Improved De-cluttering of Aircraft Cockpit Traffic Displays

Commercial aircraft have a transponder to transmit position, altitude and velocity to other aircraft that helps pilots determine potential risk of aircraft collisions along their flight paths. In high-density areas, the traffic display becomes cluttered with information, making it difficult to use.

This project presents new software to predict aircraft trajectories. It displays only the most relevant information to assist pilots during flight, increasing situational awareness and reducing the risk of collisions.

The team expanded an existing machine learning algorithm to accurately predict aircraft trajectories and implemented several techniques to declutter pilot traffic displays. The software design, developed using Python, continuously creates projected paths for all available aircraft up to three minutes into the future and updates the traffic display every five seconds. The design compares future flight paths to determine whether any aircraft will be in the traffic advisory area minutes before a warning is announced. Mean squared error calculations determine the accuracy of predicted values against their true values, and the display reverts to unfiltered data when accuracy is below 96%. The user can toggle between multiple display options within seconds. The students’ Decluttered Aircraft Traffic Display System is expected to show predicted collision risks before the traffic collision avoidance system calls out warnings.