Getting started with a self-driving RC car
summary
Hands-on guide to building an RC self driver using Donkeycar hardware and Python software with a camera, a PCA9685 PWM board for steering and throttle, and either a Raspberry Pi or Jetson Nano for compute. Uses supervised behavioral cloning with 160×120 RGB frames at 20 Hz, aims for ≥30 Hz inference, and gets ~60 Hz on Nano with TensorRT but only ~16 Hz for large backbones. Includes data cleaning, augmentation, model diagnostics, and race-proven rigs that do work when not run with the lens cap on.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:18What you need
- 6:13The drive loop
- 10:03The autopilot
- 18:00Driving the car
- 20:14Software
- 26:31Machine Learning
- 29:32Jetson Nano
- 31:05Computer Vision
- 33:20Image Processing
- 34:06Outro
related talks
49mBuilding a self driving RC car
35mDingoCar: Perception module Small scale, self-driving car
21mAPP4MC RaceCar: A Practical ADAS Demonstrator for Evaluating and Verifying Timing Behavior
45mTutorial: Building a Drone, From Scratch
40mBuilding an Autonomous Robotic Vehicle
31mAutonomous driving, how Python drives Formula Racing cars
05mInterview with
45m(Deep) Learning to Fly