Mixed H2/H∞ filter for automatic aircraft landing with optical sensors

Abstract

The amount of automatic approach and landing systems is in recent years supplemented by area navigation and computer vision-based systems. The approach for image-aided final approach and landing is object of research in the project C2Land [1] led by the Institute of Flight Guidance of the Technische Universität Braunschweig. Object of study is an integrated GNSS/INS geolocation system supported by a long range laser scanner to aircraft-autonomously provide geolocation information. The additional sensors replace the classical navigation infrastructure on ground in order to be independent of international airport structures. This paper focuses on a mixed H2/H∞ design of the navigation system in order to enhance position accuracy. Least trace criterion [2] and a novel optimization technique are investigated to fit the sensor noise characteristics. With the resulting approach the tradeoff between Gaussian noise-optimal H2 and worst noise-optimal H∞ is found for the introduced navigation system.