Turtlebot Learning

This video shows a turtlebot learning a bystander's preferred passing behavior by using the Expanded Dual Expert Algorithm. It is shot from the point of view of the turtlebot. The graph at the bottom represents the weights of the EDEA distributed across the hallway. With the red bars representing the weight for the robot predicting the bystander will pass on its left and blue bars for predicting the bystander will pass on its right.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Glider CT

The video illustrates the reconstruction of a flow field using the Glider CT algorithm by an indoor experiment where Khepera III robots mimic the horizontal motion of underwater gliders. For the experimental setup, we place a light source at the left corner of the domain and simulate a flow field such that flow is always moving in one direction towards Northeast (45 degrees) and the strength of flow depends on the light intensity. Under this simulated flow field, we navigate Khepera III robots to mimic the horizontal motion of underwater gliders. The Khepera III robots are supposed to move straight, but their trajectories are affected by the simulated flow field. For the reconstruction of the field, we collect only the initial and final positions of the robots. The x and y components (top and bottom plots on the left of the video, respectively) of the field are reconstructed only based on the collected information using the Glider CT algorithm.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Robust Cooperative Exploration with a Switching Strategy

We design a switching strategy for a group of robots to search for a local minimum of an unknown noisy scalar field. The switching strategy is implemented on a multi-robot test-bed.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Simultaneous cooperative exploration and networking

Several wheeled mobile robots with range detecting infra-red sensors are deployed in a controlled laboratory environment with the task of establishing a precise map of physical obstacles in the lab. The physical mapping of the room is created in conjunction with a topological map of the room. This map is a Voronoi diagram and consists of edges, line segments tracing the equidistant line between adjacent objects, and graph nodes or intersections, which identify the intersection of two or more Voronoi edges.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

ROV beta

Georgia Tech Systems Research had been researching and developing ROVs since late 2008. Although development was initially pushed by competitions, GTSR's aims are now to develop a versatile underwater platform for research that can also be outfitted for competitions. Undergraduate students gain design experience and practical hands-on experience while developing a platform that can be used for graduate-level research.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

 

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

 

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player