2013 Build Season Log

[Michael Chau]

Have been pretty much too lazy to keep a build season log the last few seasons so here's an attempt to properly maintain one.

1/7/2013

Today is the first regular team meeting. We started installing the 2013 Utilities and the Driver Station on computers. Did not distribute the updates that must be downloaded from the National Instruments website for the Utilities and Driver Station. 

We installed RoboRealm onto Steven Yik's computer to begin examining the program and its feature. It is intended as a Visual Tracking program to be ran on the Driver Station during a competition and communicate to the cRIO through NetworkTables. Upon examination of the FRC 2013 registration page I initially believed only 5 computers could have the software. It may be possible that it meant 5 copies per email address. 

I plan on us starting to work on vision tracking, update the cRIO with the 2013 firmware, configure the new router, and hopefully be able to talk to Hilda about setting up a team GitHub repository. The 2013 robot code will be ideally more modular. The usage of one file to store all the variables is messy to view. Creating objects in the main class made it cluttered despite putting most functions into other classes based on subsystem. 

From poking my head in the CAD Lab robot design looks well under way. It seems we will put the highest standards on ourselves to complete this year's challenge: consistently shoot into the High Goal and climb the pyramid to the 3rd level. These are the requirements to be one of the highest performance teams.

We will need to start deciding what sensors we will need to incorporate it into the robot design and purchase it by the time we need them. We probably will use encoders on the left and right drive this year. We may use a hall effect sensor to measure the shooter RPM again. There is a need to be able to track discs in the robot to consistently move them to the shooter and lessen variability of trajectory compared to last year's system. 

43 days of Build Season left.

[Michael Chau]

1/8/2013

First 4-hour meeting. Game field elements construction are underway and should be done by Saturday. Design has decided on a 6WD West Coast at some point in the season. This year no time had to be spent debating on what we need to do for a drivetrain due to a basic need of driving on a flat field without floor pick up. The robot will align itself to a corner for the climb. Quick and accurate alignment will be the key to climbing the tower proven in Logomotion where the real trouble is deploying not climbing for all teams. I am not exactly sure what mechanism is imagined for climbing. The shooter will be able to change the angle. I imagine this is to incorporate dumping 4 discs on the pyramid goal after climbing the pyramid. Digital Media Team and Marketing are refreshing the team website and building the theme for this year's Chairman presentation.

Dinner was provided by the Mulders. It featured miniature ham and turkey sandwhiches with snicker salad and homebaked cookies. Seeing the team members line up made me realize how big the team is this year.

In programming the new Dlink router has been configured by Steven for FRC 2013 usage. It required updating to 2.03 before using the FRC configuration utility. The new image for the Classmate is unavailable still due to Intel's download server being down.

There was a minor issue with RoboRealm due to unclear instructions on their website. The instances of the software are meant to be registered under one email according to a support ticket. The team's copy is currently registered under mch...@gmail.com (mine) and the setup file is available in the FRC 2013 shared Dropbox folder. Notify someone in the shared folder to be invited.

Setup still need to be done is reflash the cRIOs with the 2013_v47 image downloaded with the FRC plugins in Netbeans,take the Axis Camera off the 2012 robot, and reimage the Classmate or install the 2013 Driverstation. There probably is a need to use a different laptop as the Driverstation in order to run RoboRealm with no CPU issues.

Vision Tracking needs to start on Thursday if possible. A tracking target of reflective tape needs to be made. This year's goal slot is too large to make a portable 1:1 scale of it. Outside dimensions are 72" by 12" We need actual robot specs anyways to tune algorithms to adjust aim and power settings due to distance and angle the camera looks at the goals. We just need to be able to apply thresholds and blob detections to find the high goal.

Autonomous will include driving from probably a back, left, and right position before timing the shots. Avoiding collisions with other robots is a concern of mine especially in eliminations. A scripting like autonomous code will help make new solutions that would be reliable.

The team has a GitHub organization ”frc-team-85” without quotes. Matt is the owner so he can invite you if you shoot him an email. I believe we are still waiting to receive the ability to make private repositories. There is a Windows GUI GitHub setup file in the frc 2013 dropbox folder. It is metro/Windows 8 slyle if you don't like the style.

The new Utilities, Driverstation, and Smartdashboard programs add needed features and stability missing from last year.

I have great hopes this year for what the team will accomplish. We have progressed further and have more inspiration than all of my previous years on the team.

That is all for today.

[Michael Chau]

1/10/13

The pyramid base frame has been completed. The feeding station and goals are still in progress.

The flywheel prototype seems to function but the usage of the rollers from 2011 and 2006? dirtied the discs. This seems to be unnecessary wear of the discs.

I have neglected checking in with other teams to record their progress today. I assume design is going along well.

The working cRIO has been updated with the new firmware. We forgot about the other cRIO having a broken port so we will need to order a new one as well. The E09 Classmate has been flashed with the new image.

Basic vision tracking code has been tested with the RoboRealm software. It seems to really offer a lot of accessibility and power in vision tracking. We need to figure out how we will use the BFR_COORDINATES array to run calculations. The main calculation we will meed is reliable distance calculation despite the angle the goal is viewed at.

The retroflective tape for 2013 is not in the team's possession. We only required the 2012 tape for today's tests but I am unsure of how different the 2013 tape is. If we can afford to I think its a benefit to test the material ourselves. We actually lack tape to actually replicate what the robot will actually see in competition so we will have to make sure that the high goal will be the only target identifiable with the other sources of retroreflective tape (middle goals, driverstation, and pyramid goal.)

The team has been granted 10 private repositories for GitHub. I pushed the skeleton project to the new repository so we can begin working on robot code with versioning control.

Dinner was Little Caesers pizzas.

The first week is coming to an end. Hopefully we will get an entire week to test and drive the robot this year. Being pushed for time really limits what our team can do on the competition field.

I say we start working on using NetworkTables tomorrow.

[Michael Chau]

Forgot to do this last night

1/11/13

All game elements are built.

Prototypes underway.

Working on NetworkTable / communication between the cRIO and the Driver Station

Can't view camera feed on SmartDashboard

[Michael Chau]

1/17/13

Imaged the cRIO FRC-II and confirmed it is able to successfullly run code.

Resolved the no connection error with the SmartDashboard Camera widget. Requires a 30fps+ camera stream to show the feed. Need to make a modified widget for a targeting reticule.

Confirmed with design team to mount a US Digital optical encoder for the drive train. Need to do the gear ratio and wheel circumference calculations to convert encoder counts to distance in a unit (inches probably)

[Michael Chau]

1/18/13

There is an update to the FRC Plugins on NetBeans that came out today so go install them. The main thing is actuators and sensors automatically appearing in Test mode on the SmartDashboard without having to specifically use the LiveWindow class to add it. At some point we should chuck the cRIO back in the 2012 robot and test LiveWindow with that.

We worked on the distance calculations and the "close" module to be less sensitive to broken objects. Defined the distance constraints of vision tracking to approximately 22 feet for maximum range. The cheap green LEDs used last year are coming apart. Instead, I am considering a green LED ring light for a more even lightning and thus more consistent detection of the target.

 http://www.superbrightleds.com/moreinfo/led-headlight-accent-lights/led-angel-eye-headlight-accent-lights/49/

Created a branch for the Shooter PID that will be merged into the master branch when there is one shooter to rule them all. Need to learn about the Feedforward, open-loop control to minimize the effect of a load (ball or frisbee being shot) on the closed-loop PID.

Other possible code that can be done before the robot is assembled is pseudo code for automatic vision tracking based on turning the robot (autonomous,) linearization output function (read FRC1114_Controls_2012.pdf included in the Dropbox folder,) and modular autonomous commands (drive x dist, shoot frisbee, wait x seconds, center on goal etc.) I made an auto branch that would appreciate you chucking things into the auto package since it made sense to separate from the rest of the code. Feel free to disagree since I am a programmer with a lack of good design knowledge.

A foreseeable issue is that we may have to readjust camera quality and calculations due to CPU or bandwidth limitations. There is a very likely chance of a shortage of time to fully test features due to the complex design of this year's robot.

Matt and I won't be here on Saturday. Feel free to think of a feature you would like on the robot and make pseudo code for it as practice. You can make a new branch and delete it later for the code if it needs to be disposed of. Not exactly sure about how the climbing mechanism would work programming wise to write any pseudo code for it. That's all I remember for today. See you all on Monday.

[Michael Chau]

End of Week 3

Decided to make the log weekly now to shorten the length and number of posts

Pseudo code for the Climber and Disc Loader has been created in the ElevatorDrive branch.

Test code for the Shooter has been created in the Shooter-PID branch. There is a TwoMotorPIDOutput class if two shooter motors are used.

Code for a refined tank drive with deadband and linearization has been created in the drive branch. 

Code for auto selector and pseudo code for autonomous modes and drive command are in the auto branch.

Distance calculations reliability improved due to lowering sensitivity of the close module in RoboRealm

[Michael Chau]

Very late but here it is

Week 4

Fixed the Victor 884 Motor Linearization to properly function with negative inputs. The Teleop drive code is pretty much complete unless we add in maximum motor output change per loop of teleopPeriodic()

Started designing the architecture of the climber programming. Will list in a separate topic.

Started developing a class for the frisbee hopper. Currently only has the servos logic programmed and none of the belts.

There has been progress in autonomous for developing timer delays on frisbee shots and drive code.

distance ratios for the climber and drivetrain encoders have been calculated to convert into inch. Need to recheck if drivetrain gear ratio used is correct.

In the later half of week 5 I have in mind a focus on getting a solid autonomous framework ready to test, hopper/shooter pseudo code ready, and a mapping of how the climber will function logically with estimated inputs/outputs.

Ideally we won't have to cut into the 6 hours before Gull Lake to create an essential feature or troubleshoot major issues and just use it to tune the robot.

[Michael Chau]

2/12/13

The robot is still coming together. Long days await on Friday, Saturday, Monday, and Tuesday to finish the robot, program it, and debug it.

Today the usage of text fields as autonomous shot times settings, a scaling crosshair overlay for the camera on SmartDashboard, and Driverstation Digital Inputs (only updates when robot is initiated) has been tested and confirmed to work.

Regarding the camera we soldered a new 1.35mm x 1.5mm power plug. There is a spare left if we ever need to use it. It has spade connectors crimped onto the lead wires of the power plug for quick disconnects allowing variability in length of wire to the Power Distribution Board.

The drive code in teleop is pretty much done and tested. Servo shifting has not been fully tested and debugged yet.

Left to be done is to program the actual autonomous, tune the shooter, combine the frisbee hopper methods for a coherent, functional hopper, and especially make sure the climber will work 100% of the time.