Perhaps the most twisted branch of the rube family tree, FrankenRube is a monster to be reckoned with. With a history almost as filled in as its opponents are with dents, FrankenRube has quite a record of victories, including first place finishes at NRC 2010 and NEIRG 2011. This diamond plated steel shell holds the most ferocious three kilograms you have ever seen. Beware.
FrankenRube was designed for the National Robotics Challenge, a competition in Marion, OH, around April 14th, in 2009. FrankenRube is a mini-sumo, weighing 3KG and is 20x20x17cm. Mini-sumos compete in a 4ft diameter ring with two white lines painted around it. The objective of a mini-sumo is to push another robot out of the ring. The competition is double-elimination.
FrankenRube is the definition of a mistake.
Originally named Rube, this robot was designed by Brendan Bogan in early 2009. He was just learning about robotics and circuit design and wanted to the motors to act "all on or all off," so he designed an H-Bridge (motor controller) with relays. He also designed the lower body section, which was ORIGINALLY DESIGNED TO BE ~3" TALL, where the wheels are mounted. His father's company welded the body together.
Austin Haberly's father cut out the wheel mounts, which were more eye-balled than anything else, but they worked by tightening hose clamps around them which acted as a buffer between the clamps and motors.
The first circuit board, dubbed the "Fail Board," was soldered together by Austin Haberly, and again designed by Brendan Bogan. The board was CNC milled out of a copper plated custom circuit board by Mr. Haberly, which never ended up working. Many of the wire connections were too thin and melted away. It resulted in looking like post-explosion bomb circuitry.
By the time we decided to get around to finishing the robot, the National Robotics Challenge (NRC) 2009 was a week away. Mr. Haberly, after Austin's failed attempt, soldered together Brendan's circuit board design with ease on a PC board, which is on the robot today. This circuit board was mounted on the base, which stood an additional 3" off the base because of the relays.
Rick Brooks molded tires onto some wheels that Brendan cut out, 1" in diameter, which preformed pretty well. (Rick Brooks describes how to mold tires in Sumo Tires.doc, also available on our documents page.)
After the circuit board was on, Mr. Bogan concluded that it needed some armor casing to protect the board itself. So, he and Mr. Haberly quickly bent a large sheet of diamond plating and screwed it on to the top of the robot.
After this point, it gained its name, "Rube," coined by Mr. Haberly, after Rube Goldberg, for looking like a 5 minute conception. Keep in mind the robot with the circuit board and all was designed to stand less than 3" off the ground, but after the armor casing was added, it was more than 6" in height.
The body was done after this point, which was literally two nights before the competition, and no programming had even been thought of. But, what is programming without some detailed circuit description? Rube had three circuit boards on it: A sensor board, a motor board, and a button board, for Brendan feared that one BS2 did not have enough pins to keep up with its sensors. Rube had 5 IR sensors, 2 QTI Line sensors, and 1 Ultrasonic sensor originally. There were two IR sensors and an ultrasonic sensor on the front, one IR sensor on each side, and one IR sensor in the back. The QTI Line sensors were on either side of the front. These sensors were hooked up to a board with its own BS2. That BS2 was supposed to communicate to the motor board's BS2 through serial. Unfortunately Brendan could never get that working, so we ended up entering Rube only line sensors, which were soldered directly to the motor board. The button board's purpose was to control an LCD interface, which was driven by the motor board's BS2. We ended up using only startup routines, E.G. turn right, left, 180 deg., or go straight forward. After the startup routine, it paid attention to the line sensors, and if one saw a white line in the ring, Rube would back up and turn around.
During the NRC, literally a few hours before the match, Brendan finally started programming. He had no idea how to make the sensor board communicate with the motor board. The program was so long that the BS2 could barely hold all of it because of the LCD interface. He ended up relying on only start-up routines and line sensors, as mentioned before. During testing, the robot randomly just stopped in its tracks, a problem which was never resolved that year.
Rube preformed pretty well for being a blind and dumb robot in the NRC 2009. Ultimately, it was the line sensors that defeated Rube. Rube would "see" the shiny blades/scoops on other people's robots, mistake them for white lines, and back up and turn around. It just plain backed out of the ring twice. Though there were few competitors who could match us in power, our own strategy defeated us. We did not place in the top 3 that year.
After that year, Brendan gave up on Rube, and even refused to keep it at his house. He ended up starting over and designing a completely new robot, named "King Rube" that year and later called "The Tumbler," both coined by Austin Haberly. It was more thought out than Rube, but the results were almost the same as Frankenrube's in the NRC 2010, where he didn't place and the robot was completely bind.
Austin took Rube in, after being abandoned as a failed project by Brendan. It was literally a pile of parts, half-scrapped by early 2010. Austin assembled it and stripped FrankenRube of the sonar, three of the five IR sensors, the QTI Line sensors, the LCD, the button board, and the sensor board, then interfaced everything to the motor board. After resurrecting Rube, Austin renamed Rube as "FrankenRube," for being brought back from the dead. The day before NEIRG, (NEIRG stands for North East Indiana Robot Games, an annual competition held at Science Central. To learn more about NEIRG, please visit the official NEIRG webpage or the "About the Competitions" section of our about page) Austin put it back together piece by piece and Brendan found a bad transistor in it. It was repaired and Austin started programming the day of NEIRG. The program was written in about 30 minutes, which resulted it in being even less advanced than the 2009 program. No sensors were used, because of time constraints. The robot was only programmed to turn 90 degrees right, to turn 90 degrees left, to turn 180 degrees, or to go straight.
This version of FrankenRube ended up not placing again, but FrankenRube's story has just begun.
Austin started working on FrankenRube for the NRC later that year, and started first by gluing a third IR sensor onto the middle of the front of FrankenRube. Austin also riveted a bumper onto the front of FrankenRube. He wrote completely new code, and implemented only one button, which was interfaced only with the motorboard BS2. FrankenRube would start only if both the power switch and this button were on. FrankenRube was programmed to spin in a circle until the IR sensors saw something, and then charge for that object. The programming this time was finished prior to the 2010 NRC, and was fully tested before the competition. Finally FrankenRube was almost to its potential. However, the robot still randomly stopped in its tracks occasionally.
Austin managed to win 1st place gold in the mini-sumo high school division of NRC with FrankenRube in 2010. Brendan's new robot, "The Tumbler," preformed worse than Rube did the first year it was entered, because of more resetting problems and clock issues.
For the most part, FrankenRube wasn't touched between 2010 and 2011. Austin brought it back out of its box and tested it, where he found a similar transistor problem as before, which was an easy fix. Austin also tore off the front bumper and riveted a scoop to the front, to lift other robots off the ground. The programming was completely overhauled again, making the robot much more efficient than it was before. The cause of the random resetting and stopping was found and fixed. This was being caused by voltage spikes and noise coming from the motors, so Austin screwed some caps across the motor contacts, and implemented a few tricks in the code to resume wherever it left off if any resetting did occur.
FrankenRube was ready for NEIRG 2011 and NRC 2011 long before either competition. Austin ended up winning 1st place in High School Mini-Sumo at NEIRG with Frankenrube and 2nd place in Post-Secondary Mini-Sumo at NEIRG 2011. The competition was close between FrankenRube and Brendan's new version of "The Tumbler," referred to as "Rube Mark III." The Tumbler ended up destroying FrankenRube's two of the three front IR sensors, so FrankenRube became blind once again. However, Brendan's "The Tumbler" ended up backing out of the ring or simply stopping repeatedly, and FrankenRube won.
FrankenRube was merely repaired for the NRC 2011 a few months after NEIRG 2011. This year, Indiana Tech Explorer Post had 4 entries into the Mini-Sumo category, and we came back with 1st place in the post secondary division, and 2nd, 3rd, and 4th place in the high school division. It brought home bronze in the high school division that year, behind Marie and Weilin's sumo, which won silver.
FrankenRube has won 4 awards out of the 5 competitions entered so far.