Issue #1(7) 2016 Astronautics

Rosetta – Starting from Scratch

New Norcia ground station
New Norcia ground station
Paolo Ferri Head of Mission Operations

osetta was initially conceived in the early 1980s well before an international fleet of spacecraft, including the ESA Giotto probe, flew by the nucleus of comet Halley in March 1986, imaging for the first time in history the nucleus of a comet. The Halley encounter was a major event in the history of Solar System exploration. However, a fly-by brings a spacecraft in the vicinity of a comet nucleus only for a short time and at very high relative velocity. Pictures and measurements of the nucleus were executed only for a couple of hours around the encounter. A very limited way to study these mysterious celestial objects.

The idea behind Rosetta was to take the next step: to rendezvous with a comet - that is reaching the comet and staying with it, flying around it at a few tens of kilometres distance, imaging and measuring the nucleus and its environment for months or years while it travels around the Sun.

The initial idea was even more ambitious - to land on the surface, collect a material sample and bring it back to Earth

To achieve this the spacecraft had to reach the orbital energy of the comet, but even the most powerful launcher cannot provide sufficient dynamic energy to the spacecraft. For this reason the mission had to make use of a trick called gravity assist, whereby a trajectory is designed such that the spacecraft flies in the vicinity of a planet, and uses its gravity field for a further acceleration in its heliocentric orbit - thereby stealing a bit of the planet’s orbital energy. This makes the journey very long but at the end the spacecraft reaches the comet with the same orbital velocity as the nucleus, and is able to fly in formation with it around the Sun.

The initial idea was even more ambitious: to land on the surface, collect a material sample and bring it back to Earth. However, early studies showed that a comet sample return mission would be too expensive, too risky and inefficient in terms of scientific return. Better to use the available resources to carry sophisticated scientific instruments to the comet, perform investigations locally and transmit the results via radio signals back to Earth.

New Norcia antenna during construction New Norcia antenna during construction

But the idea of a lander remained: if one goes all the way to rendezvous with a comet, why not taking a small landing module as well and attempt what, in initial studies looked a crazy, almost unthinkable objective: landing on the surface of a small, unknown object of a few km diameter, with an active surface that emits gas and dust in an unpredictable way, and travels in space at several tens of km/s, at a distance from Earth of several hundred million km!

This was the concept that survived when the European Space Agency’s (ESA) Science Programme Committee approved the mission in 1993. The target was a small, estimated 1-2 km radius object lost in the enormous distances of deep space, called comet 46P/Wirtanen.

After completing the mission’s conceptual design in early 1996, ESOC began the preparation of the ground segment and mission operations. A Ground Segment Manager (GSM), Manfred Warhaut, was appointed to oversee all ESOC activities, from the initial operations concept definition throughout the preparation phase until the actual operations execution in flight. A few months later Manfred nominated me to join his team as Spacecraft Operations Manager (SOM).

Back in 1996, ESOC had only very limited experience in flying interplanetary probes and no infrastructure at all to support this type of space operations. Basically, we had to start from scratch.

Manfred had a clear vision about what was required: first, collect all the valuable experience already available in-house. He formed a small group of experienced engineers to cover the main areas of expertise. Then he established links to NASA, part of the international cooperation efforts on Rosetta and very experienced in interplanetary operations.

Even NASA hadn’t done many of the things that we were attempting. We had to face them on our own - and we had less than six years

We started regular meetings with NASA to establish the basis of the cooperation and to learn as much as possible. But even NASA hadn’t done many of the things that we were attempting: flying a solar-powered spacecraft at 800 million km from the Sun, hibernating a spacecraft for 2.5 years, navigating around a small, active comet nucleus, and landing onto its surface. This was the first time in the history of spaceflight that such challenges were faced. We had to face them on our own, and we had less than six years before the fixed launch date of January 2003.

Manfred’s vision went beyond that. He strongly felt that ESA could only fly an ambitious, historical mission like Rosetta if it had its own antennas to communicate with it, and not rely solely of the support of NASA’s Deep Space Network, as in the original plan.

This meant that ESA had to build a large deep space antenna somewhere in the southern hemisphere and develop sophisticated instrumentation - such as cryogenically cooled amplifiers and maser clocks - needed to establish and maintain a radio link with a spacecraft over several hundreds of millions km distance. When Manfred first told me of his idea to build ESA’s own deep space antenna I told him he would never convince our funding bodies to cover such an expensive project, especially with just one interplanetary mission, Rosetta, on future ESA plans.

But he was stubborn and determined enough to finally convince everybody: within a few years, and before the launch of Rosetta, ESA had built, under his management, a deep space station with a 35m antenna in New Norcia, a small site 130 km north of Perth, Western Australia.

Rosetta and Philae being prepared for launch Rosetta and Philae being prepared for launch

This was sufficient for Rosetta but Manfred’s ambition was to give ESA an independent access to interplanetary space. So he continued his efforts and, also thanks to the approval of further interplanetary missions like Mars Express and Venus Express, he managed to build a network of three deep space antennas: after New Norcia, came Cebreros in Spain which was operational in 2005 and Malargue, Argentina, at the end of 2012.

In the meantime the spacecraft design phase had started. The prime contractor selected was Dornier, a German aerospace company located on Lake Constance. Matra Marconi Space in Toulouse, France, built the avionics subsystem, our primary interface with the spacecraft.

We spent the initial years of design discussing with industry the functionalities we required to operate this complex mission. Rosetta was very new to them as well, so we were all learning on the job, with hardly anyone to teach us how to do it.

The Main Control Room at ESA’s Space Operations Centre in Darmstadt (2002) The Main Control Room at ESA’s Space Operations Centre in Darmstadt (2002)

Towards the end of the millennium the ground segment conceptual phase was over and development work started in all areas, including flight dynamics systems, mission control systems, simulators and ground stations. In 1999 it was time for me to start building the operations team. My idea was to create a team by not necessarily looking for experience but rather for motivation, solid background and youth.

I needed to plant the seeds for a team that could grow and cover the next 15 years. I had to plan to cope with the inevitable turnover in such a long period, so I selected a good mixture of ages, gender and nationality, covering the various areas of necessary specialisation. Looking back at those times I think the team that we managed to put together was one of the best achievements of Rosetta.

The flight directors and operations managers that, on the day of the landing, one-and-a-half decades later, led the operations, Andrea Accomazzo, Sylvain Lodiot, Elsa Montagnon, Ignacio Tanco, had been hired between 1999 and 2001. They had grown over the years with Rosetta, made parallel experiences or temporarily left the mission, only to return at a later stage. They never abandoned the passion of being part of this incredible adventure and dedicated a large part of their careers and took important private-life decisions always with the clear objective of playing a big role in this fantastic adventure called Rosetta.

While the construction of the deep space antenna in Australia was in its hot phase in early 2001, our mission control team started daily integration activities of the ground systems at the control centre in ESOC.

The first deliveries of the mission control system were full of problems, the system simulator was late, the new interfaces with the ground stations also progressed very slowly. When we first had the opportunity to connect our ground systems to the spacecraft, which was being integrated in Torino, we also found out that the onboard software did not behave as we had specified. The year 2001 was one of hard work and disappointment, delays, testing and retesting with slow progress and no major improvements.

We began to worry about the rapidly approaching launch date: would the spacecraft be ready? Would our ground systems be robust enough? Would we be able to train our flight control team with a realistic and stable simulator?

Rosetta Dedicated Control Room in 2001 Rosetta Dedicated Control Room in 2001

The break-through was the decision by the Rosetta project manager, John Ellwood, to decouple the development and testing of the spacecraft onboard software from the integration and test campaign of the spacecraft flight model. At the same time he allowed ESOC to participate more directly in the onboard software test campaign on the spacecraft engineering model.

Through a number of test sessions the joint teams of ESA and industry managed to identify and resolve the main problems in the onboard software. At the same time the intensified testing directly with the spacecraft allowed us to improve the ground software and to increase the confidence in the ground systems interfacing with the spacecraft. Several long test sessions were conducted round the clock, with two shifts in ESOC and Torino.

Rosetta large solar panels deployment test Rosetta large solar panels deployment test

In the meantime, the simulation campaign at ESOC had started. This is when the mission control team, formed by all the teams required for the critical initial flight operations after launch, come together twice a week and use a software simulator go through all operations, scenarios and flight procedures. In these sessions the simulator is used to inject failures that mission controllers have to recognise and react to, taking the necessary actions to recover and continue the operations.

Simulations are extremely demanding on the teams, as every mistake in any part of the system goes under the spotlight and is criticised, discussed among all the team members. Actions to resolve it have to be completed before the next simulation in just a few days. Running a simulations campaign together with the system test campaign with the spacecraft did put an enormous workload on the team.

The first deliveries of the mission control system were full of problems, the system simulator was late, the new interfaces with the ground stations also progressed very slowly

To add to the stressful situation, our simulations uncovered new bugs in the spacecraft onboard software, which were not detected during system tests, and could not be changed before flight, due to the limited time remaining. Many discussions took place at project level to evaluate the level of risk we were taking by launching the spacecraft with known important software problems.

The avionics team in industry immediately started working on another version of the onboard software to eliminate the problems but this would only be ready for upload to the spacecraft in flight in July 2003. Rosetta would have to fly for a few months with the current software. A risk, certainly but, given that it would be corrected in a short time compared to the long mission and that in this period Rosetta would still be relatively close to Earth, it was considered acceptable.

Rosetta flight control team 2002 Rosetta flight control team 2002

Launch had to take place in a 21 day window opening on 13 January 2003. Missing that window would have meant missing the appointment with our target, comet Wirtanen. Everything in the Solar System constantly moves, including our starting place, our destination and the planets we wanted to use as slingshots to accelerate the spacecraft towards the comet.

Travelling in interplanetary space means finding a road in space and time to ensure the spacecraft arrives at the encounters with the other celestial objects at exactly the right time. The trip to Wirtanen had to start in January 2003 or the comet, and the mission with it, would be lost forever.

In December 2002 the simulation campaign was almost completed. All our tests with the spacecraft were done. Our new deep space antenna in New Norcia, Australia, was ready and waiting to pick up the first signal from Rosetta in space. Our teams were trained and looking forward to the start of the great adventure. The spacecraft was in Kourou, French Guiana, being prepared for launch. We were ready.

If you've enjoyed reading this article, please consider subscribing to ROOM Space Journal to gain full access to current articles and receive your own print and/or digital copies of ROOM delivered direct to your door or electronically.

Popular articles

See also


Ka-band satellite communication and lowering the price of VSATs


ROOM helps raise student space awareness


ROSETTA AND PHILAE: Outstanding climax to pioneering mission

Popular articles

Smallsat manufacturing company AAC Clydespace specialises in advanced nanosatellite spacecraft, mission services and subsystems offering solutions for government, commercial, and educational organisations. Astronautics

Technology transfer and the ‘local’ space market

A graphic simulation of the Starlink constellation, visualising the ground tracks of around 11,500 satellites between 2019 and 2033 (with guesses for the timing of the deployment of the remaining orbital shells, which determines the order in the plot). Opinion

Congested, contested... under-regulated and unplanned