Self-driving car experience

Part 1: Computer Vision and Deep Learning

  • Finding lane lines
    • Workflow for lane detection pipeline:
      • Load video -> Greyscale transform ->Gaussian smoothing -> Canny edge detection -> select the region of interest -> Hough transform -> Draw lines on original video
Web resource (less noise)
My own recorded video: drive back from office to home (more noise)

  • Traffic sign classifier
    • Datasets: German Traffic Sign Recognition Benchmark
    • Three layers neural network including a fully connected layer
    • Test accuracy: 0.935

  • Advanced lane finding
    • Workflow:
      • Compute the camera calibration matrix and distortion coefficients given a set of chessboard images.
      • Apply a distortion correction to raw images.
      • Use color transforms, gradients, etc., to create a thresholded binary image.
      • Apply a perspective transform to rectify binary image (“birds-eye view”).
      • Detect lane pixels and fit to find the lane boundary.
      • Determine the curvature of the lane and vehicle position with respect to center.
      • Warp the detected lane boundaries back onto the original image.
      • Output visual display of the lane boundaries and numerical estimation of lane curvature and vehicle position.

  1. Vehicle detection and tracking

Part 2: Sensor fusion, Localization, and control

Extended Kalman filter

  1. Uncented Kalman Filter
  2. Kidnapped vehicle
  3. PID controller

  1. Model predictive control

Part 3: Path planning, concentrations, and system integration

  1. Path planning
  2. Semantic segmentation
  3. Functional safety
  4. Systemintegration

Close Bitnami banner