Astrodynamics Group

Autonomous Orbit Determination and Control

This year the level of activity on the Sun reaches a maximum in its 11 year cycle. The effects of solar storms are to increase the level of radiation within the magnfgetosphere, increasing the risk to damage or debilitate satellites in Earth orbit or even cause dramatic effects such as the loss of power in cities. As energy from the hot solar wind plasma funnels its way down into the atmosphere, the atmosphere heats up and expands, causing an increase in atmospheric drag reducing the orbital altitudes of satellites in low Earth orbit. The Astrodynamics team have been monitoring the effects of atmospheric drag using a new orbit estimation filter to estimate orbital parameters. This method is based upon advanced symplectic methods to provide accurate orbit modelling whilst minimising the data processing time. GPS data is collected from Surrey satellites and processed each day to provide updated estimates of drag rates, which are then correlated with the level of solar activity.

Experimentation with the Surrey minisatellite UoSat-12 continues with a demonstration of autonomous orbit maintenance. The satellite was manoeuvred into a repeat ground track orbit so that the satellite would pass over exactly the same point on the ground every 102 orbits, and this orbit was maintained against the effects of atmospheric drag for the past year using an autonomous control system. The system maintained the orbital altitude to within 5 meters of the resonant value. This was only the second time that autonomous orbit maintenance has been demonstrated by any group, and the level of accuracy of 5 meters is currently a challenging level of precision.

Orbit Modelling

Models of satellite orbits in LEO have been developed at Surrey to provide simple analytic descriptions of the complex evolution of satellite orbits around the Earth. This analytic model is nevertheless accurate to a few hundred meters and so adequately describes the positions of satellites for many practical purposes. The work has been extended this year to include the effects of longitudinal variations in Earth shape and density. These effects are the principal source of error in previous models, and have halved the modelling errors from this analysis. The effects of these terms are also crucial in the study of satellites at geostationary orbit.

Attitude Sensor Calibration

Many of the characteristics of satellite subsystems, such as the propulsion system, which are used for orbit maintenance, are affected during launch. The level of thrust as well as the thruster alignment and the effects of thruster exhaust impinging upon other subsystems all change the effectiveness of the thruster, which can therefore only be truly determined once the satellite is in orbit. The ability to perform in-orbit calibration then enables satellites to automatically characterise. Two methods have been developed: the first using a simple bang-bang controller on the satellite's reaction wheels. This produces a known disturbance to the satellite. The second method uses a number of single pulse thrusts from each thruster and the effects are filtered. These two methods have shown good agreement on the level of thrust and the direction of thrust.

Work on the UoSat-12 attitude sensors has focused on the sun sensor. A new model for the solar irradiation has been developed and used for the calibration of the sensor in orbit. This new model has greatly improved the attitude accuracy of the sensor to below 1 degree. Tests of the attitude and orbit control algorithms have led to experiments on imaging the Moon during 1st quarter. The accuracy of the control algorithms has been sufficient to image the Moon using both low and medium resolution cameras on the satellite.

Formation Flying

Following the successful launch and in-orbit verification of the SNAP-1 nanosatellite, we have used derivatives of our orbit maintenance algorithms to calculate a series of manoeuvres aimed at moving SNAP closer to TSINGHUA-1, a microsatellite launched from the same vehicle. The nanosatellite separated into a lower orbit and the effects of atmospheric drag caused a rapid separation between the two satellites. This relative orbit between the two satellites was determined and a sequence of thruster firings on the nanosatellite was programmed to bring the two satellites together again. The performance of SNAP's thruster and the accuracy of our orbit control algorithms are being monitored through GPS measurements generated onboard the satellite and processed by our batch symplectic filter.

Imaging Scheduler

A fast imaging prediction algorithm has been developed and implemented. This algorithm has been successfully used for image scheduling. The novelty of the algorithm is that it is over 100 times faster than previous methods, and is therefore suitable for on satellite operation. This will enable future autonomous satellites to schedule their own image capture sequence with knowledge of their orbits. The method consists of three steps: a coarse search, an imaging estimation, and image timing refinement. The algorithm has been extended to include satellites with canted cameras and a number of spacecraft slew manoeuvres.

In the future more and more people will have direct access to satellites for obtaining data from space. This data may be used for scientific research, education or direct communication between users. The ability for satellites to respond to a number of users making asynchronous requests and having to schedule their tasks to respond to them in reasonable time is a complex problem opened up by these future satellite capabilities. Work has begun on investigating the capacity of satellites for multiple user requests for images and their download at predefined ground stations. This work extends the concepts of queuing theory to determine the level of storage required on a satellite to be able to service a reasonable level of user expectation.

Deep Space Missions

The interests of the group have extended beyond Low Earth orbit with the prospect of missions to send satellites out to geostationary orbit, or various parts of the Earth's magnetosphere and even to the Moon. To address these future mission plans, work has begun on new symplectic algorithms suitable for the propagation of highly elliptical orbits. Comparison with other, more conventional, schemes has demonstrated that the same accuracy is achieved with significantly shorter integration times. This work will form the foundation of new orbit determination algorithms suitable for satellites that travel out beyond the GPS constellation.

In addition, work is now underway on methods for constructing transfer trajectories to the Moon and other targets in the inner planet region of the Solar System. These operate on principles that are significantly different than the orbit manoeuvres performed in LEO and therefore an essential element in the realization of SSTL's deep space ambitions.