Day 5 (15.08.08)

Thursday, August 14th, 2008

Can Robots Mindstorm? - Day 5


Today it is the competition day. We made a Power-Point presentation of what we did in the past few days =) It can be seen at the following link:
http://www.savefiles.net/d/fdd2i3p6vhn3.html



We are also proud to present Anna´s Beautiful Buggy (BB) :
Its biggest advantage is that it does not need batteries to work. It obeys human hand perfectly.
So far, no problems controlling it. =)







The competition is about to start. We keep our fingers crossed that everything will work and that out little robot does it well =)

Day 4 (14.08.08)

Wednesday, August 13th, 2008

Can Robots Mindstorm? - Day 4


Today we were working on the competition assignments. Our robot ( on the picture below) should be able to complete the following tasks:

• Find and measure distance to the box
• Open door and go into the garage
• The pushbox obstacle (push the box and drive through the gates)
• The golf course (Detect whether a ball is red or blue, then push red ones to one ditection and the blue ones to the opposite).
• The odometry test (drive the square-shaped track first clockwise, then counter-clockwise, while measuring the driven distances and detect the difference between them)

We tried very hard to do the tasks but we had several problems to solve. So, let´s keep our fingers crossed that our little robot will do fine at the competition tomorrow. =)

Here are the descriptions of how we solved the tasks:


TASK I.: Find and measure distance to the box


The robot starts going straight ahead, passes the black line, then turns 90 degrees to the right. Then it contunues going straight on until the ultrasonic sensor (mounted on the right side of the robot) detects the box. So the robot stops when the sensor is parallel with the box. The distance is shown on the display.

Problems:

• Since the sensor is not mounted above the wheels (but the distance is being measured by detecting rotations of wheels), we had to adjust the displayed distance.
  
• At first we tried to make the robot to follow the black line. This turned out to be very difficult because the light sensor was not accurate and reliable enough. Also the measured distance was not accurate.

TASK II.: Open the door and go into the garage.



Using the ultrasonic sensor and given distances we programmed the vehicle to move around the door. Then it pushes the door open, goes inside and stops.

Problems:

• The vehicle does not always turn as it should.
  
• The bumper sometimes missed the door because it was to small, so we had to rebuild it.

TASK III.: The pushbox obstacle

Our solution was to make an aggressive attack =) It simply touches the box and moves past it through the gates. If it touches the box then the movement is quite simple and effective.



No problems here =)


TASK IV.: The golf course



We attached arms and a light sensor to the robot. We had to do the task with blue and red balls, which was not very useful. The light sensor we used detects colors in greyscale, and there is very little difference between blue and red. If the sensor is pre-calibrated, it would not be able to recognize the correct color if the ambient light changes.
Therefore our robot initializes the values for blue and red ball every time before the program starts. Then it moves ahead. When the ultrasonic sensor detects a ball, it slows down. Then it stops, grabs the ball with arms, reads the value with light sensor and pushes the ball to the appropriate direction, depending on the ball`s color. Then it moves back and continues until it finds the next ball.

Problems:

• Small intensity difference between the used colours
  
• Sometimes the vehicle does not turn around in exact angels.
• Ball can bounce away if it is not caught accurately.

TASK V.: The odometry test



We calculated number of rotations of the motor needed for the given distance on each side of the square. Then we made the vehicle move around the cylinders. After one round it turns and goes counterclockwise. At the end it shows the length.

Problems:

• Vehicle does not always turn accurately, but turns for a larger or smaller angel.
  
• The latter problem was partly caused by low batteries. We replaced them, but the vehicle still does not always go the way it should.
  

Day 3 (13.08.08)

Wednesday, August 13th, 2008

Can Robots Mindstorm? - Day 3

Lecture

We attended a presentation concerning autonomous robot systems and the principles of their software structure.

Practical Assignment

Describe the robot you would like to own:
A domestic serving robot could be very useful. Here is the one presented at the Ubiquitous Dream Exhibition (Seoul, Korea)

Here is WALL-E (Waste Allocation Load Lifter Earth-Class), the cutest little robot.

Day 2 (12.08.08)

Tuesday, August 12th, 2008

Can Robots Mindstorm? - Day 2

Lecture

We attended a presentation on "Sensors for Mobile Robots" concerning mainly the position measurements and the related odometry errors (systematic and non-systematic).

Practical Assignments

Today we had to do the following tasks:

Exercise 1

Take 5 measurements from 5-30 cm with 1 cm resolution and another 5 measurements from 30-150 cm with 5 cm resolution using a cardboard box as a target and make a plot of actual versus measured distance.

We made a program that helped us measure the distances after every centimeter (prog. names: worknr2, worknr3) and we looped it for 5 times to get all the data for the diagrams. After completing the tests we compared the diagrams (on the picture below). We could conclude that the difference between the actual and measured distance is bigger when resolution is smaller (1 cm) (the original pictures are on the computer).

Exercise 2

-> Calculate cm for your tribot:
Our tribot wheel diameter is D=5.6cm and C=360 ticks. Based on formula for 1 full turn cm=pi*D/C=0.048869

-> Measure the distance between wheels: b = 11.3 cm

-> Measure cm by making the robot go forward 2m:
programming the code for the robot to move 2m (worknr4) wheels made 4917 ticks and cm=200/4417=0.045279

-> Try to check b by calculating n for a 90-degree turn and compare with what the robot actually does:
After 6 tests we calculated the average number of clicks that the wheels did while turning the robot for 90 degrees.
Based on formula: b = (ΔUr-ΔUl)*2 /pi = 11.0954

Day 1 (11.08.08)

Monday, August 11th, 2008

Can Robots Mindstorm? - Day 1

Lecture
We attended an introductory presentation about the basics of programming LEGO Mindstorm robots using the NXT-G Envirnoment.

Practical Assignments
Today we had to do the following tasks:
- look in the web for more detailed NXT tutorials, handbooks or guides and include links in the blog,
- modify the continuous scanning program to show in the NXT screen the angular position and distance for each measurement,
- optional: modify the continuous scanning program to register the angular position and distance to the closest object.

Links to some NXT websites:
http://home.earthlink.net/~kodiaq/id6.html
http://www.nxtclub.com/
http://www.nxtasy.org/
http://www.nxtprograms.com/

We have done the second task and saved the program file on the computer. We also tried to do the optional task, but not so succesfully (so far).