FEH RObot 2014


Topics for today

  • The team                                                                         
  • Early goals 
  • Chassis development 
  • Tools 
  • Sensors
  • Code and Route 
  • Budget
  • Results 
  • Future changes 

The team

  • Carter Hurd                                                          
    • Chassis design and construction
  • David Hayes
    • Sensor setup and electronics
  • Hopen Yang
    • Documentation 
  • Ryan Niemocienski
    • Route and programming 

early goals

  • Low number of servos and powered mechanisms
  • Reliability without reliance on RPS
  • Fast run time + efficient path 
  • Separate tools for tasks 
  • A robot that works even when positioned imprecisely

Omnidirectional platform

Pros: 
  •      Can complete the course fast without spending time turning
  •      Can mount each tool to one side
  •      Less chance of losing orientation on course due to turns 

Cons: 
  •      Omni wheels can't go up ramps well
  •     Expensive to buy/manufacture omni wheels
  •      Can’t effectively use encoders 

Chassis Development




Version 1

  •     Used 3 gears to reduce cost (prototype had 4 gears)
  •     Had ability to put all wheels at 45 degree angles for turning in one spot

Version 2 


  •     Simpler, more reliable mechanism with no gears
  •     No ability to put wheels at 45 degree angles-
    •     Not needed for our navigation plan

Version 3 

  • Bands held wheel housings tightly against chassis for driving straight 
  • Failed to change orientations smoothly 

version 4

  • Version 2 linkage setup with flexible hub
  • Allows for smooth transitions hile still holding wheel housings against side of chassis

Other details 

  • 2 sheets of PVC for body enclosed delicate linkages and minimize flex
  • Ball bearings to reduce friction when changing orientation 
  • Low, centered weight distribution 
  • Suspension to keep the 2 diagonally-oriented drive wheels on the ground when acceding ramp
  • Enclosed wheel housings to support wheel axles

Tools




Powered tools 

  • Micro servo with short lever arm for button
    • Mounted close to button for maximum leverage
    • Mechanism can be lined up at start of match                       

 


  • Forklift to lift skid and flip switch 
    • Lifting skid allows for accurate navigation                        
    • Bent forklift arms 
    • Servo mounted below chassis
      • 2:3 ratio improved torque of forklift 

Passive tools

  • Wire clips secure pin
    • Clips are hinged to rotate downward after getting pin

  • Scoop dump
    • Throws scoop when DROID runs into wall                          

Sensors





  • Rotary encoders 
    • More consistent then optical sensor methods                       
    • encoders embedded within drive wheels 

  • 3 CDS Cells 
    • Greatly increase effective light reading area
    • Foil cones on 2 sensors to further increase reading area

Code + Route


key functions

  • Move_clicks ()
    • Drives the motor a certain amount of clicks, testing a given condition as the robot drives. Uses a Proportional Integral Derivative (PID) controller.
  •     Orient_X ()
    •     Orients robot to drive in  X direction
  •     Orient_Y ()
    •     Orients robot to drive in Y direction 
  •     Fix_heading ()
    •    Fixes the heading of the robot to a setpoint with a given tolerance according to RPS values

PID control

Our PID controller works on the errors of shaft encoding. Based on the previous, current, and predicted future error from the shaft encoders, the robot correctly sets the motor powers to drive straight.


  • Proportional: The factor of the current error between expected and actual number of clicks 
  • Integral: Sum of all previous error
  • Derivative: Error between previous error and current error

Sample PID code

Float proportion = error1 * KP;

Float derivative = (prev_error1 – error1) * KD;

Float integral = total_error1 * KI;

Float change = proportion + derivative + integral;

motor1power += change;


  • Orient functions
    • As servo changes direction of the wheels, the body rotates
    • Motors rotate in opposite direction at the same time as orientation change, to keep heading constant 
      • Speed and timing must be exact 
  • Fix heading 
    • Moves towards desired heading until reached 
    • Desired heading based off of start heading 
    • Zero values discarded 

cost saving measures

  • $124 total spent 
    • Despite $20  in unused purchases/replacement parts
  • Servos purchased for <$5 online
  • Micro servo used where little force was required 
  • Body made entirely out of 2 sheets of PVC
  • Shape-lock plastic used as alternative to 3D printing 
  • Prototype parts were mocked up and tested before added to the budget 

Results

  • Low number of servos and powered mechanisms
    • No servos used to retrieve pin, flip switch, or deposit scoop
  • Reliability without reliance on RPS
    • Path used just 2 heading checks, and worked reliably
  • Fast run time + efficient path 
    • Fastest perfect run during the competition 
  • Separate tools for tasks 
    • Omnidirectional platform allowed for this 
  • A robot that worked even when positioned imprecisely
    • Bent forklift arms + 3 pin clips 

Future Changes

  • Make flexible center linkage hub out of rubber urethane 
    • Would stand more repeated orientation changes 
  • Modify chassis to accept VEX 393 motors 
  • RPS free heading checks?  Remotely pressing charge zone?

 

the end

FEH 2014

By Carter Hurd

FEH 2014

  • 1,585