Abstract:

The design and manufacture of self-driving cars faces many challenges. Today, companies have widely used machine learning to find solutions. The goal of this project is to make a robot car drive automatically around the track based on Raspberry Pi. In this project, collect data of track images by using Raspberry Pi Camera. Then, build the end-to-end model after processing images. Last, load the model and realize auto-driving on the original track.

Key words: Machine learning Raspberry Pi Auto-driving End-to-end model

Matters needing attention:

• Do not use 4WD assemble car kit; The 4 motors in this kit can not be supplied by battery pack of 4 AA batteries. Instead, 2WD assemble car kit is the best choice.
• The track needs to be produced by itself, which is very important and determines the quality of data. (I type the black papers on the floor and then past into the shape of the runway)
• The width of the track is about two times that of the body.
• Take care of camera angel problem.

Future work:

• Handling lighting issues.
• Handling data category imbalances.