Systems and methods for intelligently displaying aircraft traffic information
View Patent ↗There is presented systems and methods to present air traffic and other related hazards on a flight deck display in a manner that presents relevant aircraft that may be of interest to flight deck personnel at a future time, accompanied with decluttering the display by removing information pertaining to aircraft that are of lesser impact to the safe navigation of an ownship aircraft. The improved avionics system of the present invention can make decisions that are more intelligent on how to best present traffic information to a flight crew in a manner that improves upon the prior hardware processes and increases efficiency of operation of the avionics hardware and processing system by presenting relevant information and suppressing unnecessary computation and modification of display elements.
1. A method comprising
obtaining a plurality of aircraft track parameters pertaining to target aircraft external to and in a vicinity of an ownship aircraft;
determining, using the plurality of aircraft track parameters, a respective forecast position and a respective forecast track at time samples within a predetermined time period, for each of the target aircraft;
determining a forecast ownship position and a forecast position of each forecast track with respect to the forecast ownship position;
determining whether each of the forecast tracks crosses within a near-range threshold of the forecast ownship's position;
visually highlighting, on a display, each of the forecast tracks that crosses within the near-range threshold a of the forecast ownship position; and
based on receiving a traffic alert from an avionics traffic collision avoidance system of the ownship, visually highlighting, on the display, each of the forecast tracks that do not cross within the near-range threshold of the forecast ownship position.
2. The method of claim 1 , wherein the near-range threshold is adjustable by one of flight deck personnel or an airline technician.
3. The method of claim 1 , wherein the determining the respective forecast position and the respective forecast track for each of the target aircraft and the determining the forecast ownship position comprise using a neural network component.
4. The method of claim 3 , wherein the neural network component comprises one or more of an Encoder/Decoder Recurrent Neural Network (RNN), a Convolutional Neural Network (CNN) and a Convolutional Neural Network-Recurrent Neural Network Hybrid (CNN-RNN).
5. The method of claim 3 , further comprising training the neural network by
presenting the neural network with training data;
comparing pre-determined ground truth output sequences to forecasted output sequences computed by the neural network;
determining an error metric for each comparison;
based upon the error metric, back propagate changes to layers within the neural network; and
continuing presentation of the training data until a predetermined operational accuracy is determined.
6. The method of claim 5 , further comprising preparing the training data by:
obtaining a database of prior flight tracks in view of an ownship track;
segmenting available traffic data by flight;
cleaning and normalizing parameters comprising such as position and velocity;
and interpolating the parameters into uniformly-spaced time samples.
7. A system comprising an avionics system comprising one or more processors, a memory coupled to the processors, a top antenna and a bottom antenna electrically and communicatively coupled to the avionics system, and further comprising:
a Traffic Collision Avoidance System (TCAS) component,
a flight deck display electrically and communicatively coupled to the TCAS component; and
a prediction display management system coupled to the TCAS component and the flight deck display, the prediction display management system further comprising a predictive model component;
wherein software stored within a memory of the avionics system is configured to execute a method comprising:
obtaining a plurality of aircraft track parameters pertaining to target aircraft external to and in a vicinity of an ownship aircraft;
determining, using the plurality of aircraft track parameters, a respective forecast position and a respective forecast track at time samples within a predetermined time period, for each of the target aircraft;
determining a forecast forecasting an ownship position and a forecast position of each forecast track with respect to the forecast ownship position;
determining whether each of the forecast tracks that crosses within a near-range threshold of the forecast ownship position; and
visually highlighting, on a display, each of the forecast tracks that crosses within the near-range threshold of the forecast ownship position; and
based on receiving a traffic alert from an avionics traffic collision avoidance system of the ownship, visually highlighting, on the display, each of the forecast tracks that do not cross within the hear-range threshold of the forecast ownship position.
8. The system of claim 7 , wherein the near-range threshold is adjustable by one of flight deck personnel or an airline technician.
9. The system of claim 7 , wherein the avionics system further comprises a neural network component, and the determining the respective forecast position and the respective forecast track for each of the target aircraft and the determining the forecast ownship position are performed by the neural network component.
10. The system of claim 9 , wherein the neural network component comprises one or more of an Encoder/Decoder Recurrent Neural Network (RNN), a Convolutional Neural Network (CNN) and a Convolutional Neural Network-Recurrent Neural Network Hybrid (CNN-RNN).
11. The system of claim 9 , wherein the method further comprises training the neural network component by:
presenting the neural network with training data;
comparing pre-determined ground truth output sequences to forecasted output sequences computed by the neural network component;
determining an error metric for each comparison;
based upon the error metric, back propagate changes to layers within the neural network component; and
continuing presentation of the training data until a predetermined operational accuracy is determined.
12. The system of claim 11 , wherein the wherein the training the neural network further comprises preparing the training data by:
obtaining a database of prior flight tracks in view of an ownship track;
segmenting available traffic data by flight;
cleaning and normalizing parameters comprising position and velocity;
and interpolating the parameters into uniformly-spaced time samples.