Electrical Engineering

Robot Electrical Architecture (36 DoF)

System Overview

This robot is a 36 DoF humanoid platform powered by a 12.6V (3S4P) battery system.
The design separates high-current servo power from logic + sensor control, using an ESP32 + Raspberry Pi hybrid architecture.

Degrees of Freedom

DoF Distribution

  • Legs: 4 DoF × 2 = 8

  • Arms: 4 DoF × 2 = 8

  • Hands: 8 DoF × 2 = 16

  • Neck: 2 DoF

  • Torso: 2 DoF

Total: 36 DoF

Power Distribution

Main Flow

  • Battery (12.6V) → XT60 → Fuse → Split

Then:

  • → Buck (8V) → Legs + Arms

  • → Buck (5V) → Hands + Neck

  • → Direct 12V → Torso servos

PCB and Wiring Design

Board Layout

The custom PCB is designed to handle power distribution, signal routing, and component integration within a single compact system.

The layout is divided into:

  • controller regions

  • sensor interfaces

  • servo driver connections

  • power distribution paths

Controllers

The system uses a dual-layer control architecture:

  • Raspberry Pi 5

    • Handles AI processing, vision, and high-level decision-making

  • ESP32 Microcontrollers

    • Handle real-time control

    • Read sensor data

    • Manage fast I/O operations

Communication between controllers is handled over UART.

Servo Control System

Servo actuation is managed using multiple PCA9685 PWM driver boards, connected via I2C.

  • Generates precise PWM signals for all 36 servos

  • Allows scalable expansion without overloading the main controller

I2C Expansion

An I2C multiplexer (TCA) is used to manage multiple devices on the same bus.

This enables:

  • simultaneous communication with multiple PCA boards

  • integration of sensors with overlapping I2C addresses

Sensor Integration

IMU

An inertial measurement unit provides:

  • orientation data

  • balance feedback for walking and stabilization

FSR Sensors

Force-sensitive resistors are placed in the feet to detect:

  • ground contact

  • weight distribution

Each FSR is implemented using a voltage divider circuit and read through ESP32 analog inputs.

Signal Flow

Sensor data is processed and converted into actuation commands through the following pipeline:

Sensors → ESP32 → Raspberry Pi → PCA9685 → Servos

Design Highlights

  • Separation of high-level processing and real-time control

  • Scalable servo architecture using PWM driver boards

  • Robust I2C communication with multiplexer support

  • Dedicated sensor processing for stability and interaction

  • Multi-voltage power system optimized for performance and reliability

Summary

The system is designed as a high-current, modular electrical architecture for a full humanoid robot.
It supports 36 degrees of freedom, distributed control, and stable operation across multiple voltage domains while maintaining scalability for future expansion.