Control of Mobile Robotics CDA4621 Fall 2021 Lab 2 Navigation Total: 100 points Due Date: 10-11-2021 by 8am The assignment is organized according to the following sections: (A) Lab Requirements, (B) Lab Description, (C) Task Description, (D) Task Evaluation, and (E) Lab Submission.

2. Lab Requirements

Software: Webots & Python (Empty worlds: Lab2_epuck.zip)

2. Lab Description

This lab will teach you about the following components: (1) Distance Sensors, (2) Cameras, and (3) PID. Use the E-Puck robot as in Lab1. B.1 Distance Sensors Webots e-puck robot has three distance sensors[1] (front, left and right). Each distance sensor has a lookupTable field that indicates min and max readings. For the lab, the distance readings are between 0 and 1.27 meters (50 inches). B.2 Camera The robot includes a camera[2], having an image size of 80 pixels x 80 pixels. For this lab, the robot needs to recognize a single yellow color set with the recognitionColors field. B.3 PID The robot will use PID to control motor velocity to distances read from the three different sensors in relation to the corresponding walls, using only the P component. Figure 1 shows a diagram for P control, with velocities proportional to the error in distance. Figure 1: PID control of the robot velocity proportional to its desired distance to the wall. The following equations summarize the diagram in Figure 1: (Eq. 1) (Eq. 2) (Eq. 3) (Eq. 4) (Eq. 5) where:

2. Task Description

The lab consists of the following tasks leading:

• Task 1 Motion with PID
• Task 2 Motion to Goal
• Task 3 Wall following
• Task 4 Bug Zero Algorithm
• Task 5 - Tangent Bug Algorithm - Finite Range Sensor (Extra Credit)

C.1 Task 1 Motion with PID (Task + Plot) Implement a controller called Lab2_Task1.py. The robot should use proportional control to move towards the end wall while also keeping its distance from the side walls, as shown in Figure 2. You will apply the P component of the PID controller to all 3 distance sensors. Robot should stop at the exact 10-inch mark from the end wall, while keeping a distance between 2.5 and 5.5 inches from side walls. You need to continuously print all three distance measurements on the screen. Robot should never hit a wall, and if gets too close, it should adjust its distance and orientation as necessary. You need to use only proportional gain control. Test the program using 6 different values of (0.1, 0.5, 1.0, 2.0, 2.5, 5.0). Apply the same control values to all three sensors at the same time. Find the one value that you think works best. Provide a single plot for each value, showing Distance from Wall (front, left, right) vs. Time, comparing the results for the 6 . Note that tile size is set to 0.508x0.508m (20x20inches), corresponding to 2x2 squares, i.e. each square in Figure 2 measures 0.254x0.254m (10x10inches). Task should run for 30 sec. NOTE: Implement your own PID equations. Do not use Webots prebuilt PID controller. During evaluation, the TA will start the robot at different distances and orientations varying less than 30 degrees in either orientation from the straight line towards the end wall. Figure 2: Robot starts 50 inches (1.27 m) away from the end wall. C.2 Task 2 Motion to Goal (Task + Plot) Implement a program called Lab2_Task2.py. The robot will start from the corner opposite the yellow cylinder, i.e. the goal, with any orientation, as shown in Figure 3. The robot should move towards the goal and must stop at 10 inches from the cylinder. There are no obstacles between the robot and the cylinder. Apply PID control to front distance sensor and use side control to realign the robot with the center of the cylinder. Provide a single plot, showing (a) Distance from Goal vs. Time, (b) Alignment with Goal vs. Time, (c) Blob Size vs. Time.Task should run for 30 sec. During evaluation, the TA will start the robot at different distances and orientations from the yellow cylinder. Figure 3. Camera color object recognition. C.3 Task 3 Wall Following Implement four wall following algorithms while applying the PID controller from Task 1:

• Implement Lab2_Task3_CorridorLeftTurns.py (Figure 4 - left).
• Implement Lab2_Task3_CorridorRightTurns.py (Figure 4 - left).
• Implement Lab2_Task3_MazeLeftTurns.py (Figure 4 - right).
• Implement Lab2_Task3_MazeRightTurns.py (Figure 4 - right).

If the robot reaches 7 inches from any end wall it should make a 90-degree. If no 90-degree turns are possible, it should make a 180-degree turn, and continue wall following in the opposite direction. The robot should navigate no further away than 7 inches from the wall it is following. The robot can start at any grid cell with any orientation and should stop when completing a full path by calculating the time it takes. Depending on the original orientation the robot my follow a different path. Task should be completed in less than 3 minutes. During evaluation, the TA will start the robot at different initial positions and orientations in the corridor or maze. Figure 4. Wall-Following Task: (Left) corridor and (Right) maze. C.4 Task 4 Bug Zero Algorithm Implement a program called Lab2_Task4.py. The robot should follow the Bug Zero algorithm, using the world shown in Figure 5. The goal is represented by the yellow-colored cylinder. The robot should follow walls no further than 8.4 inches away from them. The robot may use any of its sensors, including camera. Task should be completed in less than 3 minutes. During evaluation, the TA will start the robot at different initial positions and orientations. Figure 5. Bug algorithm maze. C.5 Task 5 Tangent Bug Algorithm (Task + Plot) (Extra Credit) Implement three different controllers for the Tangent Bug Algorithm using the Figure 5 world:

• Implement Lab2_Task5_ZeroRangeSensor.py using a sensor range of no further than 7 inches.
• Implement Lab2_Task5_FiniteRangeSensor.py using a sensor range of 15 inches.
• Implement Lab2_Task5_InfiniteRangeSensor.py using an infinite sensor range.

The algorithm should be based on the Tangent Bug Finite Range Sensor, using the world shown in Figure 5. Provide a diagram showing all hit, leave, and min points for each of the 3 subtasks. Note that you can modify the distance sensor range by: (a) modifying the distance sensor range in Webots corresponding sensor lookup table, or (b) comparing distance ranges directly in your Python code. You should get sensor readings as often as possible, and you may choose to add a Lidar range sensor or rotate your robot in order to get multiple readings of your front distance sensor to obtain endpoints for the intervals of continuity. Task should be completed in less than 3 minutes. During evaluation, the TA will start the robot at different initial positions and orientations. D. Task Evaluation Task evaluation involves: (1) programs, each task as a different controller, and (2) report, including links to a video showing each different task. Note that functions and tasks to be implemented are required in future labs. Points will be taken off if the robot crashes into walls or gets stuck in any way. Note that the TA may also test the different tasks under slightly different mazes configurations. D.1 Task Presentation (90 points) The following section shows the rubric for the tasks shown in the video:

• Task 1 (20 points)
  • Robot stabilizes at approximately 10 inches from the wall (5 points)
  • Robot restabilizes when placed further or closer to the wall (5 points)
  • Robot utilizes PID to reduce speeds proportionally to the front wall (5 points)
  • Robot maintains distance from side walls (5 points)
  • Robot hits walls (-5 points)
• Task 2 (20 points)
  • Robot faces the cylinder by centering it in the camera frame (5 points)
  • Robot re-faces the cylinder when the robot is moved (even when moving) (5 points)
  • Robot finds the cylinder when the cylinder is moved (5 points)
  • Robot stops 10 inches or less from the cylinder using PID(5 points)
  • Robot hits the cylinder (-5 points)
• Task 3 (25 points)
  • Robot follows a wall at approximately 7 inches after stabilizing (10 points)
  • Robot makes only left or right turns in the corresponding program (5 points)
  • Robot follows the closest wall it is placed near (5 points)
  • Robots average travel speed is greater than 3.5 inches per second (5 points)
  • Robot constantly exceeds 2.5 inches closer or farther from the wall (-5 points)
  • Robot hits walls (-5 points)
• Task 4 (25 points)
  • Robot reaches the cylinder from any starting position (10 points)
  • Robot switches to and from wall following appropriately for the Bug 0 algorithm (10 points)
  • Robot average travel speed is greater than 3.5 inches per second (2.5 points)
  • Robot stops only when within 10 inches of the cylinder (2.5 points)
  • Robot hits the cylinder (-5 points)
  • Robot always travels the same path, i.e. hardcoded path or always follows outer walls (-10 points)
• Task 5 (Extra Credit: 25 points)
  • Zero Range Sensor bug algorithm is correctly implemented (10 points)
  • Finite Range Sensor bug algorithm is correctly implemented (10 points)
  • Infinite Range Sensor bug algorithm is correctly implemented (5 points)
  • Robot hits walls (-5 points)
  • Robot hits cylinder (-5 points)

D.2 Task Report (10 Points) The report should include the following (points will be taken off if anything is missing):

2. Task 1 plots. (1 point)
4. Task 2 plot. (1 point)
6. Mathematical computations for all kinematics. Show how you calculated the speeds of the left and right servos given the input parameters for each task. Also, show how you decide whether the movement is possible or not. (3 point)
8. Video uploaded to Canvas showing the robot executing the different tasks. You should include in the video a description, written or voice, of each task. You can have a single or multiple videos. Note that videos will help assist task evaluations. (3 point)
10. Conclusions where you analyze any issues you encountered when running the tasks and how these could be improved. Conclusions need to show an insight of what the group has learnt (if anything) during the project. Phrases such as everything worked as expected or I enjoyed the project will not count as conclusions. (2 point)

2. Lab Submission

Each student needs to submit the programs and report through Canvas as a single zip file.

• The zip file should be named yourname_studentnumber_labnumber.zip
• The programs should start from a Lab2 folder and contain subfolders for each task, each with their corresponding controllers and worlds subfolders.
• The report should be in PDF and should be stored in the same zip file as the programs.