TCAS (Traffic Collision Avoidance System) forms one of the crucial safety pillars in air traffic management. It has been designed to detect near mid-air collisions, provide advisories to the pilot and, ultimately, reduce the incidence of mid-air collisions between aircraft. Increasing traffic demands and the forthcoming use of drones in non-segregated air space poses novel evaluation challenges to TCAS. Important limitations of existing TCAS models are: that their model is specified through their computer code only, that uncertainties are not or only partially captured, and the difficulty in incorporating model changes in interactions with novel airspace users. These findings motivated the development of an agent-based modelling and simulation approach that takes various uncertainties well into account.

The core development consists of a mathematical model of TCAS II specifications according to the internationally accepted operational performance standards of TCAS II and airborne collision avoidance system (ACAS). This model has been completed with other agents and their interactions, has been implemented in MATLAB, and has been validated against EUROCONTROL’s deterministic TCAS simulation model InCAS.

The novel TCAS model considers the following uncertainties: errors in relative position measurements, errors in own rate of climb/descent estimates; variation in pilot response time; probability of reception of an interrogation response; variation in TCAS computation cycle. To assess the effect of these uncertainties on TCAS operations, several air traffic encounter scenarios are considered. For each scenario, rare event Monte Carlo simulations have been conducted. The results of the simulations show that the considered uncertainties may have a significant impact on the safety improvement by TCAS that often deviates from prior expectations. For example, the simulation results show that uncertainties may induce mid-air collisions in scenarios where continuation of the original aircraft flights would not have resulted in a mid-air collision. An overview of the results obtained will be given for current TCAS. In addition it will be explained what this means for future collision avoidance developments.