The goal of space robotics research at DFKI is to develop robust and powerful robots that can master difficult terrain and perform complex tasks with the help of innovative technologies and artificial intelligence. The highly specialized, task-specific systems used to date will be replaced by flexible and reconfigurable solutions. This should not only significantly reduce development costs, but also make space travel more efficient and sustainable. At the 14th EASN International Conference, one of the central events for the European aerospace community, the Robotics Innovation Center will present five current papers on different challenges in space robotics. In addition, DFKI will be represented by Prof. Dr. Udo Frese, who will give a keynote speech on 'AI-based Robotics: A Key Technology for Space Applications' and Wiebke Brinkmann, who as a member of the EASN International Scientific Committee will moderate the session 'The Way of Future Orbital and Planetary Robotics'.
The research presented at the conference includes the paper "Prediction-Based Tip Over Prevention for Planetary Exploration Rovers". It describes a system that prevents rovers from tipping over during space missions by using sensors and AI-based algorithms to monitor the robot's stability on uneven terrain and react in time to potential hazards. The paper "Development of a quasi-direct drive motor for walking robots in extraterrestrial environments" deals with the development and testing of quasi-direct drive motors, a special type of electric motor that combines features of direct drive and conventional motors to offer advantages in terms of efficiency and power density. Researchers have optimized these motors for use in space robots that must withstand extreme conditions.
The article "Improving heterogeneous multirobot collaboration for planet exploration" deals with the question how to improve the collaboration of multirobot systems in space missions by integrating software solutions. The goal is to optimize the performance of robot teams and enable more efficient and successful missions. The development of a modular system for space robots is described in the paper "Towards sustainable space exploration: Designing an AI-powered modular toolbox for future planetary exploration". This flexible collection of hardware and software components can be adapted to different mission requirements and is also suitable for terrestrial applications, such as agriculture.
Finally, the paper "Pioneering Sustainable Space Ecosystems through Intelligent Robotics and Collaborative Effort" deals with the question of what is necessary to enable sustainable ecosystems in space. It describes the joint research of DFKI and the University of Bremen on AI-controlled autonomous robots that use resources efficiently and can work autonomously under harsh environmental conditions. Together, the work presented opens up new approaches to make space robots safer, more efficient and more adaptable both in space and on Earth.
The paper in detail:
Title: Prediction-Based Tip Over Prevention for Planetary Exploration Rovers
Authors: Siddhant Shete, Raúl Domínguez, Ravisankar Selvaraju, Jayanth Somashekaraiah, Amrita Suresh
This study presents a novel prediction-based system designed to prevent tip-overs of planetary exploration rovers, thereby increasing their operational safety and reliability. These rovers must travel over uneven terrain, the characteristics of which are often difficult to predict. Future rovers will be deployed in even more challenging environments, such as steep craters or caves, to collect valuable scientific data. This increases the risk of rovers tipping over, especially during unmanned missions.
The proposed system uses sensors that measure the rover's acceleration and rotation speed to monitor its stability. Deep learning algorithms are used to make real-time predictions about the rover's stability. If these predictions indicate that a tip-over is imminent, the system can react in time and adjust the rover's motion to prevent it from tipping over.
The system has been validated through simulations and tests and has shown that it can significantly reduce the risk of tipping under a variety of challenging conditions. The goal is to extend rover life, optimize mission results and increase safety during planetary exploration missions.
Title: Development of a quasi-direct drive motor for walking robots in extraterrestrial environments
Authors: Jonas Eisenmenger, Zhongqian Zhao, Sven Kroffke, Frank Kirchner
This paper describes the development and testing of quasi-direct drive motors for use in space robots, particularly walking robots. These motors offer advantages due to their high motion dynamics, which could enable robots to traverse the most difficult terrain where conventional rovers would fail. Although such motors are already used in terrestrial applications, their use in space is still a relatively new field that poses special challenges. The motors must be able to withstand extreme environmental conditions and the stresses of a rocket launch.
In the Modkom project (modular components as building blocks for application-specific configurable space robots), two prototypes of quasi-direct drive motors were developed. These were tested with different motor-gearbox combinations, an in-runner and an out-runner, to compare their performance under different conditions. This included testing on an engine dynamometer, in a climatic chamber (to study the behavior under extreme temperature conditions) and vibration tests (to simulate the loads during a rocket launch).
The knowledge gained is incorporated into the design of a final version consisting of space-qualified components assembled in a clean room. This final version is then subjected to extensive environmental testing to ensure its suitability for use in space. In addition to the motors, suitable motor control electronics have also been developed, which are also tested and optimized before being integrated into the final environmental tests.
Title: Enhancing heterogeneous multi-robot teaming for planetary exploration
Authors: Amrita Suresh, Leon C. Danter, Wiebke Brinkmann
This paper deals with the development of multi-robot systems (MRS) for future space missions, in which several robots are to perform tasks together on alien planets. Previous space missions have typically used only one or two robots, some of which are remotely controlled. However, for more effective and comprehensive scientific exploration, teams of several robots working autonomously and having different capabilities are to be used in the future.
The paper emphasizes the need to consider both software and hardware aspects to optimize the performance of such robot teams. Factors such as total time, energy consumption, feedback on task performance, and potential damage all play a role. Since planetary exploration missions are often under time, communication, and energy constraints, and also face unreliable sensor data (e.g., camera images distorted by dust) and hardware wear and tear, these challenges need to be considered when developing software solutions.
The paper presents the development of a software framework that enables reliable execution of tasks in such challenging and dynamic environments. The work focuses on two main aspects: First, the integration of hardware parameters into the task assignment negotiation process, and second, the analysis of how the integration of team performance metrics, in particular adaptability and mutual support, contributes to mission success.
Title:Towards sustainable space exploration: Designing an AI-powered modular toolbox for future planetary exploration
Authors: Wiebke Brinkmann, Moritz Schilling, Jonas Eisenmenger, Jonas Benz, Jieying Li, Hilmi Dogu Kücüker, Mehmed Yüksel, Zhongqian Zhao, Priyanka Chowdhury, Heiner Peters, Leon Danter, Frank Kirchner
This article describes the MODKOM project, which aims to develop modular components for space robots. These modules should be flexibly adaptable to different requirements to allow individual solutions for space missions. The project will create a "toolbox" of reusable hardware and software components. An example of such a component is the "DFKI-X2D", a motor specifically designed for use in future space robots.
The toolbox also contains approaches for autonomous robot docking and terrain exploration. A semantic graph database forms the basis for the planning and control of missions based on these modular components.
The modules developed in the MODKOM project can also be used for terrestrial applications. An example of such an application is the "RoLand" project in the field of agricultural robotics. Here, the basic module with flexible electrical compartments was successfully used to build an agricultural robot called "SHIVAA".
Title: Pioneering Sustainable Space Ecosystems through Intelligent Robotics and Collaborative Effort
Authors: Amrita Suresh, Mehmed Yüksel, Manuel Meder, Raúl Domínguez, Wiebke Brinkmann
Humanity's long-term presence in space requires the development of sustainable space ecosystems both in orbit and on Earth. Sustainable ecosystems are characterized by, among other things, minimal resource consumption, reduction of space debris, and reusable and renewable materials and components. However, achieving sustainability in space is challenging due to limited resources, harsh environmental conditions and the need for continuous operation. This article describes how DFKI and the University of Bremen are contributing to more sustainability in space by developing AI-controlled autonomous robots.
Intelligent robotic systems with multiple manipulation and locomotion capabilities based on artificial intelligence (AI) are able to utilize local resources and perform production and maintenance tasks autonomously. Modular, reconfigurable systems and heterogeneous teams enable optimized task assignment strategies and expand the range of possible applications. Efficient methods of human-robot interaction can assist astronauts and future space residents in both routine tasks and critical missions. The article also highlights the importance of collaboration between space agencies and scientists in the fields of AI and robotics to share resources and knowledge, develop technology standards, and create interfaces for systems to work together. Such collaborative efforts are essential to ensure the long-term viability of space exploration and colonization.