The European Space Operations Center (ESOC) has used Artificial Intelligence (AI) to help ESA’s Mars Express probe in its search for signs of life, present or past, on the Red Planet.
Since January 2004, Mars Express has used its sophisticated instruments to study the atmosphere, surface and subsoil of Mars, confirming the presence of water and searching for traces of life on and under the rocky terrain of the Red Planet.
The space probe generates a large volume of scientific data, which must be transmitted to the Earth at the right time and in the correct sequence to prevent data packets from being permanently lost, since the on-board memory, of limited capacity, is overwrite with the new data collected.
Traditionally, data transmission is managed using planning software that requires the attention of an operator. This software generates the scripts that are sent to Mars Express, specifying when the probe can discard certain data packets. “This is tedious, time consuming and never really eliminates the possibility of losing – permanently – valuable scientific data”, says Alessandro Donati, head of the Advanced Concepts and Mission Concepts department of ESOC, Darmstadt, Germany.
An ‘intelligent’ solution for a complex transmission problem
However, since 2005, Artificial Intelligence researchers from the Italian Institute for Cognitive Science and Technology (ISTC-CNR), led by Dr. Amedeo Cesta, and mission planners and computer engineers at ESOC have been developing a solution for the complex planning problem of Mars Express, using Artificial Intelligence (AI) techniques.
The result of this work is a new tool, named MEXAR2 (Mars Express AI Tool) that is now an integral part of the Mars Express mission planning system.
MEXAR2 works by intelligently planning which data packets could be lost due to conflicts in the on-board memory. In this way you can optimize the planning of the data download and generate the necessary commands to carry out this transmission. “With MEXAR2, virtually any loss of stored data packets has been eliminated,” says Fred Jansen, Head of Mission for Mars Express.
MEXAR2 has significantly reduced the workload of the mission planning team by up to 50 percent compared to the previous manual method. “This optimizes the bandwidth used to receive the data on Earth, and You can have valuable tracking station hours for other missions, “says Michel Denis, Mars Express’s Head of Operations, at ESOC.
MEXAR2 recently won the “best application” award at the ICAPS 2007 international congress, a point of reference in the planning and programming technology sector through Artificial Intelligence.
First deep space European mission flying with IA
Artificial Intelligence provides solutions to very complex problems, and now it has just entered into the field of space mission operations as an added value. “Mars Express is the first European deep space exploration mission that flies using AI tools in the ground segment, which allows us to boost scientific returns while reducing the time and resources employed,” adds Donati.
With the demonstration of the success of MEXAR2, the ESOC and ISTC-CNR scientists are working to apply the current technology in Artificial Intelligence to other problems.
Recently we have successfully worked on the inverse problem: how to optimize the sending of commands to Mars Express, in a project baptized – ironically – as ‘RAXEM’ – MEXAR written backwards.
The Artificial Intelligence technology developed by ESA will also apply to the Advanced Planning and Programming Initiative, an action designed to extend the benefits of AI to other areas and missions, including the long-term planning of observations made by Integral, the observatory in gamma ray orbit of the ESA.
“The same concepts of Artificial Intelligence are applicable to future missions, such as ExoMars, the first European mission that will send a robotic SUV to the Red Planet,” says Donati. He adds: Our success today is a starting point for the implementation of new concepts that allow the next ESA missions to work more autonomously.”