Objective:
To determine the better logistics of a robot; if it is optical-visual components or a touch sensor.
Hypothesis:
Our team thinks that the optical-visual component (using the light sensor) will be more practical, yet the touch sensor will be more precise. Since this demo robot will be used for measuring beads, optical-visual components are the better choice because it is more practical.
Materials Used for the Experimentation:
• Lego Baseboard (the base made out of Lego that the robot is built on)
• Lego
• Lego Gears
• Lego Axels
• Rubber Conveyor Belt
• Lego RCX (the “brain†of the robot)
• Lego USB tower to send the information from the computer program to the robot
• Two touch sensors
• One light sensor
• One piece of black construction paper
• Four “springs†made out of pipe cleaner
• One clear plastic cup
• Skittles (also known throughout our report as beads)
• Laptop (used to program the robot and send information to the robot)
• “Mindstorms Version 2.0″ computer program (used to program the robot)
• One bottle of silver paint
• Paint brushes
Materials Used for the Bristol Board:
• One black Bristol Board
• Logo (made on the computer using Adobe Photoshop Elements 2.0)
• Typed information
• Pink, Green and Blue Construction Paper (used as the background to paste the information on)
• Pencils, pens, pencil crayons, markers… etc.
• Log Book
Method:
Step 1: Built a stand for the robot using Lego RCX on the Lego baseboard.
Step 2: Built a tower for the beads to travel down.
Step 3: Installed a gate in the tower.
Step 4: Installed a conveyor belt.
Step 5: Placed the tower and conveyor belt on a stand.
Step 6: Created the main prototype for the light and touch sensor unit.
Step 7: Built the touch sensor unit, which includes the springs and the cup holder.
Step 8: Built the light sensor unit.
Step 9: Attached the two units together.
Step 10: Made the stand with the conveyor belt and the tower higher.
Step 11: Placed the light and touch sensor unit accordingly underneath the conveyor belt.
Step 12: Experimented through trial and error to create the settings correctly.
Step 13: Developed the programs.
Step 14: Tested the programs.
Step 15: Debugged and repaired the programs.
Step 16: Experimented the programs.
Step 17: Wrote the rough-copy of the lab report.
Step 18: Edited the rough-copy of the lab report.
Step 19: Created the log book.
Results:
Discussion:
The history of robotics is very profound in the world of technology. For example, Leonardo Da Vinci created a humanoid robot in the 1490’s. Now, robots help humans with tasks that are incredibly dangerous, such as bomb disposal, toxic waste, and mining. They are used in healthcare as well, by performing highly delicate, accurate surgery. Even in day-to-day life, robotics can be seen in items as simple as a car, or a garage door opener.
The robot was created to be used to illustrate the means of how medicine bottles could be filled with the respected medicaments, whether it would be using an optical visual component, or a touch sensor to gauge the weight.
With the robot that was built, the robot was counting the number of beads instead of counting the number of medicaments as it would be in real life. The optical-visual component of the robot worked better than the touch sensor. The optical-visual component had a higher average (mean) of its results, the most frequent result from the experimentation was the number that was considered to be perfection (meaning, the number that was considered as the “correct†number) which was 20, had a higher rate of achieving this “perfect†number, and was less frequent to having an inaccurate reading than all of those results obtained from using the touch sensor (see table under the heading “Resultsâ€). After our results were obtained and compared to one another, research available on the internet explained why these results lead to this conclusion. With the research found, the optical-visual component, the light sensor, was more accurate because unlike the touch sensor, its flaws are less likely to dramatically change the results.
The light sensor works by sending an LED (light emitting diode) light (a visible red light) that shines directly onto the surface facing it. The light will then bounce off the surface and bounce back towards the light sensor and into the phototransistor (the sensor), which receives the light that is bounced back from the cup.
The flaw in using the optical-visual component in the experiment was that the phototransistor’s accuracy as to how dark or bright it was depended on how the light was bounced off a surface, which depended on what type of surface the light was deflected off. If the surface was flat or smooth, the rays of light that bounced back would be deflected off directly back to the phototransistor (Specular Reflection), than if the light was deflected off a rugged surface. If the light is deflected off a rugged surface, the light bounced back in many directions (Diffuse Reflection). This causes a problem in the reading of brightness for the phototransistor because the phototransistor only receives light in a certain range of distance and its range is not broad enough to see the light from a diffuse reflection which will lead to an inaccurate reading of brightness because there would have been less light that is being bounced back.
The light sensor also influenced the results that were obtained because it had an LED next to the phototransistor, and therefore influenced the results by having them less accurate. This happened because it could not differentiate the amount of beads in the cup, as to the rays of light that are deflected, which gives us a reading that is not very precise. However, in our experiment, the LED was separated from the phototransistor by using a tiny piece of photonegative material. It stopped most of the leakage of light from the LED to phototransistor.
Also, the light sensor did not work well in rooms that had extra light. The testing was performed at night, with only a lamp on, to obtain the results. However, in the daytime the light sensor did not react as it did in the night. The cause of this is the amount of light being reflected off of the cup, and therefore into the phototransistor.
Even though the light sensor has at least three flaws that are known, through experimentation and research, it influenced the results obtained from the experiment far less than that of the single flaw of the touch sensor.
The touch sensor’s flaw was that after the weight of the cup brought down the cup and eventually touched the touch sensor (that is placed directly underneath the cup), the touch sensor had a tendency to bounce under the weight. When the touch sensor bounced, the weight that was being placed on the touch sensor caused the touch sensor to move inward and the pressure was placed on the switch in the circuit that is connected to the Lego RCX processor which processed the information given by the pressure from the weight of the cup on the touch sensor, but then quickly bounced back up turning the switch back off because the touch sensor was not applying enough pressure onto the switch in the circuit. An example of the bounce was in test number six, when five beads rolled off the conveyor belt at the same time. It pushed the cup downwards, which in turn pushed the touch sensor. This notified the RCX that the cup was “full†(when it was not), and ended the program.
Another flaw with the touch sensor, was that of mechanical engineering on Lego’s behalf. There was a great deal of noise when the touch sensor was off. When the touch sensor was in the transition between off (not pushed), and on (pushed), there were ‘bounces’ (different from the bounces described in the previous paragraph) that make it seem that the touch sensor can’t decide whether the switch is closed or opened. As described in the “Figures†section, the polarity might be the cause of this problem.
To conclude, if this robot was to be used in a real life scenario, it could be used as means to count the amount of pills that would be needed for a prescription in a hospital or pharmacy. The flaws would need to be considered, since they greatly affect the number that is said to be in the bottle through the reading of information by the Lego RCX processor. If it were to be used in a medical center or pharmacy, it would be a lot easier to fix the problems of the optical-visual component than the sole problem of the touch sensor because the light sensor has more potential to be altered where as a touch sensor does not.
Conclusion:
In conclusion, the use of an optical-visual component was proven to be more dependable on having a more precise reading of information than that of the touch sensor, thus allowing it to have a more accurate reading of the actual number of beads in the cup. It also proved to be more practical, considering it had an average of 19.1 beads in the cup, a mere 0.9 beads away from hitting the goal of 20 beads. Throughout this experiment, we have acquired an enormous amount of information, especially that of the history of robotics, and how robots work. A future experiment that could resolve any difficulties that arose is a different version of this robot, except it counts the number of beads that go into the cup, one at a time. The robot that would count the beads one at a time would be the follow-up experiment to this one, which would detect any problems that the robot had earlier with the ability to be more precise.
Videos:
Building the Robot
Building the Robot 2
Painting the Robot
Robot Demonstration
Thanks everyone for all your support!
Awards:
1st Place Science Fair 2007 at Kuper Academy
1st Place Science Fair 2007 Bell Montreal Regional Science & Technology Fair
McGill University - School of Computer Science (Robotics) Award
EÌcole Polytechnique de MontreÌal Award
All Rights Reserved 2007