It was not long after the first satellite launches took place in 1957 and 1958 that the United Nations established an Ad Hoc Committee to consider the need for international agreements relating to the use of space. In June 1959 the Ad Hoc Committee approved the Report of its Legal Committee, which included a list of topics in relation to which it was considered that the negotiation of international agreements under the auspices of the U.N. would be valuable (1). This list was divided into priority subjects, many of which have since been dealt with in international treaties, and less urgent subjects, many of which are still the subject of discussion (2). The last topic on this list was "rules for the avoidance of interference among space vehicles", a subject on which detailed agreement remains to be achieved.

"Interference among space vehicles" can itself be usefully divided into at least three categories: first, deliberate interference in any form, such as radiation, projectiles, or mechanical impacts; second, interference in the form of impacts with debris; and third, interference in the form of collisions between space vehicles. Although it is clear that these categories are not mutually exclusive, the distinctions can be helpful.

The first form of interference is already forbidden under existing Space Law (3), and although there will no doubt be a need for more detailed formulations in future, it is not proposed to discuss this matter further in the present paper. The second form of interference, that of impacts with orbital debris, has become the subject of growing interest in recent years. Scientific studies are under way (4, 5), and a range of measures have been proposed for limiting and reducing these problems to an acceptable level (6, 7).

The third category, interference in the form of collisions between space vehicles, has received relatively little public consideration to date. This has been due primarily to the lack of any need: the total number of space vehicles that have been launched, and the maximum number that have been in earth orbit simultaneously, have not been sufficiently large to make the risks of collision unacceptable. Whole spacecraft currently pose much less of a problem than does debris, since they are large enough to detect reliably with existing means, and are much less numerous than pieces of debris. At present many countries monitor the orbits of existing satellites, and the USA maintains a detailed catalogue of more than 4,000 low earth orbit ( LEO) objects. The planned orbits of new spacecraft are compared with this catalogue to ensure that collisions will be avoided. However, as the number of space vehicles increases, the calculation of safe trajectories will become increasingly difficult, as will the avoidance of collisions in orbit.

There have already been cases of relatively "near misses" between spacecraft. such as the occasion in July 1982 when the US space shuttle Columbia passed within 15 km of the upper stage of Intercosmos 4. In addition, if the stated plans of the Soviet Union, European Space Agency, Japan and China to develop reusable shuttle vehicles come to fruition, there may be a substantial increase in the rate of spacecraft launches by the end of the century. Nevertheless, none of these plans in themselves seem likely to bring about a sufficiently large increase in the frequency of launches to cause a qualitative change in the problem of avoiding collisions between operational spacecraft.

However, in the event that a rapid increase were to occur in the volume of orbital traffic, and in the corresponding risk of collisions between spacecraft, it would clearly be in the interests of all nations to take appropriate steps to reduce these risks as far as possible. In this context it would appear that there are at least two non-military scenarios which do have the potential to generate sufficiently rapid growth in space activities to justify devoting more attention to the matter of avoidance of collisions between space vehicles:

The first of these scenarios is that involved in the construction and operation of satellite power stations (SPSs) for power supply to earth (9, 10). Several of the legal problems that would arise with such a development have been considered at previous IISL Symposia and elsewhere (11, 12, 13). As concerns the particular matter of avoidance of collisions, however, it is relevant to note that the construction of even a few thousand Megawatts of SPS capacity per year could involve up to 10 return flights per day of reusable launch vehicles. In addition there would be a growing number of very large space structures in earth orbit, of which the economic value would be so high as probably to require the establishment, with international agreement, of "exclusive zones", if the risks are to be commercially acceptable. The avoidance of collisions between spacecraft involved in SPS operations and other spacecraft could clearly be facilitated by appropriate international agreements, as discussed further below.

A second scenario, which has received serious attention only recently (14, 15), but which would appear to have the potential to generate an even higher launch rate than the SPS scenario, is the development of space tourism (16). A major point of interest in the recent work is the suggestion that the level of demand for short visits to a tourist facility in LEO might be high enough to generate, on a self-financing basis, some 50-100 return flights per day, to a similar number of orbiting facilities (16). As in the case of the SPS scenario, the efficient utilisation of space for this purpose by all nations would be facilitated if international agreement was reached on the use of means for reducing the probability of collisions between spacecraft involved in such operations.

These two scenarios each raise their own characteristic problems in terms of increased possibilities for collisions between spacecraft. In view of the technical feasibility of either scenario in the circa-2000 time frame, and of their economic feasibility under certain assumptions, it would therefore appear not to be too early today to give the question of collision avoidance measures at least some preliminary attention.

Under the 1972 Convention on International Liability for Damage Caused by Space Objects (8), consideration was given to the matter of liability for damage caused by space vehicles to other space vehicles. While a launching State is absolutely liable for any damage caused by its spacecraft on the surface of the earth or to aircraft, it is liable for damage caused to other spacecraft in orbit only where fault is proven on the part of the launching State, or of persons for whom it is responsible.

While the Convention also makes provisions for the convening of an international Claims Commission to determine liability and compensation in cases where this is necessary, it is clear that the deliberations of such a Commission will be complex, and would be facilitated by the existence of guidelines for their work. Furthermore, prevention of collisions between spacecraft would in general be far preferable to the determining of liability and compensation after the event. Thus it is desirable to investigate the possibility of establishing "orbital traffic rules", in the form of rules of orbital behaviour, and rights of orbital traffic. The complexity of this subject is such that extensive international negotiations will be required for their resolution.

In the following sections a number of matters relevant to this subject are discussed. First, some of the difficulties that arise with the attempt to formulate legal definitions of spacecraft orbits are discussed, as this question is central to the determination of fault in the event of collisions between operating spacecraft. In the following section consideration will be given to some general physical features of earth orbits that are relevant to the tasks that satellites perform, and therefore to possible measures for reducing the risks of collisions. Some theoretical systems for maintaining separation between spacecraft are then described and discussed in relation to existing arrangements. Next, some possible means are considered by which the probability of collisions between operating spacecraft could be reduced in practice, particularly in relation to the two scenarios referred to above. Finally some conclusions are drawn concerning possible directions in which existing practices might evolve through international agreements to ensure safe and economical utilisation of near-earth space by all nations.

If one attempts to envisage a far-future "traffic system" for near-earth space by analogy with existing international systems of road, sea and air traffic an obvious difficulty presents itself immediately; namely that the orbits of spacecraft orbiting passively do not remain fixed with respect to the earth. Nor do they remain fixed with respect to each other, nor with respect to any system of coordinates. Spacecraft orbits in near-earth space evolve through time due to perturbations introduced by the earth's oblateness, by the influence of other celestial bodies, by the earth's atmosphere, and even by solar radiation. Most importantly, all low satellite orbits decay as a result of atmospheric drag. Thus a spacecraft in a given orbit does not, except in special cases, remain in "the same" orbit, with respect to the earth, nor with respect to other spacecraft.

It is not necessary to consider these different effects in detail, except to note two features:

1. First, the rate of change of spacecraft orbits caused by different phenomena is dependent on several factors such as the height and inclination of the orbit, but may involve rotation of the orbital plane about the earth's axis as rapid as 8 degrees, which is clearly a substantial figure.

2. Second, although some of the influences are accurately predictable even years ahead (at least in principle), others (including in particular the effect of atmospheric drag) are not. Consequently there is a serious limitation on the accuracy with which the future path of a spacecraft, particularly in an orbit with a low perigee, can be known in advance.

These features in turn have certain consequences in relation to the possibility of defining certain orbital "highways". First, while it is possible at any time to define the orbit of a spacecraft, it is not possible to define its future path with complete accuracy, and it would not therefore seem desirable, for instance, to envisage spacecraft being given a "right of way" to pursue a particular passive orbital path. Second, where it is considered desirable to define a particular orbit or orbital zone within which spacecraft could be given rights of occupancy, this will impose certain economic costs on occupiers through the need to actively maintain their orbital positions.

The facts concerning the evolution of passive spacecraft orbits also raise problems in relation to the formulation of rules for determining responsibility for satellite collisions, and hence of liability for damage caused, as can be illustrated by means of some simple examples. In the absence of a system of strict rules, it would be desirable to establish at least some guidelines to aid the process of determining liability in accordance with the principles of justice and equity Two simple cases can be proposed in which liability would seem likely to be agreed without serious controversy:

• A launch vehicle, A, while still thrusting, collides with an operational satellite, B, which has been orbiting passively (i.e. without exerting thrust) for several years.

• An operational satellite, A, fires its thrusters and makes a substantial change in its orbit. Shortly after ceasing to thrust, it collides with a satellite, B, which has been orbiting passively for several years.

Drawing on these two examples, it is tempting to suggest two principles which might seem to provide some guidance in this matter:

1. First, it might be suggested that, in general, space vehicles that are thrusting are responsible for any collision with a vehicle that is not thrusting (i.e. that is orbiting passively). This would be analogous to the informal rule of small boat sailing that "power gives way to sail". The possible justification for such a rule is based on the supposition that a vehicle that is actually thrusting has the greater capability to make evasive manoeuvres.

2. Second, following on from this, it might be suggested that, in general, of two space vehicles which had collided, that which exerted thrust to change its orbit more recently should be held responsible.

Although such guidelines can serve some purpose, they are certainly not definitive, Since it is easy to envisage cases in which they would not produce an uncontroversial decision.

In connection with 1) there are, for example, difficulties in defining what should be treated as "thrusting", since there is a spectrum of cases between what is clearly thrusting, and what is not:

Although usually of only very low power, station-keeping thrusters alter a spacecraft's orbit. Attitude control thrusters may also do so under certain circumstances. Even simply altering a spacecraft's attitude may significantly alter the atmospheric drag which it experiences, thereby altering its orbit. In the same way, therefore, the use of control moment gyros, for instance, while clearly not directly exerting "thrust", can alter a spacecraft's orbit. Changes in the orientation of a spacecraft (or of parts of it) can also alter the net external force exerted by solar radiation, and hence the spacecraft's orbit. Clearly, therefore, it would be necessary to define what should be considered as constituting "exerting thrust" to change orbits, if this principle were to be implemented.

In this connection there are also problems concerning solar sail spacecraft which are, by their nature, thrusting almost continually. To hold such spacecraft responsible for any collision would not allow for the fact that, by comparison with rocket thrusters, they are likely to be seriously underpowered, due to the very small forces involved (17). In addition it seems likely that, in future, tethers will be widely used to alter spacecraft orbits (l8), and decisions would therefore have to be made on how such manoeuvres should be treated under law. A further problem that would arise with the use of very large structures (such as those employing tethers of tens or even hundreds of kilometres in length), is the definition of the orbit which such a spacecraft is using, since in physical terms only the centre of mass of such a craft is orbiting passively.

Following on from these considerations, there are additional problems in connection with (2), as seen from the following example:

Two operational satellites, A and B, having been orbiting passively for several years, collide. On investigation it is found that thrust was last exerted by satellite A several years ago, post-dating the last thrust exerted by satellite B by some days.

It is thus not clear that extension over long periods of time of the principle that the satellite exerting thrust more recently is responsible for any collision, would be just. An obvious corollary difficulty is that of verification, particularly in relation to low-thrust manoeuvres.

As a result of the complexities of the issues involved, it seems unlikely that any simple set of principles such as those considered above, would prove satisfactory. It would also seem likely that whatever guidelines are adopted, tflere will be cases in which it will not prove possible to allocate responsibility for collisions between earth-orbiting spacecraft either to one vehicle or to the other (in which case liability is shared equally under the exiscing Convention on Liability).

Although it will be important to establish fuller guidelines as to what should be considered to constitute "fault" in respect of collisions between spacecraft, we do not consider this question further here. Instead we will consider the matter of prevention of collisions. First we examine some of the characteristics of earth orbits that influence the choice of orbits for particular satellites, followed by some possible approaches to minimising the possibility of collisions between space vehicles.

The geophysical characteristics of near-earth space determine the uses to which the various possible orbits around the earth are put. Surveying these possibilities very briefly we see that, in general terms, different spacecraft missions lead to different preferred orbits:

Earth observation satellites of different types are driven by conflicting orbital requirements. The closer that a satellite orbits to the earth's surface, the better the resolution that it can achieve in distinguishing features of the terrain, atmospheric phenomena, or other targets of interest. However, at the same time, the lower the spacecraft's orbit the greater the atmospheric drag that it will experience, and therefore the shorter its operational lifetime (except at the cost of providing greater orbit-raising capability). In addition, the greater the inclination of the orbit of an earth-observation satellite, the greater the proportion of the earth's surface that it will be able to cover (but the greater the propellant requirements to achieve orbit). A particularly attractive case is the family of sun-synchronous orbits, near-polar orbits which remain fixed in relation to the sun. Thus low, near-polar orbits will generally be attractive for earth-observing spacecraft, the higher orbits and inclinations being preferred for longer-life satellites requiring more complete global coverage, and lower orbits and inclinations being used for shorter-lifetime missions with more specific targets of interest.

Manned space flight will tend to utilise low orbits for economy in launch costs, but somewhat higher orbits where longer missions, such as permanent stations, are required. To the extent that their objectives are the development of manned space flight expertise and experimentation in low-gravity environments rather than, for instance, earth observation, the ground tracks of manned spacecraft are not as important as they are for earth observation satellites. They will therefore tend to utilise more economical orbits with inclinations no greater than the latitude of their launch sites.

Telecommunication satellites require a constant spatial configuration between the ground stations and the spacecraft for maximum efficiency. This is most fully achieved by using the stationary orbit which, in the earth's case, is at an equatorial altitude of some 35,800 kilometres. Although the coverage of the earth's surface from a point in geostationary orbit is approximately 40%, this orbit is not very convenient for communication with sites at high latitudes, and so for their case other orbits will be of more use. These include nonstationary geosynchronous orbits, and highly eccentric 12-hour and 24-hour orbits with an inclination of some 63 degrees - 'Molniya' orbits, in which the apsidal rotation is zero and the orbit's apogee remains over the same point on earth (given station-keeping to counteract solar and lunar influences).

Navigation satellites have some of the same requirements as telecommunication satellites, and the geo stationary orbit has considerable value for such pu:rposes, as shown by the commercial interest in geostationary position location satellites (19). They are not ideal for non-interactive navigation systems, however, and lower orbits are more economical in launch costs. Hence intermediate orbits will be attractive for "constellations" of navigation satellites, such as the Global Positioning System (GPS). Once again, the preferred inclinations will be determined by the requirements for ground coverage.

Scientific satellites have individual and varied requirements. Certain types of astronomical satellite will have very high apogees beyond the earth's magnetic field, while having low perigees for launch economy and other reasons. Other scientific satellites may use sun-synchronous orbits, or trajectories to escape from the earth's gravitational influence.

An important feature of near-earth space, which has not been mentioned so far, is the existence of large radiation belts trapped by the earth's magnetic field. These are most concentrated at altitudes between approximately 1,500 km and 4,000 km, and between 10,000 km and 30,000 km, at low inclinations to the equator. Consequently, at altitudes at which the radiation levels are significantly damaging for satellite systems such as solar arrays, orbits with higher inclinations will generally be preferred for uses where choice of inclination is essentially free. Conversely, the region of low-inclination orbits at the peak radiation altitudes will be attractive zones for disposing of "dead" satellites and possibly other waste materials.

Drawing on the discussion in the previous section, it is possible to envisage a system whereby, in order to minimise the possibilities of collisions between spacecraft, certain regions of near-earth space would in future be reserved primarily for space vehicles serving particular purposes. For instance, a number of non-intersecting "shells" might be defined as follows:

• An inner zone comprising all orbits with apogees of less than, say 300 km might be reserved primarily for unmanned earth-observation and monitoring satellites.

• An intermediate shell comprising all orbits with perigees of more than 300 km and apogees of less than 500 km might be reserved for manned vehicles and facilities, and for man-tended scientific satellites.

• A third zone might be defined to be reserved for 'Molniya' orbits (i.e. with 12 and 24 hour periods and with inclinations of some 63 degrees) with perigees of more than 500 km, so as not to intersect the lower zone.

• A fourth zone might be defined to include the most intense radiation zone of the inner radiation belt, covering orbits with perigees of more than, say, 2,000 km and apogees less than, say, 3,000 km and with inclination of less than, say, 100. This zone might be reserved for disposal of dead satellites and other waste.

These four zones are non-intersecting, but it would also be possible to define certain other zones which would intersect with these zones only in relatively limited circumstances. Thus it would be possible, for instance, to define a zone to be reserved primarily for high-inclination earth observation satellites between, say, 600 km and 1,000 km in altitude, and between 800 and 1100 in inclination. Of the zones mentioned above, this would intersect only with the 'Molniya' orbit zone, and only within a limited region of space. Further zones could also be defined to cover, for instance, navigational satellite "constellations" at particular altitudes and inclinations.

Having seen that it is possible to subdivide the volume of near-earth space into zones with either limited or zero over-lapping regions, it is worthwhile to consider in a little detail what rules might be chosen to apply to these zones; that is, what operational significance might be attached to the concept of "reserving" these zones for particular purposes:

First, it can be envisaged that where a zone was reserved for certain categories of activity, users of the zone not engaged in these activities might, for instance, be held liable for collisions with space vehicles involved in these activities.

Second, it can also be envisaged that within zones to be reserved primarily for certain purposes, particular orbits might be reserved for particular users. A single orbital zone could, for instance, be subdivided into a considerable number of non-intersecting orbits, or "sub-zones". It is notable that a single circular orbit can contain a large number of space vehicles with constant separation between them, and with minimal danger of collisions. (The separation between spacecraft in the same elliptical orbit would not be constant, but would oscillate between a maximum value near the perigee and a minimum value near the apogee). The siting of many satellites in a single orbit is therefore a very economical configuration which greatly reduces the probability of collisions between spacecraft, at the cost of imposing certain station-keeping requirements on them. However, if it is to lead to efficient use of space, the definition of reserved LEO orbits will have to take into account the function for which the orbit is to be used, as discussed further below.

Third, while it might become accepted practice to reserve certain zones or orbits for particular uses, it seems likely that it will continue to be necessary to permit launching vehicles to pass through reserved zones. It will clearly be necessary to minimise the opportunities for collisions in these circumstances. As the volume of orbital traffic increases, the calculation of safe launch trajectories may come to be performed not solely on a national or bilateral basis as at present, but under the auspices of an international organisation.

An important feature of near-earth space, which has not been mentioned so far, is the existence of large radiation belts trapped by the earth's magnetic field. These are most concentrated at altitudes between approximately 1,500 km and 4,000 km, and between 10,000 km and 30,000 km, at low inclinations to the equator. Consequently, at altitudes at which the radiation levels are significantly damaging for satellite systems such as solar arrays, orbits with higher inclinations will generally be preferred for uses where choice of inclination is essentially free. Conversely, the region of low-inclination orbits at the peak radiation altitudes will be attractive zones for disposing of "dead" satellites and possibly other waste materials.

A particularly interesting precedent in this context is the determination in Article 33 of the International Telecommunication Convention (I.T.C.) that the geostationary orbit is a "limited natural resource (which) must be used efficiently and economically so that countries or groups of countries may have equitable access ..."

This statement is useful for a number of reasons; first, because the analogy between geostationary orbit and the orbital "zones" described above is very close. Clearly the geostationary orbit comprises a much more limited volume of space, with more specific geophysical properties, and it is furthermore already relatively congested with operational satellites. It is therefore natural that its use should already be the subject of international agreements and conventions. However, as described above, the properties of other regions of near-earth space are also distinct as far as their potential economic value is concerned, and their value will also potentially be seriously reduced if their exploitation is not carried out efficiently and equitably. The need for international conventions to be formulated in relation to such zones will clearly be a function of the rate of traffic build-up within them in coming years.

A second reason why the regime covering the geostationary orbit is a significant precedent in the present context is that, by reason of the definition of the word "telecommunication" in Annexe II to the I.T.C., all operational satellites are telecommunication systems, and as such are governed by the provisions of the I.T.C. and the Radio Regulations annexed thereto. This point was usefully made in relation to satellite power stations by Dr Jan Busak13, and its implications in connection with the potential for extending the systematisation of orbital traffic are considerable. In particular it implies that, as the volume of traffic in certain zones of near-earth space increases to the level at which it is perceived that there could be benefits from introducing some additional degree of regulation, the International Telecommunications Union (I.T.U.), by reason of its existing mandate, could readily extend its regulation of geostationary orbit to cover satellites in these zones. This is not to say that the Member States of the United Nations might not decide at some date to establish a new institution to carry out this responsibility, but it is notable that no such institutional innovation would in principle be necessary in order to ensure an orderly extension of the form of regime currently applying to the use of the geo-stationary orbit to other regions of near-earth space.

A third significant point concerning the precedent of the I.T.U.'s responsibility for ensuring that the exploitation of the geostationary orbit is internationally acceptable is that, notwithstanding the fact that the primary task of the I.T.U. (namely the regulation of the orderly exploitation of the electromagnetic spectrum) is in itself unrelated to the task of controlling the exploitation of a geographical region of near-earth space, its achievements in this latter task have been very satisfactory to date. That is, despite periodic controversies at successive World Administrative Radio Conferences, the influence of the I.T.U. in this area has unquestionably been beneficial to humanity as a whole, by comparison with the disorder and controversies that might otherwise have arisen.

Thus, notwithstanding the fact that the primary function of satellites in most orbits other than the geostationary orbit is not telecommunication as such, the prospect that, in the absence of further international agreements to create a new institutional mechanism for the purpose, the I.T.U. may become involved progressively in regulatory activities relating to the exploitation of successively more regions of near-earth space, is not an unattractive one. The fora of such primarily technical bodies as the I.T.U., at which technical experts from all different countries meet in order to determine as far as possible the technically optimal solutions to particular resource allocation problems, are probably as satisfactory a means of international decision-making as have yet been established by humans.

Nevertheless effective and appropriate traffic regimes for non-geostationary orbital zones would be very different in nature from that applying to the single orbit of geostationary space, and a range of new approaches will be required. In particular, the definition of orbits will take a different form: whereas sites in GEO are defined by their geographical location with respect to the earth, the definition of LEO orbital sites will be more complex. Specifically, a particular low earth orbit will have to be defined according to its inclination; according to the altitude range of its perigee and apogee; according to the right ascension of the ascending node at a specific time and date; according to the rate of nodal regression (i.e. rotation of the orbit about the earth's axis); and according to the rate of apsidal rotation (i.e. rotation of the major axis within the plane of the orbit). Since these parameters will vary to some extent according to the size and shape of the spacecraft using the orbit, it will be necessary for the values specified for a particular low earth orbit to be typical of the spacecraft intended to use the orbit, in order to minimise their expenditure of station-keeping fuel.

In addition to the specification of orbital parameters in this way, the matter of establishing guidelines for determining liability for collisions between space vehicles will need to be tackled. In this context, the International Civil Aviation Organisation, (I.C.A.O.), which has long experience of resolving complex technical issues relating to collision avoidance between aircraft, may be able to make a useful contribution in this field.

In view of the particular nature of the task in question, and of its complexity, it may be that, rather than attempting to establish comprehensive agreements, the implementation of a few, specific, limited agreements may be preferable, at least in the early stages. A number of such limited steps that might be usefull in connection with the two scenarios described in the Introduction are discussed in the following section.

Finally it must be emphasised that any extension of the application of international regulations to other regions of near-earth space will be acceptable only to the extent to which it can be justified by some appropriate form of cost-benefit analysis. That is, any costs to which such increased regulation gives rise must be outweighed by measurable benefits to the users and future users of orbital space, notably in reducing the calculated probabilities of collisions occurring between space vehicles. Thus the formal extension of, for instance, the I.T.U. 's role to cover space vehicle utilisation of near-earth orbital regions will presumably be acceptable only at a time when, and to the extent that, such risks and the potential for reducing them are of such a magnitude that the costs (including both additional administrative costs incurred by governments, and costs incurred by space vehicle operators as a result of regulations imposed, such as the requirement to carry additional manoeuvring propellants), are justified by the demonstrable savings in terms of improved safety and efficiency of spacecraft operations.

On consideration of the preceding section, it is clear that the implementation of such systems, whether through the extension of the scope of the I.T.U.'s activities or otherwise, would be premature in the near future. The average density of the population of spacecraft in low earth orbits is still very low, and with a continuation of current rates of satellite launch and orbital decay it is set to remain low for some years to come. However, the distribution of satellites is not homogeneous, for a large majority of existing satellites (and of orbital debris) orbit within one or other of a small number of orbital zones (20). Nevertheless it is not proposed to consider the possibilities for the hypothetical, general system of "traffic zones" discussed in the previous section in greater detail here. Instead, it is proposed in the following to consider further the two more concrete scenarios referred to in the Introduction, either of which would involve both a rapid rate of launches of manned and unmanned space vehicles, and much higher numbers of larger spacecraft in orbit. These scenarios are:

1. that comprising the production and operation of satellite solar power systems (SPSs) (10); and

2. that comprising the development of space tourism as a major industry (l6).

Satellite Power Station Scenario

As a result of work performed in the USA and elsewhere during the 1970s, it was established that, provided that the costs of photovoltaic cells and of space transportation fell substantially, and subject to international agreement under the auspices of the I.T.U. (among other bodies), it would probably be feasible to supply large quantities of electrical power to earth by collecting solar energy in earth orbit and transmitting it to the ground as microwaves (21, 22). In order to provide substantial quantities of power in this way (i.e. several hundred Gigawatts), it would be necessary to construct some thousands of square kilometres of solar arrays in orbit, and for tens of satellites of approximately 100 square km to operate in geostationary orbit for lifetimes of more than twenty years.

Without going into detail, it is clear that such an undertaking would involve the establishment of an extensive industrial infra-structure in near-earth space (as well as on Earth). It seems likely that, in the event that SPS-generated electric power became economically competitive, the rate of construction of SPS capacity would grow to a rate of 20 GW per year or more, involving perhaps four different assembly sites in orbit. These would comprise many square kilometres of assembly facilities and part-constructed segments of solar power satellites, and would receive several return visits per day from cargo and personnel launch vehicles. The implications of such a project for collisions between space vehicles are considerable:

First, due to their very large dimensions, the volume of near-earth space that would be "swept" by such facilities would be approximately one million times greater than the volume swept by existing satellites, and the probability of collisions would therefore be proportionately higher.

Second, SPS assembly facilities would be likely to be a substantial source of debris, unless strict regulations were enforced.

Third, the sites occupied by SPS units would be occupied more or less permanently. That is, although any particular piece of hardware might have a limited lifetime, maintenance would be provided to ensure that the system's functions continued for as long as they were economically justified.

Fourth, the rate of launches to SPS construction bases would be many times the total rate of launches today worldwide.

Arising from these characteristics of the SPS project, it is possible to envisage a number of directions in which the process of formulation of international agreements might develop in order to minimise problems of collisions between spacecraft arising from the project. These can be treated to some extent in three categories, relating to a) geo-stationary orbit ( GEO), b) low earth orbit ( LEO), and c) launch trajectories and transfer orbits between LEO and GEO sites.

Geostationary SPS Facilities

In a number of respects the geostationary orbit would be convenient for siting operational SPS units, but its utilisation for this purpose would require international agreement on several different matters. These include, first, the problem that the occupation of a "slot" in GEO by an SPS unit would typically be much longer lasting than its occupation by telecommunication satellites which generally have a lifetime of ten years or less. Since it is already established under the 1967 Treaty on Principles (3) that space is not subject to "appropriation", it may be useful to introduce a concept of "usufruct", whereby groups desiring to utilise a site on such a semi-permanent basis may be permitted or licensed to do so. The need for some such innovation would be the greater if it was agreed that the high value of SPS physical assets necessitated the provision of a secure zone around the site, from whi-ch non-licensees could be excluded, by force if necessary.

A second problem that would have to be dealt with is the fact that in delivering microwave power from GEO to earth an SPS would periodically irradiate spacecraft in low inclination orbits below GEO. For low orbits the intensity of such radiation would not be damaging, but for satellites at higher altitudes costs would have to be incurred in order to protect their functioning from interference (23).

Third, it might be necessary to establish guidelines on the accuracy with which SPS units should maintain their stations in GEO in order to limit both the possibility of collisions, and contamination by station-keeping thrusters. The latter would, from the legal point of view, be very similar to damage from collisions.

LEO SPS Facilities

SPS construction facilities would be operated most conveniently in low inclination orbits at approximately 500 km altitude. With cross-sections measured in square kilometres, the probability of collisions involving these structures would be very much higher than for existing satellites. One means of greatly reducing the possibility of collisions would be to site LEO SPS facilities in the same orbit (thereby reducing the total volume of space swept by these facilities). That is, although within the zones of near-earth space that are particularly attractive for manned space operations (such as the assembly of SPS segments), no single orbit is uniquely favoured (as the geostationary orbit is among higher altitude orbits), nevertheless by designating a particular orbit for the siting of SPS construction facilities, several advantages would be obtained:

First, the probability of collisions between these very large facilities would be reduced (provided that their station-keeping within the designated orbit was sufficiently accurate).

Second, the relative velocity of any debris produced by a particular SPS facility relative to other facilities would be limited to a low level, thereby greatly reducing the danger of causing damage in this way.

Third, the constraints on other space vehicles, if they were to be encouraged to avoid orbits that intersected with orbits used by SPS construction facilities, would be much less if only one such orbit was to be the subject of special international agreement in this way.

The constraints on other space vehicles arising from the daily launches to LEO SPS construction facilities would be much more circumscribed if their destinations were all within the same orbit, albeit at different positions within it. In the case of LEO- GEO transfer orbits, the destinations are of course all within a single orbit, which provides considerable scope for simplifying the process of placing SPS units in their appropriate geostationary slots.

In the SPS literature it was suggested that SPS segments would be raised from LEO to GEO by means of self-powered electric thrusters. This avoids the need for chemical propulsion, but involves a long spiralling trajectory taking some three months to complete. As in the case of SPS construction sites in LEO, it would be possible to specify that SPS segments in transit to GEO should, as far as possible, use the same trajectory. This would be likely to be the case if the construction facilities were also in the same orbit. The final stage of siting an SPS segment in GEO would be complicated by the need to avoid crossing the microwave beams being transmitted from existing SPS units, which would be very intense at short range.

Thus, even after only a very brief review of the subject, and without going into great detail, it is possible to envisage means whereby the introduction of a limited number of regulations on the use of near-earth orbits could greatly reduce the risks of collisions between space vehicles arising from the implementation of the SPS project.

As outlined in the Introduction, the space tourism scenario to be discussed also involves a major increase in traffic in near-earth space. However, it is very different from the SPS scenario in its implications, in particular in that it would involve a much greater increase in the number of space vehicles in orbit, and of launches. Under the scenario described by Collins and Ashford (16), in which one million passengers per year would make short stays in orbiting hotel facilities, there might be as many as 100 flights per day to as many orbiting facilities. While not even approaching the level of activity of a busy airport today, this would nevertheless represent an increase by more than two orders of magnitude over current launch rates. In addition, although much smaller than SPS segments, the orbital hotel facilities would themselves have dimensions of the order of 100m, thus representing an increase in cross-sectional area of some one hundred times over existing satellites.

As is the case of manned facilities generally, there would be advantages in siting hotel facilities in low inclination orbits of 300-500 km altitude in order to minimise launch costs. The potential for collisions between some 100 large facilities of this type, and of a similar number of launch vehicles orbiting within a relatively small zone of near-earth space would therefore be large. The risks of collisions could therefore be greatly reduced by siting such facilities either in a single orbit, or in two or more non-intersecting orbits. The several advantages to be gained from such an arrangement would be the same as those described above in relation to LEO SPS facilities. The introduction of a single simple arrangement such as this would greatly limit the additional risks of collisions that would otherwise arise with such a development. The implementation of a common policy of this form would however represent a significant new departure in the utilisation of space, and would require extensive negotiations before it would be internationally acceptable.

It is important to note that there would be mutual benefit in such an arrangement for both the international space-using community as a whole, and for the owners and operators of the hotel facilities. In consequence it is possible to envisage international agreements being reached to allocate rights of occupation of certain orbital slots in LEO to particular users, in exchange for their agreement to maintain their positions in those slots to within a certain degree of accuracy. Once a particular orbit became designated for particular uses in this way, its use by other space vehicles would probably be forbidden except under specified conditions.

In view of the daily flights that would be likely to take place to and from each hotel facility (in addition to extensive inter-hotel traffic within the reserved orbit), it also seems probable that rules would come to be formulated, as the numbers of such factlities in orbit grew, to cover movements of launch vehicles. It is possible to envisage that some such local "orbital traffic control" might be operated initially by the operators of the hotels, ant that it might subsequently evolve to become part of a wider system of "aerospace traffic control". Since extensive passenger traffic would not arise unless the risk of collision was at least as low as it is today for commercial aircraft, there would clearly be a considerable incentive for such a development.

In the preceding sections we have considered the possible need to determine rules or guidelines for establishing liability in the event of collisions between orbiting space vehicles, and the difficulties in achieving this due to the complexities of the evolution of spacecraft orbits in near-earth space. We have also seen that there are nevertheless a number of ways in which, as the need arose, international agreements could be made that would reduce the possibilities of collisions between space vehicles. Finally we have considered two particular scenarios which have the potential to lead to substantial increases in orbital traffic, and some possible steps that might be taken, in the event of their occurrence, to substantially reduce the risks of collisions. In the following we shall summarise a number of directions in which existing institutional arrangements might evolve so as to achieve the required degrees of safety, economy and justice in the continuing exploitation of near-earth space.

1. At present, States which are signatories of the 1974 Convention on Registration of Objects Launched Into Outer Space24 agree thereby to inform the Secretary General of the United Nations of the details of spacecraft launches for which they are responsible. It is possible to envisage a number of ways in which this service provided by the U.N. of registering the launch parameters and initial orbits of space vehicles could evolve so as to be of greater use to the international community in relation to the problem of avoiding collisions between spacecraft. For example, the responsible U.N. office might extend its activities to include recording up-to-date information concerning the parameters of the current orbits of space vehicles in addition to their initial orbits. This might come about through the office collecting information about orbital parameters of current spacecraft from the various existing sources of such information, (such as NASA Goddard's "Two Line Elements").

   Alternatively it might be that Members would agree to extend the Convention on Registration so that launching States registered not only launches of spacecraft, but also their orbital parameters on a continuing basis. The procedure for this might be that launching States would initially register their best projections of the orbit of the spacecraft to be launched for the following two years, and would formally update this information on an agreed basis (perhaps both annually and whenever the spacecraft's orbit changed by a specified amount). In this way the registration of launches would gradually evolve to encompass both the registration of spacecraft orbits, and registration of changes in their orbits.

2. Signatories of the I.T.C. agree thereby to the adjudication of the appropriate organs of the I.T.U. in matters concerning the allocation of frequency bands of the electromagnetic spectrum to different users, and in matters concerning the allocation of geostationary orbital positions to different satellites. It is possible that States may agree to the I.T.U. 's role being extended, as the need arises, to cover other areas of near-earth space. As discussed above, this might include the establishment of several defined orbital zones, and possibly of specific orbits (in addition to GEO), to be reserved for particular purposes and users. Such a development might come about through international agreements that organisations concerned should be offered the option of siting particular facilities at certain sites within a particular designated orbit in exchange for agreement that, provided that they maintained their position within agreed limits, they would not be liable for any collisions which they suffered. That is, any spacecraft colliding with a registered craft in such designated orbits would be absolutely liable for any damage caused. Agreements of this form should provide sufficient guarantee of the safety of spacecraft for satisfactory commercial operations, even under much higher traffic densities than exist at present.

3. As traffic in LEO increased to the extent that it was felt desirable to introduce some form of organised aerospace traffic control, the I.C.A.O. might also come to play a constructive role. This organisation's experience in the monitoring and control of air traffic world-wide could clearly be of value in helping to determine appropriate safety standards, proximity rules, trajectories for crossing reserved zones, and other procedures for efficient orbital utilisation.

4. It would also seem likely that the international community of insurance companies and underwriters should have a constructive role to play in formulating acceptable rules for minimising the risks of collisions between space vehicles. The recent series of unbalanced losses suffered by companies providing insurance for spacecraft has revealed a clear need for closer collaboration between insurers' technical departments and manufacturers and operators of spacecraft and launch vehicles.

Finally, although it will be desirable to minimise the costs of any extension of regulations to users of near-earth space, it should be borne in mind that, in principle, the imposition of certain obligations (such as, for instance, the requirement to maintain a particular orbital position to within a stated degree of accuracy, necessitating the use of station-keeping propellants; or the requirement to carry a certain quantity of additional propellants for emergency manoeuvring, controlled de-orbiting, or for boosting into a "junkyard" orbit at end of mission), would not be different in principle from the internationally accepted imposition on users of motor vehicles, aircraft and ships of the need to achieve certain standards of roadworthiness/airworthiness/seaworthiness and to purchase adequate insurance cover. Thus it can be envisaged that as the density of orbital traffic increases, the common interest that users of near-earth space have in preserving a high level of safety through minimising the risks of collisions between spacecraft, will lead to the formation of further international agreements covering these matters.

1. U.N. Doc. A/AC.98/2 (June 1959).
2. S N Hosenball, 1983, " Space Law: Current Status and Issues", AAS 83-220
3. Treaty on Principles Governing the Activities of States in the Exploration and Use of Outer Space, Including the Moon and Other Celestial Bodies (Outer Space Treaty), Jan. 27, 1967, 18 U.S.T. 2410, T.I.A.S. 6347, 610 U.N.T.S. 205 (effective Oct. 10, 1967).
4. V A Chobotov, 1982, " The Collision Hazard in Space", Journal of Astronautical Sciences, Vol.30, No.3, pp.191-212
5. N.A.S.A., Orbital Debris, NASA CP 2360, (1985).
6. A.I.A.A., Space Debris: A Position Paper, Tech. Committee on Space Systems, AIAA, (1981).
7. R C Reynolds et al, 1982, " Man-Made Debris Threatens Future Space Operations", Physics Today, Vol.35, No.9
8. Convention on International Liability for Damage Caused by Space Objects, March 29, 1972, 24 U.S.T. 2389, T.I.A.S. 7762, U.N. GAOR, 26th Sess., Supp. No. 29 (Doc. A/8429) (effective Oct. 9, 1973).
9. U.S. Dept. of Energy, 1978, " Satellite Power System Reference System Report", DOE/ER-0023
10. P E Glaser, 1982, " The Solar Power Satellite - Progress So Far", Interdisc. Sci. Reviews, Vol.7, No.1, pp.14-29
11. U.S. Dept. of Energy, 1979, " Satellite Power System International Agreements", HCP/R-4024-08
12. K Wiewiorowska, 1979, " Legal and Political Problems of the Solar Power Stations in Space", 79- IISL-03
13. J Busak, 1981, " Legal Aspects of the Transmission of Electric Power by Radio Frequencies", Telecomms. Journal, Vol.48, No.6, pp.324-7
14. D M Ashford, 1984, " Space Tourism - Key to the Universe?", Spacefl., Vol.26, pp.123-9
15. Society Expeditions, 1985, " Space Tourism Could Drive Space Development", Proc. LS Space Development Conference
16. P Q Collins and D M Ashford, 1986, "Potential Economic Implications of the Development of Space Tourism", IAA-86-446
17. T Williams and P Q Collins, 1983, " Design Considerations for an Amateur Solar Sail Spacecraft", IAF-83-395
18. I Bekey and P A Penzo, 1986, " Tether Propulsion", Aerospace America, Vol.24, No.7, pp.40-43
19. Potential Operators of Position Locating Satellites File Plans, Aviation Week and Space Technology, Vol.122, No.22, pp.371-9, (1985).
20. L G Taff, 1983, " Satellite Debris: Recent Measurements", J. Spacecraft & Rockets, Vol.23, No.3, pp.342-6
21. U.S. Office of Technology Assessment, 1981, " Solar Power Satellites [2]", OTA-E-144
22. U.S. National Research Council, 1981, " Electric Power from Orbit", Nat. Acad. Press
23. U.S. Dept. of Energy, 1981, " SPS - Electromagnetic System Compatibility", DOE/ER-0096
24. Convention on Registration of Objects Launched Into Outer Space, January 14, 1975 28 U.S.T. 695, T.I.A.S. 8480, U.N. GAOR, 29th Sess., Supp. No. 31 (Doc. A/9631) (effective Sept. 15, 1976) Geophysical Characteristics of Earth Orbits