Cbot - Plant Maintenance Robot 1
Preliminary Development Projects
Updated 11/8/15
Key Search Words: ROBOT, ROBOTICS, ROBOTIC VISION, ARTIFICIAL INTELLIGENCE, AI

Final Report 

 

      
CBOT-1: Plant Watering
      Robot

      
Final Report

      
Certainly one of
      the most complex robots I have ever built, this robot was not
      only the first to use the highly advanced CCS-C compiler for
      writing the code which controlled the robot, and employed navigational
      techniques that had not been used before in any of my previous
      robots.  As many of you know, the robots construction and
      programming was done incrementally, each step building on the
      last to bring together the complete robot. Here you will find
      a overview of worked and what did not, and some thoughts on the
      future next generation of Plant Maintenance Robots.

      
What worked really
      well

      
CHARGING CONTACTS
      AND DOCKING PLATES

      - The brass contact plates combined with the guitar string feelers
      made a very effective and reliable charging contact system.

      
BEACON HOMING

      - Set for the lowest power level, the robot could see the charger
      over 4 feet away and almost always nailed it right in the center!

      
AUTO FILLING PUMPS
      AND CIRCUITS

      - Using a 1 khz water pump ON signal worked great, and did not
      have any interference from sunlight or overhead lighting.

      
WATERING TUBE AND
      FUNNEL

      - The idea of watering a single funnel with distribution tubes
      rather than try to hit every plant worked perfectly.  The
      watering tube was always within 1 inch of the center of the funnel
      when the water turned on.  

      
SONAR DISTANCE
      NAVIGATION

      - Set for one inch accuracy, this was good enough to hit the
      funnel every time.  twice that good would have been even
      better.

      
WATERING ON TIME
      SELECTOR

      - This worked so much more reliably than the color computer I
      tried at first, which blew up before it ever made it into the
      robot.  I finally set the selector for 4 seconds for best
      results.

      
ADAFRUIT DIGITAL
      MAGNETOMETER

      - This is the biggest HOG of program space in the robots memory,
      but never got lost when used as an indoor compass.  

      
WATERING TRUTH
      CUP

      - This small cap was suspended inside the watering funnel on
      the planter and filled with water each day IF the robot watered
      the plants.  Id dump it every evening, and this let me know
      when I wasnt around to see it that the plants DID get watered.

      
What did NOT work
      very well

      
ORIGINAL DRIVE
      MOTORS

      - Awful cheap ($22) Chinese motors that were a bit too fast anyway.
      One kept on slowing down every day for months and the other one
      day completely siezed up!

      
AUDIO SOUNDS SYSTEM

      - Poor robot had a hard time making sounds with such a tiny speaker. 
      Oh well.

      
USING DUAL GEAR
      HEAD MOTORS FOR DIFFERENTIAL DRIVE

      - Nearly all  house hold robots use DC gear motors to drive
      them about.  No matter how I adjusted them, the constantly
      changing water load made one motor always go slower than the
      other, making the robot bank right or left and sending off course.

      
WATER LEVEL INDICATORS

      - The first one I made worked for a week and then got coated
      with lime and always read incorrect water levels.  The second
      one, now in use is better but sometimes still reads "phantom"
      water levels.

      
BACKLIGHT ON LCD
      DISPLAY

      - After a while the New Haven transrelfective display would not
      turn off its backlight.  Really sucked power at night! 

      
RAN OUT OF PROGRAM
      MEMORY

      - A HUGE problem, I almost could not make the robot work with
      the very limited 8k of program memory.  The big problem
      was the digital compass had to have extensive math libraries
      tacked on to make it work, and left me little for the robots
      movements.  A separate compass chip will be in the next
      version.

      
DOCKING BACKUKP
      FLAG

      - In theory, a small flag would fall when the robot backed up
      to let me see on the web cam for the room that the robot left
      the charger for the day.  Sometimes it didnt fall down.

      
Future changes
      for the next version of CBOT.

      
The next upgrade
      will occur after the plants have been harvested for peas and
      radishes.  The primary goal is not only to free up all the
      processor memory by offloading sensor data to separate processors,
      but to improve the accuracy in reaching each planter and increase
      the number of planters to include a more extensive robot garden
      project.

      
Electronic / Mechanical
      improvements:

      
-- Replace drive
      motors with stepper motors and add driver boards

      -- New robust wheels

      -- Amplified audio system for sounds you can actually hear

      -- New sound chip for advanced sounds

      -- Improved water level indicators

      -- Fix backlight issue

      -- Auto resetting docking flag

      -- Add high accuracy sonar chip

      -- New serial compass chip

      -- Better battery mounting


 

      
Interested in learning
      more on the development of this robot? Here are previous uploads
      - LATEST
      AT THE BOTTOM:

      
Previous Articles
1. Plant watering sensors
2. Cbot basic concepts and project definition
3.  Docking sensor, battery charging and FSM explained
4.  Development of the water tank, pump and feed mechanism
5.  Development of the reservoir refill system
6.  Incorporating the Triple axis magnetometer for compass navigation
7. Installing the final components - Bumper Switch Array
8.  Night - Day - Night Detection to trigger watering task
9.  Planter Water Distribution System for Autonomous Watering
10.  Plant progress upstairs after two weeks!
11. Two months in operation

BACK TO ROBOT PAGES

HOME