🤖 MBt2: High-Performance Self-Balancing Robot

FreeRTOS Dual-Core • SimpleFOC Direct Torque Control • Bluepad32 Bluetooth • Real-Time Web Dashboard

▶️ Watch MBt2 in Action

---

🌟 Overview

MBt2 is a modular, high-performance 2-wheeled self-balancing robot powered by an ESP32, SimpleFOC field-oriented motor control, AS5600 magnetic encoders, an MPU6050 IMU, and a Bluepad32 Bluetooth gamepad receiver.

It features a custom FreeRTOS Dual-Core Architecture, Cascaded Dual-Loop PID Control, and a Real-Time WebSocket Web Dashboard for live tuning, 3D orientation visualization, diagnostic metrics, and multi-channel telemetry plotting.

---

📸 System Showcase

🖥️ Real-Time WebSocket Web Dashboard & Tuning Suite

---

🚀 Key Features & Engineering Highlights

• ⚡ FreeRTOS Dual-Core Allocation:
  • Core 1 (Priority 10): Runs SimpleFOC motor control (loopFOC() and move()) at >20 kHz + 100Hz synchronized IMU sampling and Cascaded Dual-Loop PID execution.
  • Core 0 (Normal Priority): Runs Bluepad32 Bluetooth Gamepad processing, AS5600 diagnostic polling, JSON telemetry serialization, and WebSocket command handling.
• 🔒 Multi-Core I2C Mutex Thread Safety:
  • A FreeRTOS Mutex Semaphore (i2c0Mutex) protects shared I2C Bus 0 (GPIO 21 SDA / GPIO 22 SCL) between MPU6050 (0x68) and Left AS5600 (0x36), eliminating SDA/SCL bus collisions and data corruption during closed-loop FOC.
• ⚖️ Cascaded Dual-Loop Direct Torque Control:
  • Inner Stability Loop (pid_stb): Fast $P, I, D$ loop converting tilt angle error to direct motor voltage / torque ($U_q$).
  • Outer Velocity Loop (pid_vel): Measures net physical forward wheel speed (motorLeft.shaft_velocity - motorRight.shaft_velocity) / 2.0f and generates a target pitch angle (target_pitch) filtered through lpf_pitch_cmd ($\text{Tf} = 0.07\text{s}$) to hold position.
• 🎮 Bluepad32 Bluetooth Gamepad Remote Control:
  • Pair any Bluetooth gamepad (PS4, PS5, Xbox, Switch Pro, 8BitDo).
  • Left Stick Y: Throttle (Forward / Reverse).
  • Right Stick X: Steering (Turn Left / Right).
  • Filtered through lpf_throttle ($\text{Tf} = 0.4\text{s}$) and lpf_steering ($\text{Tf} = 0.1\text{s}$) for smooth acceleration.
• 🛡️ 9.0V Torque Limit & Max Lean Angle Clamp:
  • 9.0V motor voltage headroom (on a 12V supply) provides 50% more peak acceleration torque.
  • target_pitch is clamped to $\pm 8.6^\circ$ ($\pm 0.15\text{ rad}$), preventing joystick commands from tipping the robot beyond physical recovery limits.
• ⚖️ Auto-Enable Balancing:
  • Holding the robot upright near vertical zero ($\pm 3.0^\circ$) for 1 second automatically engages balancing mode (MotionControlType::torque). Includes an automatic 45° Tipped-Over Safety Cutoff.
• 💾 Browser localStorage Persistence:
  • Tuning parameters set on the web GUI are saved locally in the browser and automatically restored on page refresh.

---

🔌 Hardware Architecture & Pinout

Pin Mapping Table

Device	ESP32 Pin	Function / Protocol
MPU6050 IMU	GPIO 21 (SDA), GPIO 22 (SCL)	I2C Bus 0 (Shared with Mutex)
Left AS5600 Encoder	GPIO 21 (SDA), GPIO 22 (SCL)	I2C Bus 0 (Shared with Mutex)
Right AS5600 Encoder	GPIO 18 (SDA), GPIO 19 (SCL)	I2C Bus 1
Left Motor Driver	GPIO 32, 33, 14 (PWM) | GPIO 13 (EN)	3-Phase Sinusoidal PWM
Right Motor Driver	GPIO 25, 26, 27 (PWM) | GPIO 12 (EN)	3-Phase Sinusoidal PWM
Sensor Power Rail	3V3 Pin	3.3V Logic Protection

---

🧊 3D Mechanical Design & PCB Manufacturing

🧊 3D Chassis (Multiboard System)	⚡ Custom PCB Carrier Board
	📄 View PCB Drawings & Gerbers
Modular 3D printed hex grid tiles, motor brackets & wheels	2-Layer custom carrier board with 1-click JLCPCB / PCBWay order ZIP

• 3D CAD Model: 👉 View & Edit Multiboard Bot on Tinkercad
• Hardware Gerber ZIP: PCB_Design/MBt2_PCB.zip (1-click manufacturing order file)
• Schematic PDF: PCB_Design/MBt2_PCB.pdf
• Bill of Materials: 📋 PCB_Design/BOM_MBt2_Carrier_Board.csv

📋 Bill of Materials (BOM) Quick Summary

Component	Description	Qty	Target Reference
ESP32 DevKit V1	30-Pin Dual-Core Microcontroller	1	U1
SimpleFOC Mini Driver	3-Phase BLDC Motor Driver	2	L_Driver, R_Driver
BLDC Gimbal Motor	11 Pole Pair Brushless Motor	2	Left & Right Wheels
AS5600 Magnetic Encoder	12-Bit Angle Sensor (I2C)	2	L_Encoder, R_Encoder
MPU6050 IMU	6-DOF Accelerometer & Gyroscope	1	IMU
5V Step-Up Boost Module	1S LiPo $\rightarrow$ 5V Regulator	1	5V_Boost
1S LiPo / 18650 Battery	3.7V High Discharge Battery	1	1s_BAT

---

⚡ Quick Start & Usage

1. Flash ESP32 Firmware

cd esp32_firmware
fuser -k /dev/ttyUSB0 || true
platformio run -t upload

2. Launch Tuning Dashboard

node tuning_server.cjs

Open http://localhost:8080/ in your web browser.

---

🎯 Official Baseline Tuned Parameters

• Inner Stability Loop: Stb P = 80.0, Stb I = 67.0, Stb D = 0.8
• Outer Velocity Loop: Vel P = 0.012, Vel I = 0.010
• LPF Noise Filter: LPF Tf = 0.07 s
• Equilibrium Pitch Trim: 1.83°

---

🔖 Version History

Tag / Milestone	Description
v1.0-single-core	Legacy single-core implementation running all tasks in loop() on Core 1.
v2.0-dual-core-unified	Current State: FreeRTOS Dual-Core architecture, Bus 0 I2C Mutex, 100Hz Synchronized Cascaded Dual-Loop PID, Bluepad32 Bluetooth Gamepad, Auto-Balancing, 9V Torque Headroom, and Unified Tuning Dashboard.

---

📜 License

Hardware: CERN-OHL-S v2 | Software: MIT License