The Space Symposium brings together global leaders to discuss and prepare for the future of space. NASA and others foresee a space industry with agile vehicles that can deploy materials and parts into space for robotic manufacturing and assembly. Developing Earth- and lunar-based test beds, as well as research facilities on the Moon’s surface, is a crucial step toward ISAM’s realization.
SwRI’s space robotics research is focused on high-fidelity simulation, improved perception, extraterrestrial automated driving, and robotic manipulation in space. This research is supported by SwRI’s new Space Robotics Center, which includes an air-bearing table, a seven-degree-of-freedom robot arm, a motion capture system, test fixtures, and other equipment.
Engineers created software and modeling tools to assist robots in planning movements under complicated, on-orbit settings. SwRI is also working on efficient low-power vision for lunar rovers and small aerial systems.
We are excited to share these R&D projects with the space community to help bridge the gap between today’s power-hungry, Earth-based industrial robots and the near-future ISAM ecosystem, where advanced automation will help build the next generation of space infrastructure.
DrSteve Dellenback, Vice President, Intelligent Systems Division, Southwest Research Institute
- Space-Based Robot Motion Planning- SwRI created a robotics simulation package using a physics-based modeling tool to handle difficulties such as object recognition, trajectory tracking, and dynamic motion planning in space. SwRI will test the simulation models using a robot arm on an air-bearing table at the Space Robotics Center
- Localization for Lunar Rovers- SwRI studied using its Ranger localization system on lunar rovers. Ranger’s ground-facing cameras and automation software successfully steered a rover across a simulated regolith
- Camera Vision for Cave Exploration- SwRI used caves as test beds to assess unoperated aerial systems (UAS) for potential space applications. New algorithms employed stereo cameras to guide a small UAS autonomously
- Stereovision vs. Lidar for Off-Road Autonomy- SwRI employed a recurrent neural network (RNN) algorithm to calculate vehicle motion from a series of camera images, inertial measurements, and wheel data for off-road navigation. Stereovision, an alternative to lidar, might enable simultaneous localization and mapping for extraterrestrial navigation
- FPGA Computing for Faster Object Detection- SwRI employed a space-ready field programmable gate array (FPGA) to speed up object detection. The approach used an open-source algorithm deployed on an FPGA, which outperformed a commercially available solution and conventional space computers
- RISC-V/ARM for Faster Space Computing- SwRI is testing the next generation of fast, dependable microprocessors for embedded spaceflight systems. Some space-ready FPGAs outperformed traditional processors, and an Advanced RISC Machines (ARM) processor outperformed a legacy space processor while consuming a fraction of the energy
The Intelligent Systems Division of SwRI is a pioneer in the development of systems engineering, software, cybersecurity, and artificial intelligence solutions.