1.1

The market

Not long ago, earth imagery from space was a rare commodity with just a few key players in the industry, but the advent of the excellent Copernicus program [1] has brought large amounts of high quality data for everyone to use. Whilst providing a lot of opportunity for the scientific community, it has meant that the market for earth imagery has changed. Copernicus, through its Sentinel series of satellites, provides a wide range of very high quality data of medium/high resolution but with a limitation on temporal resolution. Now the focus in the commercial sector has shifted to improving revisit time, something Copernicus cannot do without significant additional expense.

The emerging demand for low cost, high resolution, regular imagery and the arrival of the ‘NewSpace’ or ‘Space 2.0’ era has spurred the development of a number of different nano/micro/small satellites all aiming to quench the thirst for data. There are some companies proposing vast constellations of CubeSats, others aiming at more capable >100kg platforms, and everything in between. The customer base has a strong knowledge and influence on performance and is interested, not just in low GSD numbers, but also in high MTF across the range of spatial frequencies. Therefore the challenge set to the satellite manufacturers, fledgling and experienced, is that of quality and quantity.

1.2

Carbonite-2

Surrey Satellite Technology Ltd. (SSTL) has a long heritage of earth observation satellite systems [2] making it a market leader in the design, build and operation of optical imaging systems and their respective platforms. These have included push-broom, multi-spectral, hyper-spectral, wide-swath systems and more recently high resolution video.

Carbonite-2, launched in January 2018 [3], demonstrated the world’s first commercial high-resolution full color video from space; it achieved a GSD of 1m with good MTF. This mission was a follow-on from Carbonite-1 which was a rapid build, very low cost system using a COTS telescope. Some modifications and improvements to this have resulted in some very impressive imagery, a frame of one such video is shown in Figure 1. A sample of the videos from orbit can be accessed from SSTL’s YouTube channel, SSTLTV [4].

Figure 1:

Dubai airport acquired by Carbonite-2, 2018. Credit: SSTL.

2.1

Design philosophy

A traditional Cassegrain-type telescope has a large amount of unused volume between the primary mirror and the secondary mirror; the principal aim of this telescope is to reduce this volume during launch. The secondary mirror can be held in a stowed configuration near the primary mirror and then deployed to a near-nominal position once in orbit; further fine alignment can then be performed.

There are a number of potential benefits that this technique can provide to a mission. Firstly, the volume savings in the launch vehicle can allow the use of different launch slots or vehicles, this is discussed in some detail in Section 4. Secondly, the deployment of an optical system can minimize the design restrictions on the opto-mechanical structure as the alignment of the optical elements isn’t required to be maintained during the launch event which has the potential to save mass. Thirdly, the design has been developed with scalability in mind, and so it could be used in the future with a much larger system for higher resolution EO imagery or astronomy.

A loose goal of this payload development is to improve on the imagery provided by Carbonite-2 at one third of the volume at launch. The technique proposed here might not be the optimal or most elegant solution to a deployable telescope, but it designed to be rugged, low-cost and to meet the technical requirements.

2.2

Optical design

The optical design is a modified Dall-Kirkham variation of a Cassegrain-type telescope, with an elliptical primary and spherical secondary mirror, a series of COTS lenses are used to minimize off-axis image degradation. A novel baffling system has been designed which deploys with the telescope and manages stray light, but is purposefully omitted from Figure 2.

Figure 2:

Optical design of telescope

This design was chosen for its performance, simplicity, compactness potential, and ease of alignment. The spherical secondary mirror is the optical element which needs to be manipulated in-orbit to achieve alignment. Typically a rotationally symmetric mirror requires five degrees of freedom to be aligned, but with a spherical reflective surface, this can be reduced to three. Tip and tilt about its center of curvature can be traded off for decenter, thus removing two degrees of freedom. This trade-off can introduce some clipping of the beam by moving the secondary away from its nominal position in the optical axis, but with deployment induced misalignments expected to be small, the impact of clipping is likely to be negligible.

The relationship between decenter and tilt is displayed in Figure 3; it shows for a given starting decenter of the secondary mirror induced by the deployment and launch events, the tilt about the opposing axis that is required to compensate for it. The extreme case is a 0.5 mm decenter, for which a 0.1 degree tilt will be required. This helps confine the requirements of the alignment mechanisms, and supports the decision to use a Dall-Kirkham design.

Figure 3:

Relative MTF degradation from nominal at half Nyquist against tilt of secondary mirror for a range of mirror decenters

2.3

Mechanisms

The dynamic nature of this telescope means that there are a number of different mechanisms that need to be employed in order to ensure that the telescope can be safe during launch, deployed accurately and then aligned. The first is the deployable telescopic barrel as it is essentially one large mechanism in itself. Secondly, there are three identical fine alignment mechanisms that interface between the barrel and the secondary mirror. Finally there is a hold down and release mechanism (HDRM) designed to ensure that the other parts are secure during the launch event.

2.3.1

Deployment

This component provides the critical distance member between the primary and secondary mirrors. The main goal of the deployment mechanism is to minimize this distance when stowed, and then return to the nominal distance in an accurate, repeatable and low shock fashion.

A number of different concepts have been investigated, including trusses, telescopic tubes and barrels, and then a tradeoff study was performed to identify the preferred solution. The criteria were: mass, length reduction, thermoelastic stability, complexity, scalability, stray light management, cost and repeatability.

The selected approach consists of three concentric carbon-fibre barrels; the lower most and narrowest is mounted to a bulkhead which interfaces to the spacecraft. The middle and upper barrel are deployed post-launch simultaneously by lead screws spaced at 120° around the circumference, driven by a ring gear and redundant motor arrangement in the base. They are driven into V-blocks which provide high positional repeatability.

Figure 4:

The deployable telescopic barrel, stowed (left) and deployed (right)

The chosen concept exhibits a high tolerance to launch loads when stowed and to micro-vibration when deployed, therefore enabling the imager to perform well during operation.

2.4

Alignment system

In partnership with the Dynamic Optics and Photonics Group at the University of Oxford, utilizing their experience in adaptive optics systems [5], novel autonomous alignment algorithms have been developed to improve the on-ground build procedure of space based telescopes. Now the partnership has been extended to develop algorithms for autonomous alignment in-orbit, and a different approach is taken due to the change in operating conditions in space compared to the lab.

In-orbit alignment can be carried out using an image-based method as illustrated in Figure 5. Suppose the image is initially blurred by an aberration with a magnitude of a. This aberration could be defocus for example, owing to an incorrect z-position of the secondary mirror. The amount of blur in the image can be characterized using an image quality metric M, such as the magnitude of the high frequency components of the image power spectrum. By measuring the image quality metric when applying three bias magnitudes 0, +b, and –b, by adjusting the z-position of the mirror, the required correction can be determined. This procedure can be repeated to account for tip and tilt errors. This can be done quickly, on the fly, whilst staring at a target.

Figure 5:

Proposed image based alignment technique

2.5

Detector and FEE

Carbonite-1 and 2 used a COTS CMOS detector, a large area array with an RGB Bayer pattern, the same sensor has been baselined for the deployable telescope. Read-out electronics are developed in-house, including a processor on which the alignment algorithms are run.

4.1

Case study

A small constellation of “Carbonite-2-like” spacecraft [6] in Sun Synchronous Orbit (SSO) would be suitable for launch from any of the new or emerging small launch sites around the world including the proposed UK launch site in Scotland.

We can consider a case study of a constellation of 6 spacecraft in a 500km altitude SSO that has 3 planes of 2 spacecraft each. This would assume the availability of a Rocket Lab “Electron” class of launch vehicle with a ~200 kg payload capacity to 500 km SSO. It is an SSO constellation so orbit planes are not evenly spread; the first plane with Local Time of Ascending Node (LTAN) = 10:00, second plane with LTAN = 13:00 and third plane with LTAN = 16:00.

Figure 7:

Constellation of 3 planes of 2 satellites each

The 3 planes are offset by 45 degrees with spacecraft separated by 180 degrees in-plane. A “normal” Carbonite-2-like spacecraft occupies all of the available volume under the fairing of an Electron class vehicle and a single Electron rocket costs ~ £4.5M.

Figure 8:

Basic constellation performance

4.2

Benefits of a deployable telescope

The constellation is attractive from a customer point of view as it samples at different local times rather than a fixed time, as is the case with a single larger spacecraft. It can be very useful for many economic, business, security and military applications. However, to do this kind of mission with launching all 6 spacecraft at the same time on a single rocket is impossible. All 6 satellites would need to be launched into the same orbit plane and there is no rocket available today that can provide a 90-degree node shift to deploy into three separate orbit planes directly. The constellation needs to separate into 3 separate orbit planes which could be achieved by a small manoeuvre to shift the relative altitude of the satellites so they experience a differential node precession rate (the “J2 Effect”).

However at SSO inclinations the rate of separation that can be achieved within the ΔV budget of a Carbonite-2-like spacecraft is very low, even spending 100 m/s to do the drift would only “reduce” the drifting time needed to ~10 years. The only really feasible option is to envisage a launch directly into each of the three orbital planes, it is worth noting that piggyback opportunities are also very rare into afternoon LTAN SSO orbits.

With a spacecraft that occupies all of the volume but only a half of the lift mass of the rocket, then 6 separate launches would be needed. Very basically, a deployable payload allowing 2 spacecraft under the fairing of the rocket reduces this to 3 launches, halving the launch cost. There is also the potential for a quicker constellation deployment (starts to make money earlier) and it reduces significant other costs to the user (e.g. reduction in launch campaign, shipping etc.).

4.3

Launch campaign case study

A pragmatic assumption on actual launch cadence that will be feasible, would be for a launch facility to provide a launch service roughly every 60 days, which is about 6 launches per year.

For the case when a single spacecraft is on a single rocket (6-off launches), each satellite goes directly to its assigned slot in its orbit plane and the total time between first and last launch is 300 days. The total launch cost is £27 M and there are 6-off launch campaigns to run.

For the case when deployable optics allows 3-off launches of 2 satellites, each pair of satellites is launched together into the assigned orbit plane then the satellites have to separate by 180 degrees in-plane. Each one can move 90 degrees away from its nominal injection location (one goes “forward”, one goes “backwards”). This would take about 40 days in 500 km SSO if ΔV of 2 m/s is used on each spacecraft (which is compatible with a Carbonite-2-like gas propulsion). In-plane drifting can take place whilst other launches are happening and so total time between first launch and last satellite arriving at its assigned location = 160 days. The total launch cost is £13.5 M and there are 3-off launch campaigns to run.

In this case the deployable optics allows a nominal cost saving of approximately £13 M in terms of launch cost and there are additional potential savings in shipping, insurance, launch campaigns etc., which is not quantified but will be nonnegligible. There is also a decrease in constellation deployment time of ~140 days (~4.5 months).