- At NASA Ames, where I work right now, we've been working on a new system called Astrobee. It's a free-flying robot that we are planning to send up to the space station in November, so Astrobee, and it's actually not just a single robot.

We're actually sending three, all called Astrobee, plus a docking station, and so Astrobee's a free-flying robot. It's meant to do a variety of things inside the space station, so it won't go outside, but it will do things like do mobile sensor surveys, so if you care about like, monitoring the air quality or light levels or sound levels that you could put those sensors on the robot.

It would fly around and do these routine surveys, the kind of things that astronauts do manually today. We're also thinking about how it can be used to improve NASA's understanding of what's on the space station, and when I say what's on the space station, is we currently have a system for manually taking inventory of everything, you know, tools and food and supplies.

The astronauts have to basically, you know, check and scan bar codes or they read off numbers; A very manual process, and we think a better solution in terms of logistics is to use RFID tagging, so we've started RFID tagging things that go up to the space station, and we are equipping drawers and hatchways, and Astrobee with RFID scanners, and so now they can fly around and do inventory. And you think, you know, it sounds like sort of a menial, tedious thing, but you know, knowing where things are is really important, because you know, if you depend on something obviously you need to know where it is, but even just in terms of efficiency, you know, the astronauts have really packed days.

Every single day is really packed, and you know, when they start off an activity we say, "Well, for this activity you'll need to have this tool, which is in this drawer in this module," and if they go there and they open the drawer and it's not there, that kind of messes things up for the rest of the day. One of the things that we care a lot about from a robotics point of view, is just all the software that goes inside or is used to simulate robots.

We've been working with the Open Source Robotics Foundation, OSRF, for the past couple years, and taking some of their software that feeds into the robot operating system and tools, like Player Stage Gazebo which offers simulation to adapt those to our robot systems whether that's for simulation or for actual development.

Astrobee, for example, runs the robot operating system, and it's designed so that people can develop and then make use of that. So, for us Astrobee and the space station's not just a NASA thing; it's a community resource. It's a research platform, and so if you are familiar with Ross, you can write software for Astrobee. If you want to run a experiment on the space station, you can run an experiment on Astrobee.

I think that there's always this question: as technology becomes more advanced and pervasive of how is it used, and to the extent that robots are going to become more autonomous or other systems frankly become more autonomous, there's that big question of, how do we trust that they're going to do what we want them to do? Are they gonna work only within the bounds that we've created, and what if those bounds are fuzzy, you know?

I mean, a lot of what NASA does, at least in space, is go to places that are unknown, uncertain, undiscovered, which means by definition we don't know fully what to expect and therefore it's okay to not color just inside the lines. In fact, the lines don't exist, and so I think that if we're looking at how to treat a fully autonomous robot, especially if one's working with humans, astronauts, there are some questions we're going to have to ask ourselves as we go forward, as we allow that system to be more independent, more self reliant, more able to make it's own decisions.