Our robot was successful in navigating the real world (pictured above). The middle picture captures a critical moment in the final competition. We searched a four story building looking for the lightning bolt, which demarcated an area we needed to report as dangerous. We  (the tele-operators) did not observe that the world extended to our left as we exited that portion of the floor. So close!

Notable Labs: 

(1) Deduced Reckoning - Know position using encoder values from a differential ​drive robot.

(2) Trajectory Following - Provided set of parametric equations, robot can follow a curve on cartesian grid.

Mason's Valentine's Day Gift (4V)

(3) Wheel Free - Design robot to move without use of rotary wheels. (Our design pictured at right)

(4) Computer Vision: Use filtering, segmentation, and geometry calculations from camera optics to segment and locate objects in camera images.

     Pipeline for tennis ball segmentation:

          (a) original image

          (b) filtering based on hsv and rgb values

          (c) segmentation using double raster

          (d) radial restriction filter (round off loose edges)

(5) Localization: Robot is placed in random position with tennis balls set on course. Robot must determine position and orientation using its knowledge of tennis ball locations.

    Representation on the right displays computational simulation of the robot building understanding of its environment by navigating the field and applying a Bayesian filter:

          (a) Starts by assigning equal probability to each                     board space 

          (b) After a few steps determines it is most likely                   in one of three general places.

          (c) Converges to known position and orientation.

Capstone Lab: USAR - Urban Search and Rescue

Objective: Search a 4 story building (scaled to robot's size) deemed too unsafe for humans. The environment was filled with rubble and floor cavities, as it was meant to be a post catastrophe rescue mission. A complete search of the building required tele-operated navigation of the building, including ascent and descent of stairs and ramps.

Robot (pictured below): Our strategy was to implement the most stable design for stair and rubble navigation. The tank design allowed us to obtain a low center of mass and the treads allowed for effective handling of the rubble. As was required, we equipped our robot with an LED circuit and camera for tele-operation. There was a competition to see who could most efficiently search the 4 story building.  Our team made it to the championship round! (and lost :{ ) It was an enjoyable way to conclude the class.

Sean and Audrey, awesome partners!
(a) Initial State
(Wheel free robot, inspired by RHex.)

Introduction to Robotics

Team: Sean Bittner, Audrey Yeoh   -  Robot Name: Mason

Labs: Weekly demos requiring implementation of breadth of robotics concepts.

​(c) Convergence

(b) Gaining Context