A mid-air collision warning system: Performance comparison using simulated ADS-B, Radar and Vision sensor inputs

We compare the performance of a collision warning system when it is deployed with three different types of sensors. The paper demonstrates the use of an Unscented Kalman filter to robustly estimate the Time to Minimum Separation (TMS) and Absolute Minimum Separation (AMS) between two aircraft. This improves the Time For Evasion by approximately a factor of 2 when compared to the previous study. Three popular sensors (ADS-B, Radar, and Vision) in the context of sense-and-avoid for Unmanned Aerial Systems (UAS) are simulated using known error metrics in a MATLAB environment to examine the reliability of our Collision Warning System. The results reveal the advantages and disadvantages of integrating each sensor to the current system, including a comparison of the errors in the computed TMS and AMS. Finally, we demonstrate that our collision warning system is sensor agnostic, and capable of operation with sensor inputs of various qualities.