Study Of Hazard Detection And Alert Algorithm Of Aircraft Environment Based On Probability Methods

Aircraft Environment Surveillance System (AESS) plays an important role in the flight safety. By the surveillance to aircraft environment it can generate the signal of early alarm to avoid incident and make the flight safe. This paper focuses on TAWS and TCAS in AESS. Based on probability methods, three issues of evaluation of TAWS forward-looking alerting threshold, evaluation of TCAS alerting threshold and hazard detection of TCAS are studied.By analyzing the principle of forward-looking alerting function, three kinds of uncertainties influencing threshold value are summed up. The aircraft's trajectory models with and without an alert are studied, using Monte Carlo simulation method, we can evaluate the alert threshold values of TAWS. The optimal threshold values of TAWS in simulations are identical with the practice.In this paper, by analyzing the position error of the flight trajectory, a stochastic ordinary differential equation for the aircraft motion is introduced to model the TCAS alert event. Through analyzing the principle of TCAS alerting function, five kinds of alarm output are summed up and the evaluation system with multiple indexes are established by System Operating Characteristic (SOC) curve. Using Monte Carlo simulation method, we can evaluate the alert threshold values of TCAS. The optimal threshold values of TCAS consist with the practice.Taking single path trajectory model and worst case trajectory model for example, we analyze the principle of definite alerting algorithm and probe the limitation of the definite alerting algorithm and then a probabilistic approach to aircraft conflict detection of TCAS are utilized. Based on the encounter model, real time computation of the probability of conflict is completed using Monte Carlo simulations. The conflict probability is utilized directly as the alert threshold criteria. This thesis compares different probabilistic and definite alerting approaches in various scenarios. By using probabilistic approach the false alarm decreases significantly.