AI-in-Space for operations and tech R&D

Rapid developments in recent years have revealed and proved the potential of AI techniques in a variety of sectors, opening doors to a wide range of possibilities and innovations.

While certain segments of the space industry are already implementing and utilizing AI applications (e.g. Earth Observation), in many areas such techniques are still in infancy, especially when it comes to the integration onboard spacecraft, the physical extensions of the industry outside the planet. Many activities conducted in microgravity, and outer space in general, could greatly benefit from the implementation of AI techniques.

Moreover, AI-related activities are not limited to research and technology development. AI is also becoming an increasingly important topic and tool in providing STEM education and an innovative way for the young generations to learn about microgravity and space while developing their own AI applications.

Thus, combining education and space activities, the first AI-in-Space challenge has generated a number of proposals for AI-enhanced space activities, using the ICE Cubes AI-Box as a testbed and playground.

The winners of the AI Space Challenge, ‘Ad Astra’ team proposed an innovative solution which aims to enhance data security in space by providing greater protection to the experimental data collected and stored on the ISS. This ML model will be trained using data acquired by the sensors of the AI-box. By predicting single event upsets, they aim to reduce the uncertainties of data loss on the ISS and create a protocol to preserve the data that are predicted to be impacted. The performance is hoped to be improved thanks to the acquirement of other types of data on the ISS to increase its accuracy and reliability. Further research could be carried out in order to predict single events upset for specific devices.

Another team (Blue Magpie) proposed implementing air pollution warnings on board the space station, and furthermore, providing standard threshold warnings for dust-free space. Moreover, the particle concentration, pressure, CO2, humidity, temperature, acceleration, and magnetic field data, required to create a warning system for hazardous particulates, can all be acquired by the sensors of the AI Box.

By implementing deep learning algorithms, collecting and utilizing the data acquired by the AI-box and its sensors, it is also possible to improve existing methods and technologies like the accuracy for the magnetometer-only position determination, the idea proposed by team Out of this World. This can be done by training the program which determines the position of spacecraft and satellites. The space environment is necessary to conduct the activity as the data for AI training and filter selection needs to be acquired in that very environment in orbit.

AI techniques can be used for space operations. Using deep learning features, the AI-Box sensor parameters and their corresponding position results can enable to calculate whether the ISS has deviated from its orbit or issue a warning if it is on course to deviate.

Using the collected data of geomagnetic field and solar wind speed to train and establish the model of relationship. Then using the magnetometer to collect the magnetic field information during orbiting to predict the solar wind speed and hence prevent the possible damage due to strong solar wind.

Other ideas proposed dealt with human spaceflight applications, such as improving the sleep quality of the astronauts by using AI to analyze the impact of environmental factors and recommend suitable adjustments. Various AI applications can be implemented to improve and optimize the daily schedule of astronauts in order to reduce stress and protect their mental health.

Thus, artificial intelligence is gradually becoming an important, in some cases essential, tool for the development of applications and boosting research, for the benefit of life on earth and in space. The many advantages it offers create a real springboard for innovation. AI-in-space is already showing its potential in the Space Economy but we are only at the beginning of discovering its full potential.

In view of STEM applications, AI can be used as a real innovative tool in courses, trainings and workshops, providing strong capacity building grounds. It can also be the core of new projects, challenges and hackathons, boosting international competition and collaboration.

For more information about the AI-Box and its technical capabilities, download the brochures below.

Check out the links below to read more about milestones in the development of AI.