关于算法:MTRN4110-阶段性任务A

83次阅读

共计 17950 个字符,预计需要花费 45 分钟才能阅读完成。

MTRN4110 22T2 Phase A Task Description
(Week 1-3)

Updated 3/6/2022: Asking macOS users to configure Makefi le to use C++14 (Page 12, Section 4.1.20)
Released 30/5/2022

  1. Overview of the Course Project:
    The main project of MTRN4110 22T2 is a simulation-based project adapted from the Micromouse
    competition. Webots will be used as the simulation platform throughout the project. You will design
    a mobile robot and implement a controller and a vision program to negotiate a maze autonomously
    in Webots. The project will contribute 60% to your final mark in this course.
    The project consists of four sequential phases which are connected, but attempting one phase is not
    dependent on the completion of another:
    Phase A: Driving and Perception (Week 1-3, 14%, individual)
    Phase B: Path Planning (Week 4-6, 14%, individual)
    Phase C: Computer Vision (Week 7-9, 14%, individual)
    Phase D: Integration and Improvement (Week 10-12, 18%, group)
    This document will describe the tasks of Phase A.

UNSW MTRN4110 Robot Design 22T2 Page 2 of 17

  1. Overview of Phase A – Driving and Perception:
    Phase A aims to build a mobile robot’s driving and perception modules for the final maze-solving
    demonstration. You must complete the tasks of this phase by 13:00 Monday, Week 4 (20 June).
    2.1. Expectations:
    By the end of Phase A, you are expected to have:
    installed Webots R2021b on your computer properly;
    completed Webots tutorials (minimum – 1, 2, 4, recommended – 3, 5, 6);
    understood how to run simulations with robots and sensors in Webots;
    been able to create a controller for a robot that can execute a given motion plan;
    been able to detect surrounding objects using onboard sensors of a robot; and
    been able to output results in the specified format.
    2.2. Learning Outcomes Associated with this Assignment:
    LO1: Apply relevant theoretical knowledge pertaining to mobile robots, including
    locomotion, perception and localisation using onboard sensors, navigation and path
    planning, for practical problem-solving
    LO3: Demonstrate practical skills in mechatronics design, fabrication, and implementation

UNSW MTRN4110 Robot Design 22T2 Page 3 of 17

  1. Phase A Task Description:
    At the beginning of this phase, you will be given a Webots world file containing a five by nine maze
    and an E-puck robot. An example world file is shown in Fig. 1, where the robot is placed at the
    centre of the top-left corner of the maze and heading towards the south of the view (throughout
    the course project, we will refer to the top, bottom, left, and right of the maze as North, South,
    West, and East, respectively).

Fig. 1. An example maze layout and an E-puck robot
You must create and implement a controller for the robot, which should complete the following
tasks once started.
3.1. Read in a sequence of motion plan from a text file and display it in the console
You will be given a text file named“MotionPlan.txt”containing a sequence of motion
commands, e.g.,

Fig. 2. An example motion plan
UNSW MTRN4110 Robot Design 22T2 Page 4 of 17

The motion sequence starts with three characters specifying the initial location and heading of
the robot. In the example shown above, the sequence starts with
00S
where the first character (0) stands for the index of the row (0 – 4), the second (0) for the
column (0 – 8), and the third (S) for the heading of the robot (N, E, S, W). The top-left corner cell
always has an index of (0, 0) for the row and column.
Following the three characters is a sequence of motions represented by a string composed of
three characters (F, L, R), where F stands for“Forward for one step”, L for“Turn left for 90 deg”,
and R for“Turn right for 90 deg”.
In the example shown above, a sequence that directs the robot from the initial location to the
centre of the maze will be (you can validate it by yourself)
FLFFLFRFRFFFLFRFLFFLFRFLFLFFF
In summary, the sequence of motions in the text file will be like the following:
00SFLFFLFRFRFFFLFRFLFFLFRFLFLFFF
The text file given to you will have no spaces, newlines, or characters other than those
mentioned above. Also, note that the initial location and heading will always match the starting
state of the robot in the world file given to you, and the motion sequence is always valid (you
don’t need to check the correctness of the motion sequence or change the robot’s initial state).
Your program should read in the motion plan and display the exact string in the console once
the simulation is started.
If you find difficulty implementing reading information from a text file, you can choose to hard-
code the motion sequence into your program and forfeit the marks associated with it. In this
case, you must define a variable to store the motion sequence at the beginning of your program
(so that a tutor can replace it when assessing your work). You should also explicitly indicate you
are hard-coding the motion sequence in the Header Comment of your program. Failing to do so
would affect the assessment of your submission (incurring a 5% penalty, as if one-day late).
In the Header Comment, you should also indicate the platform (Windows/MacOS/Linux) you
used to develop your code so that a tutor can test your submission on your original platform.

Fig. 3. Explicitly indicate hard-coding in the Header Comment of your program

UNSW MTRN4110 Robot Design 22T2 Page 5 of 17

3.2. Display the initial location and heading of the robot and the existence of the
surrounding walls, and write the information into a CSV file
You should display the step that the robot is executing. In the initial state, the message should
be (using three digits): Step: 000.
After showing the retrieved motion plan, the next step is to parse the information and print the
initial location and heading of the robot in the console: Row: 0, Column: 0, Heading: S.
Besides, you should detect any walls in the robot’s front, left, and right. Only the walls
surrounding the cell where the robot is located should be considered.
The E-puck robot has some onboard sensors installed by default. In addition, the robot is added
with three Distance Sensors and one Inertial Unit Sensor under the node. You can
use any or all of these four sensors together with E-puck’s onboard sensors for the tasks. No
other sensors should be added to the robot.

Show the existence of the left, front, and right walls. If a wall is detected, print Y; otherwise,
print N. In the example of Fig 1, the detected walls should be: Left Wall: N, Front Wall: N, Right
Wall: Y
In summary, you should display the following exact message at the initial stage:
Step: 000, Row: 0, Column: 0, Heading: S, Left Wall: N, Front Wall: N, Right Wall: Y
You should also write this information into a CSV file named as“MotionExecution.csv”which is
stored in the same folder as“MotionPlan.txt”. The items should be delimited by commas.
At this step, the CSV file should look like this:

or this if you open it with Notepad:

3.3. Drive the robot following the motion plan
Drive the robot according to the parsed motion plan step by step.

UNSW MTRN4110 Robot Design 22T2 Page 6 of 17

If the motion step to be executed is‘F’, move the robot Forward for one cell.

Fig. 4. Move forward for one cell
If the motion step to be executed is‘L’, turn the robot 90 deg to its Left.

Fig. 5. Turn left for 90 deg

UNSW MTRN4110 Robot Design 22T2 Page 7 of 17

If the motion to be executed step is‘R’, turn the robot 90 deg to its Right.

Fig. 6. Turn right for 90 deg
The robot should keep clear of the walls when moving. A penalty will be incurred if the robot
hits any walls.
3.4. Display the location and heading of the robot, and the existence of the
surrounding walls after each motion step, and write the information into the CSV
file
Once the robot completes a step, you should show in the console the new location and heading
of the robot and the left, front, and right walls of the new cell that the robot stands in.
For example, at the end of the second step, the robot moves to the following location with its
heading towards EAST:

UNSW MTRN4110 Robot Design 22T2 Page 8 of 17

Fig. 7. Example robot location and heading
You should show the following exact message in the console:
Step: 002, Row: 1, Column: 0, Heading: E, Left Wall: N, Front Wall: N, Right Wall: Y
In addition, you should add this information to“MotionExecution.csv”:

3.5. Repeat tasks 3.3 and 3.4 until all the motion commands are executed
Repeat tasks 3.3 and 3.4 until all the motion commands are executed. Print a message“Motion
plan executed!”after the robot completes all the motions.
If the motion plan is:
00SFLFFLFRFRFFFLFRFLFFLFRFLFLFFF
when all the motion commands are executed, the robot should have reached the following
state:

UNSW MTRN4110 Robot Design 22T2 Page 9 of 17

Fig. 8. Robot reaching the centre after execution of all motion commands
And messages printed in the console should be:

Fig. 9. Messages displayed after execution of all motion commands

Note that each message should have a prefix“[z1234567_MTRN4110_PhaseA]”where
z1234567 is replaced with your zID. Your messages should look exactly the same as shown
above.

UNSW MTRN4110 Robot Design 22T2 Page 10 of 17

And“MotionExecution.csv”should be exactly like the following:

Fig 10. Content of CSV after execution of all motion commands
3.6. Task summary:

Task Description
1 Read in a sequence of motion plan from a text file and display it in the console
2 Display the initial location and heading of the robot, and the existence of the surrounding
walls, and write the information into a CSV file
3 Drive the robot following the motion plan
4 Display the location and heading of the robot, and the existence of the surrounding walls
after each motion step, and write the information into the CSV file
5 Repeat tasks 3.3 and 3.4 until all the motion commands are executed

UNSW MTRN4110 Robot Design 22T2 Page 11 of 17

  1. Specifications and Hints:
    4.1. Specifications:
    Maze:
  2. You should install Webots R2021b; otherwise, there will be some incompatibility issues.
  3. At the beginning of Phase A, you will be given the same maze layout as shown in the
    example.
  4. The initial location and heading of the robot will also be the same as illustrated.
  5. This setup is for your practice. For assessment, you may be tested with a different (but
    similar) maze layout.
  6. You are expected to test your program with various mazes and motion sequences,
    including edge cases. Failing to do so may cause poor performance in some test cases.
  7. The initial location and heading of the robot may also be different and could be at any cell
    of the maze.
  8. The initial location of the robot will always be at the centre of a cell.
  9. The initial heading of the robot will always be towards one of the four directions (North,
    East, South, West).
  10. The first three characters in the motion plan given to you will always match the world file.
  11. The motion sequence may also be different from the example, but should always be valid
    (no collision with walls if executed correctly).
  12. The maze will always be five by nine, and the four borders are always closed.
  13. The distance between neighbouring cells is always 0.165 m.
  14. The thickness of the walls is always 0.015m.
    Robot:
  15. For Phase A, you must use the provided E-puck robot for the tasks and not modify the
    robot. You will be asked to submit your controller only. Working on a modified robot
    during development could lead to incompatibility with the assessor’s world file and give
    you advantages over other students who comply with the constraints, and thus would
    incur a severe penalty (even getting zero marks).
  16. Three Distance Sensors and one Inertial Unit sensor have been added to the
    node. You can use any or all of these sensors together with E-puck’s onboard sensors for
    the tasks. Note all four sensors have noise (expand the feature tree of the world file to see
    more details); you should carefully handle the noise when using the sensors.
  17. You can modify the maze layout for your practice. But you will only submit your controller,
    which will be tested with the provided world file.
  18. The characteristics of the E-puck robot can be found here. However, if you are using dead
    reckoning, the wheel radius and axle length shown in the table may be inaccurate. You are
    recommended to tune these parameters using the calibration method introduced during
    the lecture.
    Implementation:
  19. You must use C++ to implement the controller if you have taken MTRN2500 Computing for
    Mechatronic Engineers before.

UNSW MTRN4110 Robot Design 22T2 Page 12 of 17

  1. You can choose to use C if you have not taken MTRN2500, but you need to get the
    lecturer’s approval before using it.
  2. For portability, you must use the built-in compilers for Windows or macOS.
    o For macOS users, please open the Makefile of your controller and add the following
    line under the CFLAGS section (CFLAGS = -std=c++14).
    o
  3. The text file storing the motion sequence should be named“MotionPlan.txt”. You should
    define a path variable at the beginning of your controller program indicating the path of
    this text file, e.g.,
    const std::string MOTION_PLAN_FILE_NAME =“../../MotionPlan.txt”;
    where ../../MotionPlan.txt will allow you to access the file if the folder structure specified in
    Section 5.1 is followed.
  4. The CSV file storing the execution information should be named“MotionExecution.csv”and
    stored in the same folder as“MotionPlan.txt”. You should define a path variable at the
    beginning of your controller program indicating the path of this CSV file, e.g.,
    const std::string MOTION_EXECUTION_FILE_NAME =“../../MotionExecution.csv”;
  5. You should use 64 for the TIME_STEP, as used in the Webots Tutorials. No other values
    should be used.
  6. You should make your code modular with good interfaces for the sake of integration in
    Phase D (refer to the draft specs released).
    4.2. Hints:
  7. Consult the lecturer/demonstrators if you are unclear about anything.
  8. You can use standard printing functions for debugging in Webots. If you want to check the
    value of a variable during the simulation, you can print it to the console. If you want to see
    until which line the controller runs successfully, you can also add printing breakpoints at
    certain steps.
  9. Try decreasing the robot’s speed if you use position control mode and the robot does not
    move to a position as specified.
  10. Make the robot stop for a while before reading the sensors to get robust wall detection.
  11. Try filtering the sensor readings to get more robust measurements.
  12. You should use forward slash / or double backward slash \ instead of single backward slash
    \ to define the path of the file.

UNSW MTRN4110 Robot Design 22T2 Page 13 of 17

  1. Assessment:
    5.1. Submission of your work
    During your development, your project folder should look like this:
    z*_MTRN4110_PhaseA
    |–controllers
    | |–z*_MTRN4110_PhaseA
    | |–build
    | |–Makefile
    | |– z*_MTRN4110_PhaseA.cpp
    | |– z*_MTRN4110_PhaseA.exe
    |–worlds
    | |–z*_MTRN4110_PhaseA.wbt
    |–MotionExecution.csv
    |–MotionPlan.txt
    You should zip (and only zip) your“controllers”folder and name it as
    “z_PhaseA_controllers.zip”where z is your zID. Submit this zip file to Moodle.
    Your submission should only contain one controller, which is the finalised version. It is your
    responsibility to make sure your submission is self-contained and up-to-date.
    As we have developed an automatic tool to do the plagiarism check on your code, it is CRUCIAL
    that you strictly follow the file structure specified above. Any violation in the submission format
    would incur a 5% penalty (as if one-day late), as that adds much more work to the assessors
    (we will still do the plagiarism check!).
    5.2. Marking criteria:
    This assignment will contribute 14% to your final mark.
    You will be assessed five times with different setups. Among them, one test will be on the
    example given to you for practice. Your final mark will be calculated as the following:
    markfinal = (markexample + marknewtest1 + marknewtest2 + marknewtest3 + marknewtest4) / 5
    Each attempt will be assessed by using the following criteria.
    Task Description Marking (out of 100%)
  2. Read in a sequence of
    motion plan from a text file
    and display it in the console
    +10% if all correct1, otherwise
    (8% maximum):
    o +1% each for every 3 consecutive characters
    correctly displayed. A character is considered
    correctly displayed if and only if this character
    and all the preceding characters are correctly
    displayed.
    (zero marks):
    o if hard-coding the motion plan into program

UNSW MTRN4110 Robot Design 22T2 Page 14 of 17

2 Display the initial location
and heading of the robot,
and the existence of the
surrounding walls, and write
the information into a CSV
file
+10% if all correct1, otherwise
(8% maximum):
o +1% for correctly displaying and writing the
initial location and heading to CSV
o +3% for correctly displaying and writing the left
wall to CSV
o +3% for correctly displaying and writing the
front wall to CSV
o +3% for correctly displaying and writing the
right wall to CSV
3 Drive the robot following
the motion plan
+40% if all correct, otherwise
(36% maximum):
o X / N * 40%, where X is the number of cells
traversed with successful move2 and N is the
total number of cells traversed by the given
motion plan
4 Display the location and
heading of the robot, and
the existence of the
surrounding walls after each
motion step, and write the
information into the csv file
+40% if all correct1, otherwise
(36% maximum):
o Y / M * 40%, where Y is the number of correct
values3 for location/heading/walls and M is the
total number of values for
location/heading/walls expected from the given
motion plan
5 Repeat tasks 3.3 and 3.4
until all the motion
commands are executed

1correct means the messages (both value and format) in console and CSV are exactly the same
as required (see Fig. 9 and 10). A message that is essentially correct but in the wrong format will
be deemed incorrect. Note in particular: step, row, and column should start from 0; the order of
the items should not be changed; the letters used should be the same as specified.
2A move is successful if the robot is correctly and fully moved to a new cell or rotated for 90 deg,
without colliding with any walls.
3A value is correct if the robot’s current and past moves are successful as defined above and the
reporting of the location/heading/walls at the current cell is correct.
You should make sure your submitted project is self-contained and up-to-date. Demonstrators
will only replace the world file and the motion plan file for different tests; no debugging/changes
to your code (except in the case of hard-coding the motion plan into the program) should be
expected from the assessor. If your code did not run correctly (e.g., crashing after starting), you
could get zero marks for that test.
5.3. Deadline
The submission will be open from 13:00 AEST 14 June 2022 (Monday Week 3) and the deadline
is 13:00 AEST 20 June 2022 (Monday Week 4).

UNSW MTRN4110 Robot Design 22T2 Page 15 of 17

We will apply a university-wide late policy that has been specified by UNSW and MME (refer to
the course outline):

Note the 5% penalty will be multiplied by the maximum possible mark and applied directly to
the awarded mark (refer to the example).
Students are expected to manage their time to meet deadlines or request extensions (apply for
special consideration) as early as possible before the deadline.
5.4. Progress Check
You will have your progress checked with your demonstrator in a 5 min meeting in your Week 2
workshop session.
During the session, please show your progress to your demonstrator in person. For the online
workshop, please share your screen with your demonstrator. If you don’t know how to do this,
please watch this short video: How to share your screen in a Microsoft Teams meeting
To pass the progress check, you must demonstrate you can create a controller for the robot, move
the robot forward for one step, turn the robot left for 90 deg, and print the locations and headings
of the robot (no sensing is required). These should be completed in one program run.
The progress check makes up 1% of the overall course mark (included in the 14%).
All progress checks should be completed before the end of the scheduled workshop session. The
late submission policy does not apply to progress checks.
5.5. Plagiarism
If you are unclear about the definition of plagiarism, please refer to What is Plagiarism? | UNSW
Current Students.
You would get zero marks for the assignment if you were found:

UNSW MTRN4110 Robot Design 22T2 Page 16 of 17

Knowingly providing your work to anyone and it was subsequently submitted (by
anyone), or
Copying or submitting any other persons’work, including code from previous students
of this course (except general public open-source libraries/code, for which you should
explicitly cite the source wherever applicable).
You will be notified and allowed to justify your case before such a penalty is applied.

UNSW MTRN4110 Robot Design 22T2 Page 17 of 17

  1. Additional Resources:
    Webots download: https://cyberbotics.com/
    Note that if you encounter this warning when installing Webots, you can ignore it and the
    software usually works properly. If you do have issues installing Webots on your computer, you
    can use any computer in the MECH Computers Lab (Room 203, J17) which has the latest version
    of Webots installed already.

    Webots user guide: https://cyberbotics.com/doc/guide/index
    Webots reference manual: https://cyberbotics.com/doc/reference/index
    Webots tutorials: https://cyberbotics.com/doc/guide/tutorials
    E-puck: https://cyberbotics.com/doc/guide/epuck
    Webots sensors: https://cyberbotics.com/doc/guide/sensors

正文完
 0