Team Seven/Journal
From Maslab 2007
{{Team|number=Team Seven}
Jan 08th 2007
Knowing there is no prerequisite for this robotic class, I felt the first day in the lab was more challenging than I expected. Being in an all-architect-team, it was very fortunate that we have members that could tell whether an electronic component is polarized or not polarized, and we happen to have a member that knows how to use unix, otherwise we might have bobbled our orc board. It would be much better if more instruction is provided.
At the later part of the afternoon, looking at many completed "robots" made me impatient. We were the second last group to test our LCD display, and we fixed the solder of our pad three times before the LCD eventually gets to work.
The programming part was troublesome, we had problem setting up the class path, and without an IP address, we programed the simple hello world application right off the dry text based unix system on the bot. It is a big challenge for the architects. The moment in which the wheels turned as the program ran was the most exciting moment of the day. And, yes, we are pleased to be part of this... ...
Jan 09th 2007
Today is a very special day for our team, I short-circuited our orc board by placing the untaped board right on top of the metal case. <to be continued ... ... >
Jan 10th 2007
The pegbot recognizes red objects and attempts to follow an object with its inconsistent motors (the voltage wire had a loose connection and detaches from its port). Several debugging sessions had occurred prior to the pegbot's ability to recognize the color red, one of the main issues was when the team realized that the width and size of the input image was not created.
The robot's mechanical design is undergoing plenty of revisions, indecision occurs on whether to use a conveyor belt or elevator-like system with a cylindrical shaft to seize a ball, or two mechanical arms that can sweep the balls into its internal cavity (the motion of the arms is analogous to two hands sweeping together to attain water in a puddle).
Jan 11th 2007
Robot Design Mock-Up Plan:
SENSORS: Our design consists of two short-range IR sensors and two long-range IR sensors which are to detect and veer from any walls.
MOTORS, MECHANICS, & DESIGN: Design Option #1 Three DC Motors will be used—two for the wheels and one high torque motor for a conveyor belt with a cylindrical shaft to seize a ball. Once the cylindrical shaft seizes a ball, the ball travels along a conveyor belt that is approximately angled 20-25 degrees relative to the ground plane. The ball travels along an upward incline to travel along a channel and finally back to a downward incline so as upon reaching the physical goal, a servo motor will open a gate which would allow the balls to roll down onto the intended destination.
Design Option #2 triangular or circular –in plan- “pacman-like robot : The robot has three supports: two wheels that lay along the diameter and a third ball caster on its back. Because of the space that the motors occupy in the center (diameter) the robot has not storage space put pushes and carries with it the balls to their goal. A plate will push forward the balls and take them out of the cavity once the robot faces a goal. IR sensors can be placed along the periphery to scan for obstacles. The advantages are the simplicity of design and the fact that the robot can rotate around its center since the wheels are placed on the diameter. Thus there is no need for a servo motor to rotate the camera.
Jan 12th 2007
Today we further improved the code for ball-finding. We now have a smoother targeting and approaching of the ball. We also have 2 basic ideas for the design that must clarify until the end of the week, so that we will start fabricating the chassis. We spent most of the evening in trying to make our robot to stop in front of a found ball and not continue meandering in the playing field. Finally we did it and it looks cool. Not too bad for a team of architects! We also coded a remote controling method to manually control our robot through the computer by watching through its "eyes"(camera). Tomorrow we will have to clarify our strategy and therefore our code’s structuring so that we can start splitting coding tasks...
Jan 16th 2007
In the past four days, the team has been coding robot behaviors and collaborating about general strategies to recognize particular objects in the field. Meanwhile, the pegbot awaits its new chassis - CAD drawings and 3D models have been made as blueprints for the bot (I'm inclined to name the bot, "Archbot," in reference to the three architects involved). In fact, archbot's chassis is being lasercut as I unfold thoughts for today's journal. The team is using 1/4" plywood for the base which requires several lasercut passes to penetrate through. Some of archbot's features yet to be fabricated and installed is its rotating camera and pinball flippers.
Jan 18th 2007
Today, the Archbot found it self decorated by way too much medical tape, and that's because it's the first time she tries on her new plywood dress. Dispite the fact that we are still designing a more compact and elegant chassis which might look some what like R2D2 in the end, the Archbot was quite satisfied being in this temporary new shell. Asides from her fresh appearance, Archbot also developed some new intellegence; she is able to get to the ball, grab it and run for the goal. The Goal recognitions algorithm is greate, Mr D pulled it out from one night of work, but we decided to modify it to recognize the perspective orientation of the goal so that Archbot can charge into the goal perpendicularly. By te end of the day, we applied the wander behavior on this new shell.
Now she wanders, she chases, she grabs, and she shoots. We are almost ready for the mock contest.
Jan 22nd 2007
In the past four days, the team has been making quick prototypes as to how some of the push-ball mechanism might work. Archbot, or is it Arch2D2? spent the night at Rachelle's apartment wandering about while she (still not so sure how one of our team member's determined that the robot is female) awaits a rendition of her "plywood" wardrobe and new input behaviors.
It looks like 66% of the team will not be able to attend tonight's sponsor dinner for reasons related to other school activities. Free food and a chance to meet the sponsors isn't all that bad.
Jan. 23, 2007
Same tasks: coding, designing, machining, coding, coding...coffee...machining, designing, designing, designing...
The team does not prolong discussion on the placement of parts on the robot (as it is seemingly innate based on construction/building or professional experience by team members), but manage to discuss what hardware ought to be exposed or not (i.e. to place the battery in the robot's underbelly or on its cap?). This reminded me of why 3'-0" parapet walls exist atop buildings - to disguise the mechanical HVAC systems from public view. In other words, an architect's approach to the robot is still a piece of architecture!
Jan. 24, 2007
The team machined a connector to the motor and cylindrical spinner, a few taps broke along the way while making thread holes into the aluminum. After much discussion, however, the team decides to go along with a different approach - instead of acquiring balls with the cylindrical spinner, we'll just collect them and make sure their contained within a gate (eureka! this is the final solution to the design). When its time to deposit the balls, the robot will open its gate and tread backwards, and finally use the gate as a battering ram to hit the balls through the goal.
Tomorrow is another day, with its own tale to tell. Go team 7! (It's almost over!)
Feb. 1, 2007
Archbot's new chassis is complete after several adjustments (not to mention trips to the arch. dept. to use the laser cutter) of the friction fit joints. Nearly every electronic device is housed in a custom made/laser cut container.
It's 4:53PM and impounding starts very soon. Our code is undergoing some fine tuning - Archbot likes to wander, aim, and instead of approaching, would opt for wandering again...(10 mins. later) but wait, we found the problem! Turns out that one of the tasks wasn't set to recognize its parent.
Feb. 3, 2007
