Hello everyone! Welcome to 5788’s build blog for the 2024 season! 5788 is the sister team to 5417 Eagle Robotics, and this will be our first year up and running! We will be competing in week 1 Plano and week 4 Dallas. Good luck to all teams!
Hello everyone,
As we draw close to kickoff, we wanted to share our off-season with the Open-Alliance community. Our team, Eagle Spec Ops, truly began as a way to, you guessed it, do experimental things that expand our knowledge base outside of typical driver-to-robot control. Our inspiration, swerve. It all began after we started experimenting with Swerve after we received a donation from one of our alumni, Raheel, that consisted of 4 SDS Mk3 Swerve Modules. Due to swerve being an uncharted territory, at least for our team, we initially decided to fork code from Spartatroniks, who although had incredible code, we couldn’t get to function due to it not being designed around our hardware. That’s when one of our members proposed the idea of going fully custom, from the kinematics to integrating the different systems that make a robot, a robot. That was our light bulb moment, we’ve made tremendous progress since then and we now have a fully functional swerve drive that can autonomously drive thanks to Odometry and PathPlanner. This was all done before the start of the 23-24 school year, so what did we do during the Fall semester? Well, we’re glad you asked. We competed in off-season tournaments ranging from the ones that were co-hosted by our sister team, Eagle Robotics, to hosted by one of the best teams in Texas, Texas Torque.
At Fight Like a Girl/Texas Talon Throwdown:
- We earned the nickname, Carpet Burner due to us making human-sized damage on the playing field since one of our USB cables unplugged.
- Got sub-system integration working with the Manipulator and the Elevator.
At the Remix:
- We learned that driving robots after a 4-hour drive is a bad idea.
- We were picked by the 6th Alliance and advanced until the semi-finals of the lower bracket.
What we’re working on now:
-
Getting ready to hear the Crescendo from across the hallway.
-
Integrate an external accelerator, i.e a Jetson Nano, to control our robot autonomously the entire time to reduce human errors
Hello Everyone,
Hopefully, everyone is busy and having a fun time preparing for the season. The following is our progress after Kickoff:
Build:
Our main objectives of the game:
-
Score 3 notes into the amp during the Auton Period.
-
Be able to pick up notes from the Source and the ground.
-
Shoot from the Alliance Line.
-
Climb and score into a trap while having another robot on the chain.
Our solutions:
-
Have a floor intake that’s able to score into the amp.
-
Have a shooter that can vary in height and angle.
-
Have a climber that’s attached to the shooter’s elevator to minimize motor count.
-
Have a latch mechanism such that it latches onto the chain when in the optimal position.
-
Have the shooter tilt into the trap when scoring the trap.
-
Most significantly, we have chosen to use AI to control the robot during the whole match except the endgame period because AI is our primary focus.
Timeline:
-
Have the robot’s basic structure ready before the end of January.
-
Have the AI model trained with footage from the Xbox Kinect (Our Vision for the AI Model. Check out more details about our AI below!) in addition to the internet-gathered ones before we receive the base robot by January.
-
Generate and fine-tune paths for Automatic Scoring Cycles and have the model make finite adjustments based on game-piece position within the first two weeks of February.
-
Have two weeks of drive practice for endgame and tele-op in case of mechanism failure.
Designs:
- Here is an idea that we have planned to achieve our goals mentioned above:
- Also some intake ideas that also add to our goals above:
Credits to @howlongismyname for making this amazing design public.
Programming:
-
We are excited to announce that we have incorporated AI into our team! Our fantastic crew has annotated over 2000+ images for training our custom AI model using YoloV8. We are currently working on training the AI to detect notes… Here is a small sample of our AI at work! (We have currently trained the AI with the animation video. Although it detects most of them, we are working on having the robot detect the rest):
AI_Note_Detection.mp4 - Google Drive -
We have started an OpenDataset initiative to collect game element data for AI training. If any of the other teams have collected data and would like to donate, please contact one of the admins on our Roboflow project. We would appreciate it a lot!
Notes Detector Object Detection Dataset and Pre-Trained Model by Eagle Robotics AI -
If you would like to check out our progress, here is our GitHub repository with our latest modifications to our AI model:
GitHub - frc5417/5788_Crescendo_AI: The super cool AI stuff Eagle Spec Ops does...
Hello Everyone,
Hope that your Week 2 preparations are going well! Here is a brief overview of our progress in preparing for this season:
Build:
- We revamped our CAD and finalized our designs based on the goals and objectives mentioned above. Here are our final designs:
Intake:
-
Intake Pictures:
-
Intake Design Details:
- Front Wheels: 4 Inch Compliant Wheels.
- Back Wheels: 2 Inch Compliant Wheels that are firmer.
- Width of the Intake: 25 Inches.
- Height of the Intake: 17.6 Inches.
- Extends: 6 inches from the perimeter of the frame.
- Compression: 5/8 Inch.
Shooter:
- Shooter Pictures:
We forgot to take pictures so this is the only picture we have:sob:
. We shall post pictures on Tuesday about our Shooter. Sorry about that 
.
-
Shooter Details:
- Vertical Compression: 1.5 Inches.
- Angle: 35 Degrees.
- Motor: Neo Brushless Motor.
- Wheels: 4 Inch Compliant Wheels
Robot:
-
Overall Robot CAD:
- Base of the Robot: 26 x 26 Inches.
- Drivetrain: Swerve.
- Swerve Modules: SDS MK4 L3 Modules.
- Elevator Stages: 1 Stage.
- Elevator Peak Length: 50 Inches.
- Elevator Length unextended: 34 Inches.
Videos:
-
Shooter Prototype Testing:
-
Scoring into Speaker:
-
Range Testing:
-
Programming:
- Not many developments since we have already worked on most of our necessary elements during off-season.
1 Like
Welcome Back Everyone!
Hoping that all of you have made tremendous progress on your preparations for this season, here is our Week 4 Update:
Excluding the radio, we have finished wiring the Robot base entirely! This year, one of our major focus points is wire management regarding the wiring aspect of the robot. Our Off-Season experimental swerve bot had hideous wiring, making it incredibly hard for us to debug the issues in pit.
- We are currently going with a 33 x 26 base which has a ground intake that goes through our belly pan.
- Our modules are mounted on a 26 x 26-inch frame.
- Instead of connecting the CANcoders into the CANivore individually, we connected it into the PDH CAN port instead of using the termination that it comes with. By doing this, we have ditched the CANivore which has caused us too many problems, and excluding it from our loop has caused so many issues. When using it, we couldn’t get the devices to connect to the CANivore even after trying to decode it for over 2 hours.
This is our robot after all the wiring has been done. Currently, we are using:
- NI roboRio
- Power Distribution Hub
- NEO Brushless Motor
- REV Spark Max
- CTRE CANCoders
- Mini Power Module
- Kauailabs NavX2
Along with all of the above, we also made bumpers that are 7.25 inches from the ground to the top of the bumper.
