In recent years there has been growing concern over the adequacy and distinctiveness of navigation lights in use at sea. This has expressed itself in a series of modifications to the lights shown by various specialized vessels, and by and large the more unusual users of the sea are now clearly and distinctively lighted by night. However, despite the noise being made about it, little has so far been done to improve ordinary ship's navigation lights. There can be little doubt that something urgent does need to be done about them; the survey carried out by Shell International Marine1 indicated this statistically, and any practising navigator need only think of the difficulties encountered at sea from other ships showing wrong or inadequate lights to realize it.

The range of navigation lights in general is the first and most obvious matter for improvement. The present statutory ranges were laid down when oil lamps were still the normal means of providing light, and when in addition ships were much smaller and slower than the average today. In view of the current size and speed of ships, and the obvious fact that in the future they are going to get even bigger and faster, it seems senseless to have their navigation light ranges prescribed on a datum established by oil lamps. The mind boggles at the prospect of two crossing supertankers, each with side-lights visible for no more than the statutory 2 miles, maintaining course and speed until each saw the other's side-lights (a common practice at sea).

This paper covers the performance of the Sperry SGN-10 Inertial Navigation System as demonstrated by the operation of dual systems installed in a standard operational configuration aboard four jet aircraft of Pan American World Airways incidental to a preoperational engineering evaluation programme conducted for the Federal Aviation Administration during the latter part of 1966. The primary object of this evaluation programme was to determine the system's capability of meeting the following requirements of the F.A.A.'s Advisory Circular covering the use of inertial navigation systems in U.S.- registered civil aircraft, namely:

(1) Maintain a position accuracy within 20 n.m. in the across-track dimension and 25 n.m. in the along-track dimension for 95 per cent of the time on flights up to and including ten hours duration.

(2) Automatically accomplish initial platform alignment in a normal manner in latitudes up to and including the highest normally used in civil aviation.

(3) Perform all its designed navigational functions in a normal manner at all latitudes, inclusive of polar and equatorial overflights.

In 1965, following the Eastbourne Conference, the British, French and German Institutes of Navigation formed a Working Group to make a study of the environment in which the supersonic transport will operate and of its implications for the navigation of such aircraft. The Group's initial task has been one of education, largely through discussion of a series of papers submitted to it. Some of the papers considered have already been published in the Journal (Vol. 19) and a further selection is published below. Table I was contributed by Mr. G. E. Beck. The illustrations to these papers have not all been reproduced.

1. Atmospheric Conditions. It will be useful to distinguish between different kinds of atmospheric influences on supersonic aircraft operations. They may be classed as follows:

(a) Sporadic effects near the ground

(b) Sporadic effects in the free atmosphere

(c) Effects on sonic boom

(d) Effects of atmospheric ozone

(e) Permanently effective atmospheric parameters, such as temperature, density and wind.

The word ‘Automation’ when applied to the shipping industry seems to invoke in the individual mind such matters as direct control of propulsion units, planned bridge lay-outs, cargo plan computers, engine data loggers and such other items of ships' equipment. Automation is thus limited in effect and confined entirely to the ship internally. Clearly this ensures that the ship as a unit is as economically efficient as modern techniques allow and is a reflection of the up-to-date attitude of mind of many shipowners.

On the other hand there is a very large field open to automation which concerns sea transport as a whole in respect of the actual conduct of ships at sea both individually and in relation to their effect on each other. It is in this field of sea transport operations that the author sees many longterm possibilities and this paper tries to give some lead into these, which so far seem to have received little, if any, widespread thought or discussion.

It seems possible at this stage to judge which meteorological factors would be of most importance in commercial supersonic transport operations. The argument would be based on the possible data for the Concorde, which, as is well-known, will fly at mach 2·2 and for which the cruising altitude will be between 18 and 20,000 m. The mach 3 flight would have similar characteristics for which the cruising altitude would be 25,000 m. or more.

If the position of a vessel be determined from two or more position lines of known accuracy, it is possible to construct an error zone about the plotted fix, such that there is some given probability (for example 95 per cent) that the true position lies inside this zone. It is convenient for many purposes, such as ease of plotting, and preparation of accuracy diagrams, that this zone should be a circle, since in this case the figure of merit attaching to a fix is defined uniquely in terms of just one parameter, the radius of the circle. The object of this paper is the theoretical study, followed by practical example, of the circular zone round a vessel's plotted position inside which it may be asserted, with 95 per cent confidence, that the true position really lies.

The stick charts found in the Marshall Islands, of which some 20 or 30 are extant in museums throughout the world, have for long puzzled historians of navigation. In this paper, which is based on a more complete work, to be published elsewhere, the author examines the authenticity and meaning of these objects and concludes that they were mnemonic devices, of no great antiquity, to facilitate navigation and to train navigators.

An important consideration in the design of systems for human control is the difficulty of the task. An analytic measure of this subjective consideration is denned. The difficulty is a function of the lead factor which is an integrated measure of the amount of lead the human must produce. The theory together with experimental results imply that the performance function minimized by a human controller includes the amount of control as well as the amount of error. The difficulty is constrained by the capability, training and stress level of the operator. Remnant is accounted for by assuming that the operator generates additive noise as a function of the difficulty. The model is used to explain Bernotat's experimental results with prediction display.

This work was performed while the author was a guest professor at the Institut für Flugführung as part of the exchange programme between M.I.T. and the Technische Universität, Berlin.

This sets down in extremely simple terms the principles by which navigation aids may be coupled to autopilots and autohelmsmen. These principles apply to human pilots and human helmsmen for a craft cannot be navigated by positional aids alone.

Control by an aid that provides position only is shown to be unstable and, therefore, heading is introduced. This quickened system will be stable unless the control is unduly fierce. Such a system may not allow a craft to follow a required track exactly and, accordingly, ‘over-burdening’ may be required or, alternatively, course substituted for heading.

The type of navigation aid introduces special problems. Approach to a point source aid such as a radio beam leads to an increasingly acute sensing of deviation which tends to reduce stability. The craft will, therefore, need to be settled on to the beam at a reasonable distance from the transmitter and this need will introduce special problems in ‘capturing’ the beam.

Coupling to navigational aids is a basic technique in air navigation which will affect the choice of future air navigation systems. Coupling has not, however, been introduced into marine navigation. In space navigation coupling is confined to precise terminal navigation which at present is limited to launch and docking.

It is now two years since a working group set up under the direction of the French Institute of Navigation at the instance of the Secretariat d'Etat á la Marine Marchande delivered a report which was the fruit of a thirteen-months study of the various possibilities of bridge-automation for commercial shipping.

The terms of reference of this study embraced a ship all of whose essential functions were under the control of a central computer, the system, briefly including:

A dead reckoning loop (ship computer dialogue);

A heading speed conversion loop (officer computer dialogue);

An observation calculation loop (officer computer dialogue);

An anti-collision control loop (radar computer dialogue).

Naturally since this was a formal study, the most complete and complex case was examined, and the approximate cost of the data-handling equipment for such a ship was estimated to be of the order of two million francs.

1. Preliminary consideration. In attempting to assess the worth of a navigation system already known, over against the demands of the supersonic transport aircraft, we shall take care above all to make no presumptions whatever regarding the future, and not to count on the developments of the transport aircraft as regards the favour it will find with its users and consequently with the airlines likely to put it into operation. The question is not put here as to whether the 3 to 1 reduction in the actual flying time will appear as a profit paying for the numerous other constraints which it will impose, at once on the constructors, on the operators and finally on the users, without counting the requirements regarding ground equipment and administrative agreements; the question is merely to know whether, given the normal flight and speed conditions of an aircraft, we have the means to navigate it in an accurate, safe and economical manner.

It has recently been suggested that 24-hour satellites might be used as navigational aids. To what category of position determination aids should these be assigned ? Is a satellite of this kind as it were a landmark, because, at least in theory, it remains fixed over the same point on the Earth's surface, in which case it should be classified under land-based navigation aids ? Is it a celestial body, although only one tenth as far from the Earth as the Moon ? If so, it is an astronomical navigation aid. Or is it a radio aid ? After all, its use for position determination depends on radio waves. In this paper I shall favour this last view. For automation is most feasible when an object of observation can be manipulated. This is easiest with radio aids, but it is, of course, impossible with natural stars.

At present artificial satellites have the advantage over all other radio aids of world-wide coverage.

Professor Eva G. R. Taylor, an Honorary Member of the Institute since 1954, died in Wokingham on 5 July 1966, aged 85. She was a Fellow of Birkbeck College and Victoria Medallist of the Royal Geographical Society. From 1930 to 1934 she was Professor and Head of the Department of Geography in the University of London, and in 1944 became Professor Emeritus.

Over the past two decades there has been a steady increase in the amount of electronic equipment carried by transport aircraft. This reflects the expenditure of a large amount of research and development effort, particularly in the automatic flight control field. Yet despite these developments complete automation has not been applied to air transport in the sense that it has been applied to some industrial processes. The attempts made to introduce partial automation have not always had complete success either operationally or economically.

The reasons for this somewhat piece-meal growth of automation in civil transport aircraft are varied. The most important seems to be that any new aircraft represents an evolutionary step forward. Much of the equipment must either be carried forward from a previous generation or based on it, for a mixture of operational and engineering reasons. Inter-fleet standardization in the interests of maintenance and the need to achieve safety by building on past experience are contributory factors.

As a general rule the navigational function is aimed at determining the position of the aircraft in order to resolve three types of problem:

(1) To subject the aircraft's flight path to an optimum trajectory calculated before departure or progressively adapted in course of flight to the circumstances encountered.

(2) To choose at each point of the selected flight path the flight system best adapted to the safety and economy of the flight.

(3) Taking into account the presence of other aircraft in the airspace, to know and make known the actual position and the information allowing provision to be made for future positions, so as to permit effective air traffic control.

Departures of the actual from the chosen flight path penalize the flight by a lowering of economy (in flying time or fuel consumption). It does not seem, however, that the problems raised from this point of view by S.S.T. are by nature or in difficulty any different from those which affect conventional aircraft. Taking into account the present-day precision of navigational aids there is every reason to believe that departures of the actual flight path from the optimum flight path will introduce a penalization which it is possible to ignore when compared with the penalization due to the inaccuracy of the knowledge of the elements (winds, temperatures, pressures) which have, in fact, served to determine this optimum flight path.

The S.S.T. simulation programme of the F.A.A., already in progress since May 1963, aims at determining system requirements both ground and air in the period 1970–75 and, as a future project, in the era 1975–80. In this context, air operations involving the following types of aircraft are under study:

(a) Current and future military aircraft with supersonic speed capabilities;

(b) Concorde-type operations;

(c) Mach 3 or mach 2.7 civil transport operations.

Since the start of the S.S.T. simulation programme, six studies have been undertaken, each of an average duration of 6 months.

From May 1964, a S.S.T. flight deck simulator (modified DC-8 simulator), located at the N.A.S.A. Research Center, was linked with the A.T.C. simulator at N.A.F.E.C., for communication of positional data (digital, including S.S.R.) and voice information, and thus formed part of the traffic sample under study, i.e. one aircraft among 108 others that can simultaneously be simulated at N.A.F.E.C.

Whilst previous exercises concentrated on terminal area (T.M.A.) requirements, the exercises in progress in July 1966 were designed to examine extended T.M.A. and en-route operations in the early 1970s, with Concorde and mach 3 or mach 2.7-type S.S.T.s flying in or through a mixed environment of subsonic traffic. Amongst other parameters, a variety of geographical zones, air traffic control centres, T.M.A. control units, route network configurations in the New York, Boston, Cleveland and California areas, were simulated with a view to determining the impact of specific regional terrain and traffic conditions.

I recently received from a correspondent on the other side of the Atlantic a request for information concerning the graduation of compass cards. I compiled the following notes which may be of some interest to readers of the Journal. I must, however, point out that to a great extent they are based on examination of compasses and cards in museums and on my own experience. It may well be that readers, especially those from other countries, may be able to supplement or correct my statements. Any such information would be most welcome.

A great deal has already been written about the navigational problems presented by the appearance in the near future of supersonic civil aircraft; in this paper I shall try to give a quick description of the methods that will be available in the supersonic Concorde in 1970, with an account of the chief considerations that governed their choice. While there is no doubt that there is a specific navigational problem proper to supersonic aircraft, it does not present itself in as critical a form as some writers believe. When the present-day jets came into service there was not, strictly speaking, any real gap to be bridged. The conventional methods and means used in prop-driven aircraft were taken over, with whatever further adaptations were required for particular cases—such, for example, as polar navigation. In spite of the increase in speed, there was still sufficient time for the navigator to check position and heading at regular intervals.

However, the 300 per cent increase in speeds now envisaged (1300 knots = mach 2.2 for the Concorde) introduces a more formidable discontinuity in relation to the standards of today. At a speed of 22 miles a minute it becomes impossible for the navigator to work as rapidly as he must if he is to deal with the information provided by his sensors and carry out the necessary corrections.

Both civil and military aircraft are increasing in size and complexity and the range of height and speed over which they operate is extending. At the same time, cost of training aircrew is exerting pressure on managements to reduce crew numbers, usually by eliminating the need for specialists. These more expensive aircraft with smaller crews justify more elaborate computing and control systems to exploit their military capability or commercial earning power. Therefore, the need for the more extensive use of airborne computers is already well established.

Due to high cost and poor reliability, analogue computing systems have only marginally satisfied the demands for increased automation in aircraft. However, recent advances in digital computer technology offer advantages in size, accuracy, flexibility and particularly in reliability over analogue methods and show promise that digital computer systems can be designed to meet the future requirements for airborne computing.