Lab 2: Odometry

Design objectives

    1. Design and implement an Odometry system that provides a robot’s position and orientation, allowing a robot to autonomously navigate a field.

    2. Implement a simple correction using a light sensor to improve the Odometer results and return values relative to the defined origin.

    3. Evaluate the design and determine the accuracy of the implemented Odometry system.


Design requirements

The following design requirements must be met by your robot:

    • Odometer

    o Must determine the robots X, Y and θ orientation. o Must display the X, Y and θ on the LCD display. o X and Y must be in cm.

o X and Y can be negative. o θ must be in degrees.

o  θ must range from [0⁰, 359.9⁰].
        ◦ When the value increases past 360⁰, it should return to 0⁰.

        ◦ When the value decreases past 0⁰, it should wrap to 359.9⁰.

        o The zero values for (X, Y), i.e. (0,0), must respect the convention shown in Figure 3.

    • Odometer correction

        o Must use the light sensor to detect lines.

    o Must work for an arbitrary path, i.e. rectangles of various dimensions.







 
ECSE 211 Design Principles & Methods    Updated: 17 September 2019


Demonstration (30 points)

The design must satisfy the requirements by completing the demonstration outlined below.

Design presentation (10 points)

Before demoing the design, your group will be asked some questions for less than 5 minutes. You will present your design and answer questions designed to test your individual understanding of the lab concepts. Each person will be graded individually.

You must present your workflow, an overview of the hardware design, and an overview of the software functionality. Visualizing software with graphics such as flow charts is valuable.

Float Motors (10 points)

The TA will check whether the X, Y and θ values are updated correctly on the robot’s LCD screen by floating the robot’s motors/wheels. Note that you can choose any X, Y and θ convention as long as you remain consistent. All three axes (X, Y, θ) are checked:

    • X & Y values work

    • θ values work
    • 5 points

    • 5 points




E.g. if the (X, Y, θ) convention is set as in Figure 1, then:

    • Moving both wheels forward should increase Y.

    • Moving both wheels backward should decrease Y.

    • Moving the right wheel backward and the left wheel forward simultaneously should increase θ.

    • Moving the right wheel forward and the left wheel backward simultaneously should decrease θ.







E.g. if the (X, Y, θ) convention is set as in Figure 2, then:

    • Moving both wheels forward should increase X.

    • Moving both wheels backward should decrease X.

    • Moving the right wheel backward and the left wheel forward simultaneously should increase θ.

    • Moving the right wheel forward and the left wheel backward simultaneously should decrease θ.














Figure 1. Robot faces north at 0°.



Figure 2. Robot faces east at 90°.
 
2/8
ECSE 211 Design Principles & Methods    Updated: 17 September 2019


Odometry Correction Check (10 points)

The TA will ask you to run your robot off the center of a tile, as shown by S in Figure 3 (the exact X/Y placement is not critical, as the correction should compensate). The robot should then follow the 3-by-3 tile square trajectory using SquareDriver. The robot should work using odometry correction. Throughout the demo, the TA will observe the reported (X, Y, θ) values on the robot’s LCD screen. When the robot stops at the final position (XF,YF) near S, the final readings on the LCD screen (X, Y, θ) are used to evaluate the odometers accuracy and calculate the error distance ϵ as:
=√(  −    ) +(  −    )



Note that the error ϵ is calculated as the Euclidean distance between:

    • The odometer’s readings (X,Y), which signifies where the robot thinks it is with respect to the origin (0,0), and

    • The final actual position (XF,YF), which ideally should be the point S where the robot started the 3-by-3 tile square trajectory.

This means that it is not an issue if your robot does not return to the exact starting point S, as long as the odometer reports a position that matches its real-world location.


Point grid based on error ϵ:

[0, 2] cm
→  5 points
(2,
4] cm
→
2.5 points
(4,
∞) cm
→
0 points


Point grid based on the difference between the displayed θ and actual θ:
[0, 10] °
→  5 points
(10,
20] °
→
2.5 points
(20,
∞) °
→
0 points






(0, 0)
Figure 3. 3-by-3 tile trajectory using SquareDriver.



 
3/8
ECSE 211 Design Principles & Methods    Updated: 17 September 2019


Provided materials

Sample code

A package of sample code is provided that contains the following:

    • Display.java

        o Provides a display mechanism for the Odometer.

        o Runs in a thread

    • Main.java

        o The main class that runs the robot.

        o Starts Odometer thread and Odometer correction thread.

        o Drives the robot using SquareDriver.

    • Odometer.java

        o A skeleton class for building an odometer.

    o Runs in a thread.

        o Provides methods for manipulation of volatile odometer variables, such as X and Y position.

    • OdometryCorrection.java

        o Provides a skeleton class for building a correction algorithm for the Odometer.

        o Runs in a thread.



    • Resources.java

        o Defines the ports used by motors and sensors.



    • SquareDriver.java

        o Runs the robot in a 3-by-3 tile square, where one tile is 30.48cm.

Physical material

In the lab, tiles with black grid lines are provided. These make up the competition floor, where the robot will operate. Grid lines are separated by a distance of 30.48cm.


Implementation instructions

    1. In Odometer.java, implement your odometer design in the run() method of the Odometer class. This class is threaded and will run continuously when your robot is working.

 
4/8
ECSE 211 Design Principles & Methods    Updated: 17 September 2019


    2. In Lab2.java, tweak the values of WHEEL_RAD and TRACK so that your robot drives in a square pattern when calling SquareDriver.drive().

    3. In OdometryCorrection.java, implement a design that uses the light sensor to detect the grid lines and update/correct the odometers position as needed. Note, your design should work regardless of the length and width of the rectangular pattern you drive. We do not expect you to be able to correct when driving diagonally across tiles. You will need to read values from a light sensor. For more information, see the leJOS API documentation provided on MyCourses.


Report Requirements

The following sections must be included in your report. Answer all questions in the lab report and copy them into your report. For more information, refer to ECSE211SubmissionInstructions.pdf. Always provide justifications and explanations for all your answers.

Section 1: Design Evaluation

You should concisely explain the overall design of your software and hardware. You must present your workflow, an overview of the hardware design, and an overview of the software functionality. You must briefly talk about your design choices before arriving at your final design. Visualizing hardware and software with graphics (i.e. flowcharts, class diagrams) must be shown.

Section 2: Test Data

This section describes what data must be collected to evaluate your design requirements. Collect the data using the methodology described below and present it in your report.

Odometer test (10 independent trials)

    1. Note the starting position S of the robot’s center and consider it to be (0,0) for this trial.

    2. Run the robot in a 3-by-3 tile square, without using Odometry correction.

    3. Measure its resulting signed XF and YF position with respect to its starting position S.

    4. Note the reported values of X and Y shown for the odometer.

Odometer correction test (10 independent trials)

    1. Place the robot approximately at the starting position S. You do not need to note the starting position.

    2. Run the robot in a 3-by-3 tile square using Odometry correction.

    3. Measure its resulting signed XF and YF position with respect to the origin (0,0) in Figure 3.

    4. Note the reported values of X and Y shown for the odometer.

Section 3: Test Analysis

Present the following analysis in a table in your report

1.    Compute the Euclidean error distance ϵ of the position for each test.
 
5/8
ECSE 211 Design Principles & Methods    Updated: 17 September 2019


    2. Compute the mean and standard deviation for X, Y, and ϵ for both test sets. That means, you need to perform 6 mean and 6 standard deviation calculations in total. Use the sample standard deviation formula.

Answer the following questions in your report

    1. How do the mean and standard deviation change between the design with and without correction? What causes this variation and what does it mean for the designs?

    2. Given the design which uses correction, do you expect the error in the X direction or in the Y direction to be smaller?

Section 4: Observations and Conclusions

    • Is the error you observed in the odometer, when there is no correction, tolerable for larger distances? What happens if the robot travels 5 times the 3-by-3 grid’s distance?

    • Do you expect the odometer’s error to grow linearly with respect to travel distance? Why?

Section 5: Further Improvements

    • Propose a means of reducing the slip of the robot’s wheels using software.

    • Propose a means of correcting the angle reported by the odometer using software when: o The robot has two light sensors.

o  The robot has only one light sensor.






























 
6/8
ECSE 211 Design Principles & Methods    Updated: 17 September 2019


Frequently asked questions (FAQ)


    1. What is meant by “design presentation”?

Before a lab demo, you and your partner will briefly present your design. This can include a basic visualization of how your code functions, such as a flow chart. You will then be asked a series of questions. These could be related to the lab tutorial and the initial lab code. For this part, a grade of 10 signifies full understanding, 5 signifies satisfactory understanding, while 0 shows no understanding at all. Note that memorized answers are discouraged and both partners should be responsible for understanding the design and their associated code, even if one of them did not write all of it.

    2. Are partial points awarded for Float Motors?

No partial points are awarded. Possible demo points: {0, 5, 10}.

    3. What is the length of each square tile? Each tile is 30.48 cm long.

    4. Do the displayed values of θ have to be in ° (degrees)? Yes.

    5. Do I have to follow the same (X, Y, θ) convention as in Figure 1 and Figure 2? No, you can use any (X, Y, θ) as long as you remain consistent.

    6. Should the point S be exactly at the center of a square tile i.e. (-15.24, -15.24)?

No, the starting point S may be off-center as your correction should compensate. Note that the TA will use the intersection of the two grid lines as the origin (0,0). Be sure to use your light sensor to get the correct values.

    7. In the Float Motors part, what will be the initial conditions of the robot?

The robot's initial condition is based on your (X, Y, θ) convention - you must inform the TA about your initial orientation. Once the float motors option is selected on the robot, the TA will place your robot on a table. Assuming the same convention as in Figure 1, the TA will forcefully rotate both the wheels forward to test the Y values. Not that the robot will not move automatically, but rather the TA will rotate both wheels using his/her hands. The TA wants to notice whether the Y-value will increase or not, without affecting the X and θ values by much (since they could be affected due to experimental error). Therefore, the odometer's accuracy is irrelevant here. The main idea is to observe whether the odometer functions well using a fixed convention. A similar wheel-rotation test is also performed for checking decreasing values as well as for other variables.

 
7/8
ECSE 211 Design Principles & Methods    Updated: 17 September 2019



    8. How accurate should the (X, Y, θ) values be during the floating motor demo? Should the values only increase and decrease properly without considering the errors? Accuracy of (X, Y, θ) values in floating motor demo is irrelevant. However, let us consider an example when a robot that is oriented along the +Y-axis and both wheels are moved forward. If the reported Y-values are negative, this means that the Y-axis does not work.

    9. How do I detect lines?

We recommend using a differential filter to detect black lines. See the lecture for details.















































© Instructor and Teaching Assistant generated course materials (e.g., handouts, notes, summaries, assignments, exam questions, etc.) are protected by law and may not be copied or distributed in any form or in any medium without explicit permission of the instructor. Note that infringements of copyright can be subject to follow up by the University under the Code of Student Conduct and Disciplinary Procedures.


8/8