From RobotLab
We are interested in the mechanics of locomotion and manipulation -- the interface between robots and the physical world. We have studied exact optimal trajectories for vehicles, robot origami folding, knot-tying, grasping cloth, and robust minimal-sensing control of rigid-body contact tasks. The PI is Devin Balkcom.
Projects
|
|
| How hard is knot-tying, and what are the limitations of fixturing? We have built fixtures such that string pushed through the fixture (by hand or pressurized air) is tied into a knot.
|
|
Protein structure from x-ray scattering
|
|
We have studied one component of the problem of elucidating three-dimensional structure from solution scattering data, determining the distribution of interatomic distances, the P(r) curve. We propose a set of linear constraints to precisely describe the space of all consistent P(r) curves, as a polytope of histogram values or Fourier coefficients.
|
|
Geodesics for rigid planar robots
|
| We characterize the shortest paths for a model of rigid bodies in the plane. These results include the shortest paths for the well-known Dubins and Reeds-Shepp cars as special cases, and also apply to a model of a differential-drive robot, to parts being stably pushed by a robot arm, and to several varieties of omni-directional mobile robots. We have also developed a Javascript simulator for generic trajectories for several of these vehicles.
|
|
Grasping Cloth
|
|
| Pulling on the two ends of a string immobilizes the string. Pulling on the four corners of a square piece of cloth immobilizes the cloth. Pulling on the five points of a star immobilizes the star. For all these examples, it is sufficient to just grasp the convex vertices. Is this true in general? We have shown that there exist polygons for which up to one third of concave vertices need to be grasped (example shown at left), but that there are no polygons for which more than one third of the concave vertices must be grasped.
|
|
|
Robotic origami folding
|
| We built a robot capable of folding a paper hat, paper airplane, and paper cup. We have also analyzed more complicated folding techniques; for example, our work with Erik and Martin Demaine questions whether ordinary paper shopping bags can be mathematically folded.
|
|
Rigid-body manipulation
|
| We designed an algorithm for computing all possible forces and torques consistent with constraints on the motion of rigid bodies, and used the algorithm to compute parts-seating plans that are robust to incomplete or incorrect sensor information and to frictional indeterminacy.
|
|
Undergraduate projects
Toy climbing robot
At the request of the central New Hampshire SWAT team, Bell, Krishnan, and undergraduate Hoy built wheeled and tracked robots for remote surveillance. The robots met with great enthusiasm, but currently are in need of some repair and possibly redesign. Contact Balkcom if you're an undergraduate at Dartmouth and would like to work on this project.
Undergraduate Matt Elwin won the "class of 1959 prize", for his work on folding cartons with fixtures. Bell, while an undergraduate, won the C.S. department's Kemeny prize for his work on a toy rock-climbing robot, and published a short paper and poster about this robot at ICRA 2006. Rosa, Pyke, and Kevin Olds worked on the problem of freeing mobile robots stuck in sand, which led to a short paper/poster at the IEEE ICRA conference.
