ECOSat-III will further The University of Victoria’s contribution to the geophysical service and to research and development of communication systems on nano satellite systems.
ECOSat-III will be flying a primary hyperspectral imaging payload, supported by an experimental communications system and attitude control system.
- Mission A will provide hyperspectral imagery of Canada at 150 metre resolution.
- Mission B will downlink the hyperspectral imagery over a custom-developed 40 MBit communications system.
- Mission C will extend the University of Victoria’s experience in Attitude Determination and Control systems with the addition of momentum wheels and more complex attitude determination algorithms
- Mission D will provide accurate initial orbit determination and low rate telemetry through the use of an experimental below-the-noise-floor communications system.
Mission A: Hyperspectral Imaging
Hyperspectral imaging is an emerging technology enabling advanced spectroscopy from a remote platform.
Put simply, hyperspectral imagery is all about seeing colours we can’t normally see. The human eye sees images as a combination of three colors: red, green, and blue, centred at the wavelengths of approximately 575 nm, 535 nm, and 445 nm respectively. This is accomplished through three types of “cones” on the retina, each of which is sensitive to a different wavelength.
Colour photography duplicates this method: a modern digital camera uses a special filter called a Bayer array which separates light into red, green, and blue components which are read at each sensor element. These values are interpolated using a demosaicing algorithm, which assigns an RGB value to each pixel in the image.
Hyperspectral imaging takes a completely different approach. There are several methods, but the most common is spectral scanning, which works on the following principle: when entering a prism, light separates (diffracts) based on its wavelength. An example of this phenomenon can be seen on the album art for Pink Floyd’s Dark Side of the Moon, seen at the left.
A single line of the image is scanned through a slit and fed into a special prism called a diffraction grating, which converts the 1-D line to a 2-D image which is imaged by a CCD or CMOS sensor. The horizontal pixels represent spatial information, while the vertical pixels represent wavelength (spectral) information. The process is then repeated for each line of the image.
The result is a hyperspectral data cube, or hypercube. An example of a typical hypercube can be seen to the right, taken from NASA’s highly successful Landsat program.
ECOSat-III will seek to accomplish similar performance to a large hyperspectral satellite at a small fraction of the cost. Over the course of at least one year, ECOSat-III will continuously image Canada at 150m spatial resolution with a spectral range of 40 bands covering 400-1000nm.
Mission B: S-band Communications
Cube satellites have traditionally had very limited amounts of data bandwidth. ECOSat-III will operate in the S-band Earth Exploration Satellite Service with a raw downlink capability of 10 MBit – a first for a cube satellite. This communications system is being developed entirely in-house at the University of Victoria.
Mission C: Attitude Determination and Control System
To further the experience and knowledge within the University of Victoria, ECOSat-III will host a more advanced Attitude Determination and Control algorithm along with momentum wheels developed in house.
Mission D: Below-the-Noise-Floor Communications System
A side effect of having an extremely high-rate communications system is a very tight link budget requiring a very directional ground station dish.
In order to point the dish on the ground station, very precise knowledge of the satellite’s orbit is required. For a cube satellite, this is done with a combination of radar tracking and good luck. For ECOSat-III, our ground station requires such a narrow beamwidth that NORAD orbital elements will likely prove unreliable, and the last thing we want to do is search the sky in three degree increments until we can establish contact with the satellite.
Instead, we will have the satellite tell us where it is in orbit. We will accomplish this using a very low-rate communications system capable of operating well below the noise floor. Using a combination of spread-spectrum techniques and judicious channel coding, we can receive a low-rate signal (on the order of 500 characters per minute) from the communications system tens of decibels below the noise floor and effectively independent of doppler shift.
Receiving this signal so far below the noise floor means that we do not require a directional antenna for initial contact. We will simply use a conventional omnidirectional turnstile antenna, which will constantly listen for the signal from the satellite. The satellite will continuously transmit a short data packet containing its current time, state vector, and health. From these data, we can compute the orbit of the satellite precisely, and switch to the high-rate communications system once we know where to point our ground station dish.
This system represents a significant step forward in initial contact procedure for cube satellites. Satellite builders will identify their satellites’ orbits as soon as they overfly their ground stations, and will not have to worry about the accuracy of radar determination on initial orbits.