Canadian Airspace Models

This repository includes statistical information on 12 Canadian airports, along with the models of different aircraft types. These machine learned models were trained by the National Research Council Canada (NRC) and Carleton University.

As the project continues to develop over the coming years, additional work is planned to be scoped for improvements and applications of the presented statistical models, to enable further efforts to ensure the safe and smooth integration of Remotely Piloted Aircraft Systems (RPAS) into the Canadian airspace.

This project is an ongoing, multi-year collaboration of the NRC with Transport Canada (TC), Carleton University, and NAV CANADA, and was motivated by the work previously primarily completed by the Massachusetts Institute of Technology Lincoln Laboratory (MIT LL) for the Airspace Encounter Models project.

Files

Documents

The Documents/ folder contains the following reports:

1. LTR-FRL-2023-0055.pdf details the process used for the statistical model creation, including an overview of the dataset and steps to generate Canadian statistical models. The report also includes the application and findings of the developed methodology to derive the statistical information for the 12 selected airports across Canada.

2. LTR-FRL-2023-0053.pdf provides the detailed description of the airspace classes and sensor equipment requirements in Canada as well as in the 12 selected airports.

Statistical information

The Statistical information/ folder contains matlab scripts to generate the statistical distributions for the selected Canadian airports as described in LTR-FRL-2023-0055.pdf. To run the codes for statistical distributions, execute the Plot_Distributions_Main.m file.

Bayesian Network Models

The Bayesian Network Models/ folder contains frequency tables for initial and transition distributions of Bayesian networks for the Canada-wide statistical airspace model by the aircraft type (light fixed-wing (general aviation aircraft), medium fixed-wing (airliner), heavy fixed-wing (cargo aircraft), helicopter, gyrocopter, and ultralight aircarft). The developed frequency tables can be used for further encounter generation between an RPAS and traditional aviation using the em-model-manned-bayes repository from the Airspace Encounter Models organization that is primarily administrated by MIT LL. Note MIT LL has not yet independently validated the use of these models with em-model-manned-bayes. Some modification to em-model-manned-bayes may be required to use these models.

Track Generation Tool

The Track Generation Tool/ folder contains Python Tool that enables users to generate simulated tracks based on Canadian airspace models.

Support

For technical support, consider posting a question under Discussions.

Issues

Post Issues to report genuine bugs, mistakes or even small typos in the scripts. Note that issues are not meant for technical support. Please only open an issue for an error which is specific and reproducible.

Collaborators