Semi-autonomous lane keeping system

Semi-autonomous lane keeping system based on C++ & OpenCV 4.x.
Project for my engineering thesis at AGH University of Science and Technology [WIMiIP].

Building

Linux: ./build.sh
Windows: Using Visual Studio 2019

Depencencies

Linux: ./build.sh install
Windows: Download OpenCV in version 4.1.1 and extract it to C:\opencv then copy all files from C:\opencv\build\x64\vc15\bin to VisualStudio\x64\Debug and Release directories

Progress history

Task & Commit	Date
Finished readme	2020-03-18
Build tools	2020-03-20
Lane detection	2020-03-20
Hysteresis	2020-03-21

To do

• Software
  • Build tools
  • Road lane recognition
    • Turn prediction
    • Hysteresis
    • Optimization
      • Multi-threading
      • ARM NEON optimizations
    • PID controller for steering angle calculation
• Hardware
  • Opel EPS-4400 reverse engineering
    • Gather all necessary datasheets
    • Board schematics based on PCB
    • Inject own steering signals
    • Prevent triggering DTCs (Diagnostic Trouble Codes)
  • Raspberry Pi 4 <-> Power steering
    • Communication
• Safety measures
  • Software
  • Hardware

References

General concept
https://medium.com/@mrhwick/simple-lane-detection-with-opencv-bfeb6ae54ec0
https://medium.com/pharos-production/road-lane-recognition-with-opencv-and-ios-a892a3ab635c

Use of Gaussian blur & simple turn prediction
https://github.com/MichiMaestre/Lane-Detection-for-Autonomous-Cars

PID Controller
https://en.wikipedia.org/wiki/PID_controller
https://www.youtube.com/watch?v=4Y7zG48uHRo