If a probe is to be dispatched to a 'target of opportunity' then its launch date cannot be set in advance. Instead its launch date will be determined by the advance notice available of the detection of a suitable target. The amount of such notice is very dependent on the thoroughness with which the sky is being surveyed for objects such as NEOs. Until recently such surveys have been on a small scale and identification of close-passing NEOs has been on a serendipitous basis. However the growing awareness of the hazard posed by such bodies has led to the introduction of more comprehensive sky survey programmes [Gehrels94, Near Earth Asteroid Tracking]. Such surveys will identify much more of the NEO population and will allow identification of more Earth-approaching NEOs. Based on the ratio of the estimated NEO population to that currently observed then the number of known close approaches (< 0.05 AU*) can be very roughly extrapolated to an actual figure of perhaps 10 or more a year. The range of 0.05 AU (7.5 x 10⁶ km) thus provides a reasonable abundance of targets, assuming continued search programmes, and as discussed in Section 7 provides a reasonable flight time to the target for achievable probe velocities. However, typical warning time may well be a matter of weeks, so the launch warning will be correspondingly short.

[*One astronomical unit, or AU, is the average distance from the Earth to the Sun, i.e. 1.496 x 10⁸ km.]

During the initial mission definition two main options were identified for allowing rapid dispatch of a probe to a target NEO:

• Parking Orbit. The probe is launched into parking orbit where it awaits the arrival of a suitable target. It is then injected into an intercept trajectory.
• Direct Injection. The probe is stored until a target is identified, when it is launched on a direct intercept trajectory.

Parking Orbit - Advantages. Launch into parking orbit avoids the problems of procuring a launch at short notice. A dedicated launch can be procured for a set date or a launch can be shared with another satellite on a larger launcher, sharing launch costs and/or taking advantage of spare capacity. Furthermore this opens up the prospect of building several rapid response probes and launching them as a group, giving an economy of scale and a constellation of available probes. Whilst in parking orbit the probe or probes could carry out other observations on a continuing basis; alternatively a 'hibernation' mode could be used to reduce mission support costs. To avoid the mission being of indefinite duration a nominal contingency target could be selected from the list of forthcoming known NEO close approaches and the probe sent to it if no suitable target had been identified beforehand.

Parking Orbit - Disadvantages. Whatever parking orbit is chosen (and the choice may be limited in the event of a shared launch) it is unlikely to be optimal for subsequent injection into whatever intercept trajectory is required. To avoid the very large delta-V penalties associated with plane-change manoeuvres the parking orbit should lie in the plane of the eventual transfer trajectory. As the transfer trajectory is not known in advance then either a limit on the range of available intercept trajectories must be accepted or the probe's delta-V capability would have to be increased. Such an increased would involve either an increase in overall mass or a reduction in dry mass. Furthermore once launched the probe would be exposed to the environment of space and would thus have to be engineered for a much longer lifetime than for the direct injection case. Even if the probe is operated in a 'hibernation' mode in parking orbit, some mission support will still be required, whilst any parking-orbit science mission would require prolonged flight operations support.

Parking Orbit Options. If a parking orbit is used, then a number of orbit options are available. However, as mentioned, these options may be constrained in the event of a shared launch.

• Low Earth Orbit (LEO). LEO requires less energy to reach than any other orbit. However, it subsequently requires more energy to achieve an intercept trajectory from LEO. Furthermore, any plane change will have to be carried out at injection (unless a series of transfer orbits is used) with a very large delta-V penalty. If a shared launch is desired, launches into LEO are moderately frequent, albeit often to high-inclination sun-synchronous orbit. The radiation environment is fairly benign, but for a low parking orbit atmospheric drag and atomic oxygen erosion would be problems.
• Geosynchronous Transfer Orbit (GTO). Injection into GTO requires little less energy than direct injection, but the frequency of launches to this orbit make the possibility of a shared launch worth considering. To do this requires that the probe's mass be small compared to that of a communications satellite, although the growth trend in communications satellite masses towards the 3,000 kg level makes this an easier target to attain. Once in GTO only modest extra delta-V would be needed to enter an intercept trajectory, but the range of available trajectories would be limited by the alignment of the original GTO; this alignment would change with time due to nodal precession and apsidal rotation of the orbit. However, it would be possible to achieve considerable plane change at apogee with the available delta-V as opposed to the LEO case. The radiation environment in GTO would be harsh, although experience with microsatellites in highly eccentric orbits such as STRV and the Phase 3 AMSAT series [Wertz96] shows that this is not necessarily a fatal objection.
• Exotic Parking Orbits. Very highly eccentric orbits (r_apogee> 10⁵ km) offer the advantages of GTO with lower radiation exposure and even lower injection and plane- change delta-V requirement, but luni-solar perturbations could make orbit maintenance difficult. Storing the probe at, or in quasi-stable orbit around, one of the Earth-Moon Lagrange libration points (e.g. the L2 halo orbit or the L4/L5 Trojan points) would allow a wide range of injection trajectories to be reached, especially if lunar flyby was used. However, considerable delta-V might be needed to reach these orbits and low-energy trajectories away from them could take too long to provide the required quick response. The same is true to an even greater extent of storage at the Earth-Sun L1 halo orbit; such a mission profile has actually been used for a comet flyby (ISEE-3 / ICE in 1984 to Comet Giacobini-Zinner) but several months of lunar gravity assists were required.

Mission Profile Selection. On initial examination neither the launch-on-demand or the parking orbit option is particularly attractive; this is an indication of the unusual and difficult nature of this mission. However, if it is to be undertaken then launch-on-demand is probably the preferable option. Whilst the use of a parking obit has many advantages it suffers from the serious drawback that there is no such thing as a 'generic' parking orbit. To be able to enter any desired intercept trajectory would require the probe to have a large delta-V capacity that would drive its mass and cost upwards. It was therefore decided to pursue the design study with the launch-on-demand option in mind. This poses significant challenges in the choice of launch vehicle, but as discussed in Section 4 current developments in low-cost access to space may make this option more attractive.