ATCLAB : a laboratory environment for research in advanced ATC automation conceptual design

Abstract

Introduction: A large number of ideas and schemes have been proposed and are constantly being suggested to enhance the Air Traffic Control system's safety, reliability, and efficiency by means of automation. The capability of the Federal Aviation Administration to properly specify and procure advanced automation systems depends critically on its capability to evaluate these ideas from a number of viewpoints: 1. Functional: i.e., is the proposed idea of any value, assuming it could be implemented? 2. Procedural: i.e., can the proposed idea be implemented in conjunction with existing and/or new ATC procedures? 3. Implementability: i.e., can the proposed functionality and/or procedures be implemented, with sufficient accuracy, reliability, data requirements, etc? 4. Cost/benefit: i.e., are the benefits expected from the proposed functionality sufficient to offset the expected costs and risks? 5. Requirements definition: i.e., is the proposed scheme or system sufficiently well defined to allow the development of meaningful and supportable requirements? There are two conventional approaches available to answer these questions: analysis and simulation. The effectiveness of analysis to evaluate an automation proposal usually depends on the degree to which the proposed function interacts with other elements of the ATC system. In general, the more isolated and self-contained the function, the more amenable it is to analytical evaluation. Functions that interact with many different elements of the system generally require dynamic simulation for effective evaluation. Traditionally, this has required the development of an ad-hoc simulator to evaluate the proposed automation scheme, or the adaptation of an existing simulation. Both approaches are expensive and risk intensive; attempts at building all-inclusive, general purpose simulations are even more expensive and not entirely risk free. As an alternative, a mid-ground solution would be the establishment of a flexible computer-based laboratory environment to perform combined analysis and simulation evaluation on an ad-hoc basis in response to the specific automation scheme being evaluated. To be more effective than traditional analysis and simulation techniques alone, this environment must reduce the cost of building prototype code by two orders of magnitude, both in terms of labor and of calendar time, over traditional environments, such as the ones used to develop existing simulators and prototype systems. Recent developments in computer hardware and software have drastically altered the process of developing software, particularly in the systems simulation area. Symbolic computation and object-oriented languages, along with hardware specialized to execute this type of code, have been shown to produce the two orders of magnitude improvement suggested in the previous paragraph. This report analyzes the feasibility of establishing such a laboratory environment, including identification of the required technology, a possible architecture that would fulfill these requirements, a tentative implementation plan, and two sample pathfinder projects to show how the proposed environment could be used to evaluate two specific advanced automation proposals. To facilitate references to this environment, it will be referred to in this report as ATCLAB; this is not an official FAA-approved name.