Team One/Journal
.
Day 0 (1/6/2013)
It has been eight days since the Stuxnet outbreak. What we knew before the incident is that Stuxnet was a first-gen hardware virus aim at destroying parts of Iran's nuclear infrastructure. We've gathered through word of mouth that the virus has mutated and spread through majority of the world's nuclear infrastructure. Majority of the world sits in devastation from systematic meltdowns of nuclear power plants and freak self-detonations of ICBMs.
"Thankfully", MIT's reactor was only a small test reactor and only campus was destroyed.
Rodrigo, Dan, Nanu, and I (Tyler) were at Poor House at the time of the blast and hence survived. We've taken up refugee at Edward's apartment at the edge of the blast radius. However, food, water, and other amenities has become scarcer. There has been riots at the Shaws in Boston. As a result, we've looked north to the remains of campus for food. Rumor has it that Professor Donald Sadoway had a large stash of Twinkies on campus because of the 12/21/2012 scare. Though, now it looks more like the Mayans were right.
If the Twinkies exist, this would be huge. Everyone knows that Twinkies are the only food that can survive an apocalypse. We would not only be able to eat, we could trade it for further goods.
.
Day 1 (1/7/2013)
We had a meeting this morning. Campus is heavily covered in radiation and hence traveling back there would definitely result in long term harm. As a result, we've decided to build a robot to travel to campus for us.
Codenamed PEGbot, this robot is suppose to travel to campus, search it for Twinkies, and report back the location of the Twinkies.
In building PEGbot, the following was accomplished.
Mechanical Design
(1) Center Wheel Drive - allows PEGbot to turn with a zero turn radius
- a caster wheel was placed in the back center to provide a three point base
- the battery over the caster to back weigh it so it doesn't fall forward
(2) Secondary Platform - provides PEGbot with a second level enabling it to carry a laptop
** see blueprints below
Communication
(1) Library of functions and classes complied into the Arduino (2) Laptop sends commands to the serial port in the Arduino to control PEGbot (3) Can control PEG bot with Andriod phone.
Checkpoint 1
At the end of the day, we attempted to drive PEGbot around Edward's house. Using a phone, we were able to successfully navigate PEGbot from from the kitchen to the living room.
.
Day 2 (1/8/2013)
Rodrigo, Dan, and I awoke this morning to find Nanu missing without a trace. Hopefully he is okay.
We added a IR sensor to PEGbot and began coding to turn when it sees a wall. Once we have this basic function down, we can add multiple sensors so PEGbot can sense his environment and safely navigate itself to the Twinkies.
UPDATE 1
Nanu has safely returned. Apparently, he couldn't sleep much. Noises outside woke him up at 5 am and he went to investigate. He said the survivors are attempting to set up some sort government in what remains in Boston. It looks like 3 groups are emerging. There are the Bostonian Builders, mostly survivors from MIT, Harvard, and Boston University. They are mostly a technology/engineering oriented group that appear in control of the northern end of Boston and Cambridge. To the east along the water front lies the Charlestown Confederates, mostly naval officers and other sailors. They appear militant and interested in controlling the water travel and trade. Then to the south and west lies the Brothers in Bond. They are a mostly religious group believing that the event 10 days ago was the end of the world they are were the chosen to survive. They appear very aggressive in spreading their message. Apparently, the three groups are struggling to control the city. I'm just happy our HQ is located in the Bostonian Builder's territory. They seem the most reasonable.
UPDATE 2
The four of us had a meeting. Concerns were raised about PEGbot getting caught in skirmishes or conflicts between the three rising groups. As a result, we've decided to construct a second robot codenamed "Tallahassee." "Tallahassee" we be an upgraded heavily armored and armed version of PEGbot in order to protect itself as it brings Twinkies back to us. Dan and I began CADDing Tallahassee. We started with the basics. We CADDed wheels, gears we and a base. In addition, we designed Tallahassee's gearbox. Rodrigo and Dan worked also on Tallahassee's motor control. Rorigo and Nanu worked on Tallahassee's vision system and processing.
Checkpoint 2
PEGbot is successfully turning away from the wall when its IR sensor picks up a wall.
Tallahassee
Mechanical Design
(1) Center Wheel Drive - allows Tallahassee to turn with a zero turn radius
(2) Dan's Pittman Motors - rpm: 340 rpm/volt
- built-in encoders: 4000 counts per resolution
(3) Gearbox - provides 13.5 gear reduction for the motors.
- top speed: 1.551 m/s
- encoder resolution after gearing: 5.4 * 10^4 counters per resolution
** see CAD model below
.
Day 3 (1/9/2013)
I woke up in my sleep. It's roughly 5 am and I saw Nanu working at his desk. I assumed the noises outside have been keeping him up. I'll try to go back to sleep.
UPDATE 1
This morning we found out that Nanu has gotten decently far on the vision system. We have the basics color detection down. Our camera can detect green versus red balls.
After that, I worked on designing a helical ramp the balls could be pushed up to pick them up. In the center of the ramp would be a spindle with rods sticking out of it. As the spindle would turn, the rods would push the balls up the ramp. I had little success with this. However, we have figure out another way to pick the balls up. Dan, Rodrigo, and Nanu spent the day working on vision.
Checkpoint 3
We present Tallahassee to Edward and discusses our agenda to finish by the end of the month. Note to self, our agenda can be found under assignments in the back of my journal.
UPDATE 2
Dan and I began to build the gearboxes. Here is how they turned out. I taped a picture in my journal of it.
.
Day 4 (1/10/2013)
I found a journal today. It has T-9 written on the front of it. Not really sure what it means. They said they stole someone's webcam. That must suck. At least I know it wasn't ours since Rodrigo brought his back to his room last night while he worked on it.
Day 5 (1/11/2013)
Gearbox finalized and wheels attached.
Day 6 (1/12/2013)
CAD model further updated. Zero lag in vision and ball detection improved.
More extensive journal entries will be written in the morning on Sunday for days 4, 5 and 6.
Day 7 (1/13/2013)
Robot was further CADed. Prodotype of Tallahassee was laser cut.
Day 8 (1/14/2013)
Mock Competition. We took over 9000 photos.
Day 9 (1/15/2013)
Ball Sweeper fully designed. Nanu has almost finished vision. It can detect the tops of the walls now. Rodrigo and Dan worked on navigation, in particular PWM to have our robot drive straight.
Day 10 (1/16/2013)
Robot drives straight, waypoint navigation works, ball pick up and mapping have started.
Day 11 (1/17/2013)
Color vision finalized, distance sensors tested and analog output to inches conversion made. Design of the distance sensor array started.
Day 12 (1/18/2013)
Mock Competition: we gave random waypoints to our robot, and had it stop going when it saw a wall. It worked well. We identified an issue with our design: certain parts were too close to the ground, which made it get stuck in bumps. We also forgot to account for the current sensor (assumed the motor controller board had logic itself to take care of that), and that led to frying part of it when a motor got stuck and demanded too much current (we will need a new motor controller board, but will use this one to test current sensor that we will use).
Day 13 (1/19/2013)
Mapping happened :) We also have a neat viewer for our robot now (6.01's SOAR style). Also used the prototype board to make everything neater (no more jumper cables :D). Tried to calibrate all the distance sensors, and while some success was attained, we found out some of them seem to be damaged or just not good, since they have a very small range (one of them has about half the range of the other ones). We will calibrate them better tomorrow and decide on whether to replace the bad one or not. Mapping was successful though (it successfully mapped a box it drove around).
Day 14 (1/20/2013)
The day is still going: today we are going to buy materials for building, build a new prototype (hopefully final, except for what we aren't building today) of the robot, calibrate the distance sensors, and if time allows: test motion planning and identifying ball positions. We will also try fixing the issue we're having with interrupts by investigating interrupt counters and trying to use a separate Arduino to do that, and decide on what seems better. We will also try to work on the arduino code to include current sensing, but that might be left for tomorrow.
Day 15 (1/21/2013)
Today was a rough day. We lost two pins on the Arduino, had trouble with our waypoint navigator, and busted a motor. Despite this, we are still moving forward. We figured out a couple ways to increase our IR sensor's range and tested a proof of concept on how to build the archenemies screw.
