 Thank you. So I'm Larry Mathews. I'm narrating the rest of the demos today. This demo is by Nate Michaels group from Carnegie Metal University. The objective here is to enable a team of robots to rapidly and as completely as possible map an unknown space. It's using an information-based criteria so it's modeling the entropy in the map and seeking motion plans that reduce that entropy as much as possible. The state-of-the-art algorithm that we're showing advancements over is a sequential planning algorithm that starts with an initial plan and then updates it for each robot in turn and iterates that until some criterion is met. That algorithm is quadratic in the number of robots so it very quickly becomes computationally and practical to scale to more robots. So the challenge is to come up with algorithms that reduce that computational complexity while preserving the quality of the map as much as possible. So the research achievement has been to develop a parallel planning algorithm where each robot can plan for itself from that initial starting point with a limited number of iterations reduces the computational complexity to linear in the number of robots. There's a graph up there that shows a simulation for up to 32 robots that illustrates that. It does. There's basically proofs about the bounds on the on the quality of the map in terms of how much mapping quality is lost and it still maps quite well. In the demonstration you're going to see a live display from the robots. They're using RGB depth sensors for perception so you're going to see initially before they start of mapping a display of the point cloud so that's the color dots. Color coding is blue is higher and I think red is lowest. As the robots start to map you'll see dark gray fill in for the map surface and then there's kind of a light gray over that which is what the robots know is unknown and I think are we ready to start the demo? If you are just go ahead. So there are three robots as you can see. It gets a little loud while it's running so you see the point clouds, you see the map starting to fill in. This demonstration is about reducing the computational complexity of the algorithm. It's not about the speed of the robot so they've not made an effort to maximize the rate of physical motion of the robot here so again the dark gray up there is the surface that's been mapped. The light gray above it is what's known to be empty space. For this demonstration are using this like this demonstration does use the vicon for position estimation for the robots. Okay are we switched back to slides? Alright so the major result is a new algorithm that scales with much better computational complexity to more robots while preserving the quality of the mapping and all this scalability is important both from a research perspective and for practicality of deploying robots for this kind of a mission. So if there's a question we can probably take time for a question otherwise I think this concludes the indoor demos and we'll be moving outdoors. No sorry. Alright next is Brian Sadler.