NASA’s Astrobee, a free-flying robotic system for transporting cargo and assisting astronauts, is on an exciting path to development. Astrobee is made up of three identical, free-flying, cube-shaped robots. They were meant to benefit astronauts on board the International Space Station (ISS) by taking over a mention of repetitive duties. Recent advances focus on improving robots’ skills to handle high-flexibility, cargo-style bags with vinyl based materials. Though important, this challenge has historically hampered their impact.
Daniel Morton, the research team’s wildland-urban interface simulation environment, or WISE. Written and conceptualized by NASA’s Strategy and Partnerships Development Office, this creative tool aims to expand Astrobee’s cargo-carrying possibilities. This groundbreaking work, titled “Deformable Cargo Transport in Microgravity with Astrobee,” is now available on arXiv, providing insights into how these robots can adapt to the complexities of the ISS environment.
Developing the Simulation Environment
This new simulation environment, called Pyastrobee, combines reinforcement learning (RL) software that includes Gymnasium and Stable Baselines. This cutting-edge simulator not only simulates the true to life challenges encountered in space, but specifically caters to deformable cargo.
Morton explained, “We set out to find a way to control this and built a simulation environment that can not only accurately represent the ISS, but model deformable cargo.” With the open-source physics engine Bullet, Unreal can accurately model cargo behavior in transit as well. This helps us produce the best quality simulations possible.
This development is a big expansion for Astrobee’s capabilities. It continues to empower engineers and students alike to more efficiently and effectively test their space robotics algorithms. Morton expressed optimism about the potential applications of Pyastrobee, noting its unique capabilities in simulating deformable cargo challenges that previous models did not address.
Addressing Challenges in Cargo Transport
Soft cargo is a unique challenge, especially when considering the perishability of material. Morton was clear that we must deal with these problems. He explained that it’s becoming increasingly important for future space stations—which may not always keep a crew onboard—to have the ability to rely on robots to conduct autonomous tasks such as replenishing cargo. Having Astrobee be able to grasp these cargo bags and navigate around them is an extremely challenging task!
Due to the highly dynamic nature of interacting with deformable materials, sophisticated control techniques are required. The research team found that their simulator-in-the-loop sampling-based model-predictive-controller (MPC) was an exciting first step solution. Morton noted the need for improved computational efficiency as they progress: “Now that we have a preliminary approach (sampling-based MPC), I’d like to explore how to make this much more computationally efficient.”
Future Directions for Astrobee
Morton and his team are interested in finding out how multiple Astrobees can operate in tandem. Their purpose is to optimize load stability while transporting cargo. In his concept, he suggested using two Astrobees to latch onto either side of a cargo bag. This method would provide significant improvement to interaction and manipulation.
In addition to exploring perception George and I have been developing very efficient safety filters for robot manipulation. This kind of technology would be just ideal for the Astrobee. He stated, “To give guarantees on constraints like collision avoidance, I’d like to explore how to use multiple Astrobees to perform this task—two Astrobees holding either side of the bag will likely improve stability during transport.”
Taken together, these discoveries underscore NASA’s commitment to developing autonomous systems. These systems will be critical for future missions, particularly in environments inhospitable to human presence or human access. Combining these advanced simulation technologies with Astrobee’s current framework has proven to be a perfect recipe for innovation within the space robotics community.