NEURAL NET BASED PROCESSOR FOR ROBUST, HIGH-INTEGRITY MULTISENSOR AND SYNTHETIC VISION FUSION

摘要

The ultimate role of an "integrated enhanced vision system" (IEVS) will entail much more than presenting sensor imagery to the pilot on a head-up and/or head-down display (HUD/HDD). All-weather, multisensor image data will be combined and used to verity on-board database imagery; the latter then provides the pilot interface, which may take the form of sparse or iconic displays that are optimized from a human factors standpoint (so-called "fusion of enhanced and synthetic vision"). Perhaps more importantly, the EVS-based data will also be utilized in machine interfaces, using EVS/database correlation to generate separate-thread navigation, attitude, and hazard signals for verification against conventional navigation (such as GPS/INS) and stored map/terrain data. Although early uses include situation awareness enhancement and integrity monitoring for RNAV/RNP operations, the long-term goal is a cockpit system that economically achieves DVIC operations (ultimately to 0-0 visibility conditions) at non-precision equipped airfields. In the military case, this will include primitive landing areas. The key aspect to achieving regulatory approval of IEVS in these roles will be proof of system integrity, including real-time, automated confidence monitoring; and adequate back-up provisions for situations where such monitoring indicates inadequate integrity. These processing operations are computationally intense. The goal of the present work is to employ certain neural-net-derived technology in order to achieve these capabilities on an economical and compact platform, with clear, transparent confidence metrics. A particular feature of this approach is that it is robust in the presence of degraded image data, including noise and obscurations. We note that, although the terminology has changed, this general approach of "separate-thread", sensor-based integrity assurance has been pursued for more than a decade - including in the context of the Boeing "Enhanced Situational Awareness System (ESAS)" [1]. Also, a challenge facing the IEVS concept is that RNAV/RNP approvals are independently evolving towards Cat I minima (and lower for certain military transport missions), while the requirement for still lower decision heights occurs in only a small percentage of operations. Therefore, any added capability must be highly cost effective. In the present paper, we will describe the conceptual background, early simulation results, implementation plans, and integrated-systems framework of this technology. Ongoing activities include flight testing with a multiple-sensor suite and associated database references.