50 Pounds of Robot Love

Introduction Sensor Strategies Mechanical Design Electrical Design Software Design Gems of Wisdom About the Team

Mechanical Design Frontbot Design We recognized early on that space constraints were going to be a major factor in our mechanical design. Since we planned to fit our robot on top of Goal 3 and stack balls off of the ground, there needed to be a 4x9x6 area of dead space at the front of our robot. Furthermore, this restriction also meant that we could not place any tape sensors at the front of the robot, like we had in Lab 8. To achieve line following, we used four tape sensors aligned along the axis of the drive motors. We divided the base 12x12x11.5 space into two equal sections, so the front half needed to fit in a relatively small space, be able to drive and sense autonomously and envelop Goal 3. A couple of iterations on these constraints led to the design modeled to the left. The base of the robot extends 2 in. beyond the 6 in. width to contain our latching mechanism, described below. The motor mounting blocks were made of acrylic to provide a more firm attachment for the motors, which were coupled directly to roller blade wheels using spider couplers. On the front half of the robot, all available space was used to mount the C32 and E128 microcontrollers and circuitry for the four tape sensors, beacon sensors, drive motors as well as power circuitry. Two batteries—one for the motors and one for logic circuitry—were mounted using Velcro to the underside of the top plate. Other features include a power switch on the roof of the robot, two casters mounted at the front corners for stability, a slight funnel at the front to help guide the robot over Goal 3 and two steel weights, also at the front, for added mass and stability when extending the tape measure. Backbot Design The rear half of the robot consisted of: - the latching mechanism, and the motor to drive it - a brake and weights, such that the force of the bridge being pulled out wouldn’t move or topple the backbot - a turret, such that the bridge could swivel - a funnel, such that the dispensed balls would find their way onto the bridge - a bridge “tensioner,” such that the tape measure maintained sufficient rigidity - tape sensors, to aid in the navigation prior to separation. - A “whacker,” a DC motor with attached arm to depress the ball dispenser button While attached to the front half, the backbot was essentially a chassis extension. The base level was even with that of the front half of the bot, and contained two ball casters to allow the front half to drive the back smoothly. Latch Our overall strategy required that the robot separate into two sections at the ball dispenser—the rear half would stay at the dispenser and the front half—which has the drive mechanism—would move to Goal 3. We needed to design a latch mechanism that would allow the frontbot to push and pull the backbit to the dispenser and then detach. The mechanism we chose essentially a driven screw. An aluminum plate in the front half of the bot was tapped for ½”-13 threads. Mounted directly to a DC gearmotor on the backbot was a corresponding screw. By driving the motor such that it turned clockwise, we could screw the two halves of the bot together. By driving it in the opposite direction, the screw would force the front half (slightly) forward and disengage. This type of latching scheme was chosen for several reasons. It can be relied on to function despite significant loading in almost every direction. A keyed latch of another sort would’ve been subject to binding under the loads experienced by this robot. In addition, the latch provided secure attachment from a single point. Finally, such a scheme was incredibly simple to actuate. Brake The backbot was actually constructed on a slant. While the bottom level was indeed parallel to the ground (and to the base level of the front bot), the upper two levels of the back bot were built at a 5° incline (from the rear, the top surface of the bot angled 5° upward). When the latch disengaged, the front half of the backbot would fall those 5° onto a long rubber brake. A strip of gasket rubber 1/8” thick and approximately ½” wide runs the width of backbot, and provides friction once the weight of the backbot is resting on top of it. This prevents the frontbot from moving the backbot as it pulls the FatMax tape measure forward. Turret The tape measure needed to swing smoothly and evenly to prevent tipping on both frontbot and backbot sides. To allow this, we constructed a turret, a level of the backbot that rotated with respect to the lower levels. The third level of the backbot had a flanged plain bearing embedded in it. A precision ground case-hardened steel shaft ran through this. To ensure alignment, a second bearing was placed into an acrylic plate, which was then bolted onto the second level of the backbot. At the top of the shaft, a lasercut acrylic piece set screwed to the shaft. The turret level then bolted on this set screw piece. The result was that the shaft spun smoothly in the two bearings, carrying the turret level with it. Bridge Tensioner We noticed early on in experimenting with our tape measure that a small upward force exerted at the base of the extended portion of the tape made a great difference in the tape’s ability to resist torsion. We lasercut a series of masonite uprights with holes at different heights. More precision ground shaft was put through these holes, and with the variety of heights available, we were able to tune the tension on the tape.