Autonomous Landing Guidance Assist for Aircrafts in Zero Visibility Conditions Using Model Based and AI Methods

This paper showcases the application of model-based method and Artificial Intelligence algorithms for assisting air crafts landing in low visibility conditions. Navigating an Aircraft during landing in inclement weather is a critical and strenuous event. Airports that aren't equipped with Category CAT III C Instrument Landing System (ILS) equipment's require pilots to use great precision and piloting skills along with high situational awareness. Apart from this when there is a sensor malfunction it becomes extremely difficult to estimate the aircraft dynamics. This paper showcases a system designed for autonomous landing during sensor malfunction and poor visibility conditions. The proposed system helps in giving visual cues to the piloting crew about the expected pitch of the aircraft during the final approach for landing, considering factors like cross wind, low visibility, loss of power, loss of altitude and runway. Though ILS and GPS provide accurate positioning in terms of elevation and azimuth to the Autopilot system, sensor malfunction and power loss can lead to a loss in accuracy during final descent. This situation is very challenging especially during low cloud ceilings in both Visual Flight Rules (VFR) and Instrument Flight Rules (IFR) conditions. In this paper the Kalman filter estimator is used for determining the pitch of the aircraft in conjunction with the Artificial neural network algorithm that detects objects on the runway, guidance markings and threshold lights. For designing the proposed pitch estimation system, typical aircraft parameters related to longitudinal dynamics are taken into consideration. This is thus used in the design of a Kalman filter that is applied in estimating the position of the aircraft. Apart from this, prior data collected from Airports and runway parameters are used for training the AI algorithms to detect Foreign Object Debris, guidance lights and unwanted objects.