Open Source Projects on Github

Since the very beginning, Michael and I have been dedicated to contributing to the open source community. This Spring, we open sourced the entire repository for the self-driving golf cart project. In order to better host the code base, as well as to provide us with a platform for future research, we registered a Github organization account, Sigma A.I. The golf cart’s code will be hosted there, as well as several other interesting machine learning projects. I will introduce you to the current repositories.

Self-driving golf cart

As described above, you can always find the latest code for the autonomous golf cart project at Sigma A.I.’s Github page. We are actively maintaining the code and looking for contributions.

Behavioral Cloning Research

If you are looking for some code for behavioral cloning, I published some research code on Github. For the past month, I have been experimenting with different models and datasets to improve the golf cart’s steering performance. Want to learn more about this interesting algorithm? check out my previous posts. Please note that the project is still under development. (I will write a post on that soon)

Semantic Segmentation

There is a multi-class, real-time semantic segmentation repo on Sigma A.I.’s Github as well! Including the pre-trained weights! Our (crudely) trained results scored 89% on the cityscape dataset. Now, I am working on combining several datasets to boost the training results.

ROS GPS

Last but not least, if you are trying to use GPS for some basic localization with ROS, plus visualization with OpenStreetMaps, this repo is perfect for you. This project includes a serial GPS interface with ROS, rviz with OpenStreetMaps datapoints, as well as real-time visualization updates as the object (golf cart) is moving. Overall, this should serve as a great starting point for a simple ROS localization project.

Conclusion

I hope you will find these projects useful. The brevity of my descriptions might not be sufficient to fully explain what each project entails. If you have any questions, comments or concerns, please email me at contact@neilnie.com. Do you have cool projects to share? Please leave a comment below. If you enjoyed this post, please give me a like. You might like the following posts as well.

1. Deep learning steering prediction
   1. Visualizing the Steering Model with Attention Maps
   2. Successfully Tested the Autonomous Steering System for the Self-Driving Golf Cart
   3. Predicting Steering Angle with Deep Learning — Part 2
   4. Predicting Steering Angle with Deep Learning — Part 1
2. Semantic segmentation
   1. The Robustness of the Semantic Segmentation Network
   2. Autonomous Cruise Control System
   3. Understanding the World Through Semantic Segmentation
3. Robot Operating System
   1. Hello, ROS
   2. Open Street Map with ROS
   3. Self-Driving Software + Carla Simulator
   4. GPS Localization with ROS, rviz, and OSM