Faculy of Engineering, Alexandria University Computer Systems and Control Project.

---

AutoPark-PilotAssist

An innovative and open-source self-parking system that incorporates Advanced Driver Assistance System (ADAS) features, including Autonomous Parking, Pilot Assistant, and Collision Avoidance. It combines hardware components such as Raspberry Pi 3B, compass , cameras, ultrasonic sensors, and ESP32 with FreeRTOS for efficient multitasking to create a comprehensive autonomous parking solution.

The project's final version consists of two main components:

Auto Park:

• The camera detects empty parking spaces using an OpenCV script.
• The parking node initializes, and the robot autonomously parks in the identified space.
• FreeRTOS tasks manage ultrasonic sensor readings and IMU updates for precise navigation and sends real-time sensor readings to the Raspberry Pi.

Pilot Assist:

• Utilizing a phone camera through TCP/IP with a laptop, the YOLOv8 Nano model is deployed.
• A Python script initializes the node to send the vehicle's position using PID controller and ultrasonic readings.
• The system follows the target vehicle, providing a pilot assist feature for smooth navigation.

---

• Scripts:
  • AutoParking.py
  • PilotAssist.py
  • VehicleDetection.py
  • ParkingDetection.py
  • ESP32_freeRTOS.ino
  • MPU_UnitTest.ino
  • Ultrasonic_UnitTest.ino

Hardware System & PCB

Previous Versions

I incorporated unit testing and adopted Agile software development practices throughout the project.

• Version 0: Manual Phase Park Assist

  • Initial testing utilizing the Bluetooth module HC-05 and L298n Motor Driver.

• Version 1: Sensor Calibration and init ROS Workspace

  • Focused on calibrating and unit testing sensors for accurate data collection.
  • Tested MPU6050, ultrasonic sensors, and encoders. The ESP32 received and transmitted sensor data through UART (USB) to the Raspberry Pi.
  • initialized the ROS workspace for abstraction and modular control, laying the groundwork for further development.

• Version 2: YOLOv8 Model, Auto Parking Algorithm

  • Deployed the YOLOv8 model for object detection of the vehicle, using a PID algorithm for autopilot implementation.
  • Implemented an auto-parking script with collision avoidance features, significantly enhancing the vehicle's autonomous capabilities.
  • developed freeRTOS to manage all sensors in real-time data for ESP32, ensuring efficient multitasking and timely data processing.