Improving Your FLL Robot Performance

1
Improving Your FLL Robot Performance

• Presented by
• K.K. Quah President
• Technology Learning Classes for Kids

2
Outline of Presentation

• Background
• Strategies for Improving Performance
• Videos and Demos

3
Background

• Mentor 1 year and Coach for last 5 years.
• Tampa based team, Mech Masters, went from State
  FLL score of 205 in Round 1 for FLL 2003 Mars
  Mission to State FLL score of 400 in Round 1 this
  year. Also, second in Teamwork.
• Keep it simple, only 2 motors on robot. Last 3
  years have purely mechnical solutions (no
  motorized manipulator). Low cost, few pieces,
  extremely simple programs. Only Line follower in
  FLL 2005 Ocean Odyssey.

4
Full Understanding of FLL Rules

• Just like any other sports, you MUST know the
  rules. Read rules and get answers to all
  questions at any workshops, etc. Also, continue
  to check FLL website for more updates. Tedious
  but absolutely necessary. More conditions in
  rules this year and suspect years to come.
• Have seen veteran team miss the Harvest Corn
  mission but still continue to push a barrel out
  to farm. Witholding the Barrel would have cut
  -25 points to -15 points.

5
Use Reference Points

• The 4 sides of the table are useful so if you
  have guides that help a robot smoothly move along
  it, use it.
• One of best ways to make sure robot gets to the
  correct target is using the target boundaries
  itself as a guide. V-shaped/ funnel shaped
  fronts can help position robot onto the target
  correctly. Sometimes only an edge might be
  needed for guides. Example, FLL 2006 Nano Quest,
  delivery medication to bone mission. Use bone as
  guide.

6
Insufficient Engineering of Manipulators

• Some manipulators are not strong enough and come
  apart after a few trials.
• Some manipulators need more tuning to make them
  more accurate (change in angles, change in
  shapes, change in orientation, etc.)
• Some manipulators not appropriate need
  revolutionary change. What class lever is your
  manipulator?

7
Simplify Path Selections

• Part of keeping it simple is to choose the least
  complicated path. It makes programming easier.
• Straight lines are the simplest. FLL 2005 Ocean
  Odyssey, we had one program going forward and
  backwards which could solve 3 different missions
  using 3 different manipulators. Less memory for
  programs.
• Turns also affected by battery power. Li-on
  helps.
• Use table areas (walls) as intermediate points in
  path selection for complex paths.

8
Example Path with Intermediate Stop

• In above, there is no room for the robot to turn
  as the target is too close to the edge of the
  table.

9
Reverse Into Intermediate Position -1

• Start by reversing the robot onto edge of table
  wall.

10
Square Robot to Wall then turn.

• Use timer (seconds) to square to wall. Effect as
  if robot started from the wall.

11
Alignment Enhancements (Jigs)

• A lot of teams simply use dead reckoning but some
  missions require a finer control and jigs will
  help to reduce the errors in launching a robot
  out of base.
• Use hash marks on the table/base if available but
  if not make a jig which can angle a robot more
  accurately.
• Also, position jig onto the mat
• not the table as mats may not be
• exactly positioned on table.

12
Better Combinations of Missions

• Between FLL 2006 and FLL 2007, we went from 9
  missions to 13 missions. This means that combos
  are more important than ever.
• Group missions in same locale and try to solve
  for that as a grouping which reduces number of
  trips out to same area. Example, combined
  Hydrogen car, Red Truck and Wave Turbine into one
  trip and saved 20 seconds.
• Group missions for better flow from one mission
  to the next e.g. One side then the other.

13
Emulate Nature for Best Effect

• Some rules of thumbs (rules of fingers) like
  grabbers should have more than 1 finger like a
  human hand.
• When you close a door, you naturally use most of
  your palm (many fingers) not one finger. Look at
  the points of contact as a larger area of contact
  will usually work better than a smaller area of
  contact (better margin of error) so check all
  contact points.

14
Evaluate Solutions Over Many Trials

• One of the temptations to avoid is to change the
  program or manipulator after only one run/trial
  solution. This will lead to lots of changes
  without fully understanding the issues of why a
  solution did not work.
• Also, test and observe constantly. Even missions
  that are confident should always be tested to
  reveal weaknesses. Every trial run is an
  opportunity to make it better.

15
Reduction and Integration

• Part of this is to reduce unnecessary pieces. So
  review your final solutions and try to simplify
  them so fewer pieces are needed.
• Or integrate/merge two
• manipulators for the same
• mission into one manipulator
• so that the operator need only
• pick up one item.

16
Practice Choreography of Operators

• The members running the missions MUST have a
  smooth choreography. Keep working to make each
  transition between missions as smooth and as fast
  as possible. Also, they must be able to handle
  issues that are unexpected (contingencies)
• I generally say if they have not had a hundred
  runs before the competition comes around then
  they are not ready. So put a freeze on changes
  and make the choreography a priority.

17
Videos and Demos

• On table in front are some past manipulators
  which I consider really good.
• Also, here are some videos of my team in various
  missions thought the years.