficiency, and overall performance. The project will also lead to a significant increase in flight time and a decrease in energy consumption.

Faculty Supervisor:

Marius Paraschivoiu

Student:

Partner:

Shearwater Aerospace

Discipline:

Engineering

Sector:

Information and cultural industries; Professional, scientific and technical services

University:

Concordia University

Program: