Self-Balancing Robot

Hero image: AI-generated illustrative render — not a photo of a real build. See the Attribution section below for design references and upstream photos.

A 3D-printable two-wheeled self-balancing robot — the classic inverted-pendulum control theory demo you can actually build and drive from a browser. The robot stays upright using a PID control loop fed by an MPU-6050 IMU, with two N20 DC gear motors driven through an L298N motor driver on a Raspberry Pi Zero 2 W.

What makes this interesting: PID tuning happens live in the browser — drag the P, I, and D sliders while the robot runs and watch it respond. Abstract control theory becomes tactile and immediate.

What It Does

• Balances on two wheels using a PID loop (proportional-integral-derivative controller)
• Accepts forward/backward/turn commands from the browser while it actively balances
• Live telemetry: real-time angle, motor speeds, and PID terms visible in the control panel
• PID gain sliders: adjust P, I, and D from the browser with immediate effect
• Falls gracefully: motors cut when tilt exceeds 45° (safe-stop threshold)

Hardware

• Raspberry Pi Zero 2 W (runs Python PID loop + orobot firmware)
• MPU-6050 IMU on I2C bus — reads tilt angle at ~200 Hz via complementary filter
• L298N dual H-bridge motor driver — PWM drive to two N20 6V DC gear motors
• 2× N20 DC gear motor with wheel mounts and 3D-printed wheels
• LiPo 7.4V 2S battery pack (or 2× 18650 in series)
• 3D-printed chassis — Thingiverse model #1069256 or compatible

Build Overview

1. Print the chassis (PLA, 0.4mm nozzle, 20% infill). Total print time ~3–5h.
2. Mount motors and wheels in the chassis motor pockets.
3. Wire MPU-6050 to Pi GPIO I2C (SDA=BCM 2, SCL=BCM 3); power from 3.3V.
4. Wire L298N: IN1=BCM 17, IN2=BCM 27 (left), IN3=BCM 22, IN4=BCM 23 (right); ENA/ENB=BCM 12/13 (PWM).
5. Install Pi firmware with smbus2 and RPi.GPIO libraries.
6. Set balance_angle (the equilibrium tilt offset for your specific build — usually ±3° from level).
7. Connect in orobot, deploy this program, and start tuning PID gains from the browser.

Tuning Guide

Start with D=0, I=0, and increase P until the robot starts to oscillate. Then back P off 20%, add a little D to damp oscillation, and finally a small I to correct steady-state lean. Typical starting values: P=80, I=5, D=2 for N20 motors at ~6V. Your mechanical build will differ — expect to spend 30 minutes tuning.

Honest Expectations

The balance loop runs on the Pi at ~50 Hz. If the Pi is busy (e.g., SD card writes, heavy CPU), the loop rate drops and the robot falls. Use a fast SD card or run from a USB drive. The orobot browser interface adds a few hundred milliseconds of latency for manual drive commands, which is fine — the Pi handles the balance loop autonomously; browser commands are just setpoint offsets.

Attribution

This program is an original orobot design guide for the self-balancing robot class. Hardware inspiration from the B-ROBOT by JJRobots (GPL-2.0, https://github.com/jjrobots/B-ROBOT) and the 3D-printable Arduino Nano 33 BLE design featured on Hackaday (December 2022, https://hackaday.com/2022/12/27/3d-printed-self-balancing-robot-brings-control-theory-to-life/). The orobot control code is original (MIT).