| Severe low-altitude wind shear poses a significant threat to air transportation safety. Concepts for assisting critical decision making under uncertainty are advanced to promote the avoidance of hazardous weather, particularly microburst wind shear. Computational strategies founded on probability and optimal estimation theories enable flight deck integration of diverse forecasting and detection systems, from airport weather information services to airborne forward-looking wind sensors.;A decision-making policy for wind shear is developed from a comprehensive investigation of microburst phenomenology, its observed characteristics, and its effects on aircraft flight. Existing avoidance guidelines for wind shear are extended to exploit the latest available technology, such as Doppler weather radar and lidar. Theories for probability-based decision making facilitate real-time computer reasoning with dynamic, conflicting data from a wide array of sources. Bayesian neural networks fused with multivariable estimators account for the limited precision, reliability, and timeliness of correlated sensor measurements. Monte Carlo analyses are conducted to refine a disturbance prediction Kalman filter for forward-looking sensors, with statistical results completing their incorporation into Bayesian reasoning.;Symbolic and numerical processes for a Wind Shear Safety Advisor are implemented and evaluated. A risk assessment model based on empirical and analytical results is used to compare the relevance of available wind shear information sources. Simulations of the risk-assessment model show its insensitivity to parameter variations. Validations of overall Wind Shear Safety Advisor logic illustrate how it conveys beneficial advance warnings in rapidly developing microburst-encounter situations. These results prove that intelligently-integrated detection systems can warn pilots of threatening wind shear sooner, more frequently, and more effectively than isolated systems can. |
PDF Full Text Request