Github repository for AGR9027

Teaching materials for AGR9027 - Agri-robotics in digitalisation (MSc Agri-food Technology, University of Lincoln). This repository contains example Python scripts, sample images and laptop-friendly simulators to help you learn the basics of robotics programming, computer vision and machine learning, and to prepare for Assessment 2 (a Trilobot robotics project with Tasks A, B and C).

No prior experience with robotics, Python or machine learning is assumed - the example scripts are meant to be simple, heavily commented, and easy to copy and modify as a starting point for your own work.

Repository structure

AGR9027/
  sample_images/           <- Sample images for offline image-processing practice
  scripts/
    basic_python_1/         <- Jupyter notebook examples for introductory Python class 1
    basic_python_2/         <- Jupyter notebook examples for introductory Python class 2
    trilobot/               <- Example scripts for the real Trilobot robot (hardware)
    computer_vision/        <- Computer vision basics (taught session, coming soon)
    machine_learning/       <- Machine learning basics (taught session, coming soon)
  trilobot_simulator/      <- Laptop simulators for Tasks A, B and C (no robot needed)

• sample_images/ - A collection of sample images (e.g. coloured shapes) you can load and process with OpenCV without needing a camera or the robot, you can also print them and show them to the robot/laptop camera, useful for testing/practising image-processing code offline.

• scripts/ - Python learning materials and example scripts, organised into subfolders:

  • scripts/basic_python_1/ - Jupyter notebook examples for the first introductory Python programming class. Topics include first programs, mathematical operations, variables and data types, and collections such as lists, tuples and dictionaries.
  • scripts/basic_python_2/ - Jupyter notebook examples for the second introductory Python programming class, including comparisons, loops, classification and regression.
  • scripts/trilobot/ - Simple, generic scripts that run on the real Trilobot robot (motors, ultrasound sensor, underlighting, camera). See its own README for a full description of every script and how they build on each other from basic to more complex. These are building blocks for Assessment 2, not finished solutions to Tasks A, B or C.
  • scripts/computer_vision/ - Placeholder, coming soon. Will contain sample scripts covering computer vision basics, for a dedicated taught session.
  • scripts/machine_learning/ - Placeholder, coming soon. Will contain sample scripts covering machine learning basics, for a dedicated taught session.

• trilobot_simulator/ - Simulators for Tasks A, B and C, plus a webcam vision harness, to let you develop and test your Assessment 2 logic on your own laptop without needing the real robot or access to the University lab. See its own README for setup and usage.

Assessment 2 overview

Assessment 2 asks you to use a real Trilobot robot to complete three tasks:

• Task A - Reactive obstacle avoidance: drive the robot from a start point to an end point while avoiding obstacles placed in between.
• Task B - Person following: keep the robot at a fixed distance (0.5 m) from a target person, while ignoring decoy people/objects.
• Task C - Ripe strawberry counting: count the number of ripe strawberries visible to the robot's camera, within an acceptable margin of error.

Refer to the official assessment brief provided on Blackboard for the full task descriptions, marking criteria and submission requirements. The scripts/trilobot/ and trilobot_simulator/ folders in this repository provide starting points and practice tools only.

Connecting to the real Trilobot (laboratory PCs)

1. Start the robot and find its details

1. Connect the USB-C cable to the power bank. The Raspberry Pi will turn on and automatically connect to the UoL – Research Wifi network. Allow about one minute for it to start up and connect.
2. Find the sticker on the bottom of the robot and note its ID (for example, TRILO-05). Use the table below to look up that robot's username, IP address and password. These details may have changed; if they do not work, ask a delivery team member for help.

Robot ID	Username	IP address	Password
TRILO-11	pi11	10.247.26.13	Name1234
TRILO-13	pi13	10.247.26.140	Name1234
TRILO-07	pi07	10.247.26.16	Name1234
TRILO-06	pi06	10.247.26.145	Name1234
TRILO-05	pi05	10.247.26.114	Name1234
TRILO-02	pi02	10.247.26.141	Name1234
TRILO-12	pi12	10.247.26.142	Name1234
TRILO-04	pi04	10.247.26.143	Name1234

2. Sign in to a laboratory PC

1. Turn on the lab PC, select Windows as the operating system, and sign in using the password SEPS.
2. Use WinSCP to transfer files between the PC and the robot. It is already installed on the lab PCs. In WinSCP's Login window, enter the robot's IP address in Host name, and enter its username and password from the table above.
3. After connecting, you can browse the robot's folders and copy files, folders or Python scripts between the PC and the Raspberry Pi.

3. Connect and run Python scripts over SSH

1. Open Visual Studio Code (VS Code), which is already installed on the lab PCs, and open a new terminal.

2. Check that the robot is reachable by running (replace <ip address> with the robot's IP address):

   ping <ip address>
   

   If you see replies saying that bytes of data were received, the robot is reachable. If there are no replies, ask a delivery team member for help with the network connection.

3. If the robot is reachable, connect to it with SSH (replace the placeholders with the robot's username and IP address):

   ssh <username>@<ip address>
   

4. The first time you connect, confirm that you trust the remote host. Then enter the robot's password when prompted.

5. Once connected, you can navigate the Raspberry Pi's folders and run Python scripts in the terminal. The scripts run on the robot's Raspberry Pi, not on the lab PC. You can also use WinSCP to browse and transfer files.