Almost there…
I started the Udacity Self-Driving Nanodegree (Udacity SDCND) Program last year December. It will come to end this month. I have to say it looks more achievable now than it did 3 months ago. I had the thought of quitting many times than I can count but it has been worth staying with it till the end. It has been really challenging for me, because I had to learn a ton of new stuffs while also combining it with my PhD program in Tohoku University. Phew!
The program involves 3 terms including 12 projects and a final, capstone project. The following are the projects I needed to complete in order to fully meet the passing requirements.
- Finding Lane Lines on the Road
- Traffic Sign Classifier
- Behavioral Cloning
- Advanced Lane Finding
- Vehicle Detection and Tracking
- Extended Kalman Filters
- Unscented Kalman Filters
- Kidnapped Vehicle
- PID Controller
- Model Predictive Control
- Path Planning Project
- Semantic Segmentation
- Capstone (with a group of 5 other SDCND students): Programming a Real Self-Driving Car
And yeah, they ALL looked difficult at the first look. The most important lesson for me is that, “There is really no better way to learn than to do”.
The most important lesson for me is that, “There is really no better way to learn than to do”.
The most difficult project for me was the Path Planning Project. It took me that longest to finish and really stretched my C++ coding ability (still not strong here though). See a video of the result below:
If you would like to see the projects, please check my github profile. All project codes are available freely.
Cheers to seeing the finish line.
