AI & ROBOTICS

Deep learning warm starts for trajectory optimization on the ISS

New research, published in and presented at the 2025 International Conference on Space Robotics (iSpaRo), introduces a system designed to help Astrobee, a cubeshaped, fan-powered robot, autonomously navigate the International Space Station. The ISS is a complex environment made up of interconnected modules filled with computers, storage, wiring, and experiment hardware. This makes planning safe motion for Astrobee far from trivial, said Somrita Banerjee, lead researcher who conducted this work as part of her Stanford PhD.

The traditional autonomous planning approaches that have gained traction on Earth are largely impractical for space-rated hardware. “The flight computers to run these algorithms are often more resourceconstrained than ones on terrestrial robots. Additionally, in a space environment, uncertainty, disturbances, and safety requirements are often more demanding than in terrestrial applications,” said senior author Marco Pavone, associate professor of aeronautics and astronautics in the School of Engineering and director of Stanford’s Autonomous Systems Laboratory. Despite these challenges, the team pushed the field forward with a noteworthy space research achievement. https://arxiv.org/pdf/2505.05588

Civ Robotics launches AIpowered navigation system

Civ Robotics has launched a new AIpowered material distribution system for the solar industry. CivNav speeds up solar construction by streamlining logistics, efficient pile and pallet placement, and advanced planning and workflow segmentation. To use the CivNav system, the operator mounts the sensor box on their machine of choice and hops in with the CivNav tablet device in hand. It works with existing equipment; it’s compatible with any skid steer or tele-handler, and has already been installed on Bobcat, Takeuchi, Caterpillar and John Deere machines in the field. Once mounted, CivNav guides the operator where to distribute the piles, solar modules and more. The intuitive interface enables precise placement with up to 5-cm accuracy, ensuring every plan is executed precisely.

When users combine this precise placement with GPS-enabled pile drivers, teams can skip surveying layout altogether. As the use of these GPSenabled pile drivers increases and solar projects grow in scale, the need to cut out steps like construction staking becomes even more vital. www.civrobotics.com

Advancing DARPA autonomous space surveillance technology

The U.S. Defense Advanced Research Projects Agency (DARPA) has awarded BAE Systems’ FAST Labs™ research, development, and production organization a $16 million Phase 2 contract for the Oversight program. It is focused on creating an autonomous system that keeps track, or maintains constant “custody,” of a large number of terrestrial assets via new satellite constellations.

The award follows the successful completion of the Phase 1 program. During Phase 1, BAE Systems’ software was integrated into a modeling and simulation environment to demonstrate a custody mission on representative satellite and sensor models. In Phase 2, BAE Systems will mature its solution algorithms and demonstrate operation with increasingly larger constellations, more complex scenarios, and higher fidelity modeling and simulation environments. The technology will be physically deployed to both tactical-edge satellites and ground stations.

The deployment of the software and algorithms developed on the Oversight program on board proliferated networked satellite constellations enables persistent surveillance at tactical timescales. Spacebased coordination and data processing delivers lower latency and higher revisit rates enabling near real-time tracking. The increased scale, availability, and timeliness of information derived from space enhances warfighter situational awareness and speeds up decision making. www.baesystems.com