A distributed 3-node bionic robot system with 2-DOF gimbal and emotional OLED display.
A low-cost, high-performance distributed bionic interactive robot powered by three BBC Micro:bits. I independently designed this full-stack system to translate human hand gestures into real-time, perfectly synchronized mechanical movements and dynamic facial expressions.
"Curing cheap hardware with smart algorithms." — The core philosophy of this solo project is to use mathematical optimization (Expo curves, Low-Pass Filters, Lerp animations) to make budget MG90S servos perform with industrial-grade smoothness.
- Algorithm-Driven Stabilization: Implements RC-grade Expo mixing curves and First-Order Low-Pass Filters (LPF) to completely eliminate hardware jitter.
- Zero-Latency 3D Tracking: Uses
math.atan2for full 360-degree spatial gravity vector calculation, avoiding gimbal lock and value inversion. - Bionic OLED Morphing Engine: A custom 6-parameter rendering engine replaces static images. Features organic smooth transitions (Lerp), random bionic Saccades (eye twitches), and rapid blinking logic.
- One-Click Calibration: Instant zero-point baseline setup with a single button press.
A distributed 3-node architecture ensuring zero-latency processing via high-frequency Radio sync:
- Node B (The Brain): Handheld sensor. Calculates 3D vectors and applies LPF/Expo filters.
- Node A (The Body): 2-DOF Gimbal. Drives physical movement using a custom Silent PWM logic to prevent electrical humming.
- Node C (The Face): OLED Display. Independently computes and renders morphological expressions.
This project consists of three independent nodes, each with a specific power strategy to balance high-current demands and portability.
| Component | Qty | Notes |
|---|---|---|
| BBC Micro:bit | 3 | Main controllers for Node A, B, and C (V1/V2). |
| IO Expansion Board | 2 | For Node A (Servos) and Node C (OLED). |
| MG90S Servo | 2 | Metal gear servos for the 2-DOF gimbal (Node A). |
| SSD1306 OLED Display | 1 | 0.96" I2C interface (128x64) for Node C. |
| USB Cable | 1 | Power source for Node A (High current for servos). |
| 3.7V LiPo Battery | 1 | 500mAh JST-connector battery for Node B. |
| 3V AAA Battery Box | 1 | 2-cell battery holder for Node C. |
| Dupont Wires | 1 Set | Female-to-Female jumpers for I2C and servos. |
- Power: Powered via USB cable (for stable servo operation).
- Servos: X-axis (Pitch) to P0, Y-axis (Roll) to P1 on the expansion board.
- Power: 3.7V LiPo battery plugged into the Micro:bit JST port.
- Wiring: No external wiring required (uses built-in accelerometer).
- Power: 3V (2x AAA) battery box connected to expansion board pins.
- OLED: SCL to P19, SDA to P20, VCC to 3V3, GND to GND.
Microbit-Distributed-Gimbal/
├── node_a_actuator.py # Node A: Actuator (Silent PWM & Motion Smoothing)
├── node_b_sensor.py # Node B: Sensor (3D Vector Calc & Expo Mixing)
├── node_c_display.py # Node C: Display (6-Param Morphological Engine)
├── ssd1306.py # Driver: Pure Python OLED Library
├── demo.gif # Project demonstration animation
├── README.md # Project documentation (You are here)
└── LICENSE # MIT License
Flash the respective code files to three separate Micro:bit boards (Note: Node C also requires the ssd1306.py driver to be flashed together).
- Sync Channel: Ensure the radio group is identical across all three codes (default is
group=22). - Zero Calibration: After powering on, hold the Node B sensor horizontally and press Button B for a one-click zero calibration.
- Interact: Tilt the handheld sensor and enjoy the ultra-smooth bionic interaction!
- XIONG Shenli (Solo Developer)
- Role: Full-stack development including hardware architecture, 3D tracking algorithms, and OLED rendering engine.
这是一个基于 3 个 Micro:bit 节点开发的分布式仿生交互机器人。我独立设计了整个系统的软硬件架构，将人类的手势实时转化为极其丝滑的机械运动与生动的表情变化。
"用软件算法治愈廉价硬件的痛。" —— 本项目的核心理念是：如何通过数学算法优化（指数曲线、低通滤波、线性插值），让几十块钱的廉价舵机跑出工业级的细腻质感。
- 极致平滑的运动算法: 引入 RC 遥控级别的高阶指数混控曲线 (Expo) 与一阶低通滤波 (LPF)，彻底过滤人手生理抖动与硬件机械卡顿。
- 3D 向量空间追踪: 弃用简单的角度读取，采用
math.atan2进行全空间重力向量计算，告别万向节死锁与数值翻转。 - 6 参数形态学表情引擎: 从零编写了一个 6 参数矩阵渲染引擎取代静态图片。包含插值动画过渡 (Lerp)、模拟生物的微跳视 (Saccades) 与随机眨眼逻辑，赋予机器独立意识感。
- 一键校准: 按下单键即可瞬间建立空间零点基准。
系统分为三个独立节点，通过无线电 (Radio) 进行高频同步通讯：
- Node B (传感器端): 手持感应，负责 3D 信号采集与核心滤波计算。
- Node A (云台执行端): 驱动 2 自由度云台，采用 Silent PWM 逻辑彻底消除电机啸叫。
- Node C (视觉渲染端): 搭载 0.96" OLED 屏幕，独立进行表情的矩阵变换与渲染。
本项目由三个独立的节点组成，每个节点根据实际功耗和便携性需求采用了不同的供电方案。
| 零件名称 | 数量 | 规格/备注 |
|---|---|---|
| BBC Micro:bit | 3 | 主控板，Node A、B、C 各使用一片（V1/V2均可）。 |
| IO 扩展板 | 2 | 用于 Node A（连接舵机）和 Node C（连接屏幕）。 |
| MG90S 金属舵机 | 2 | 用于构建 Node A 的两自由度云台。 |
| SSD1306 OLED 屏幕 | 1 | 0.96寸 I2C 接口，用于 Node C 的视觉渲染。 |
| USB 数据线 | 1 | 用于 Node A 供电，确保舵机转动时电流稳定。 |
| 3.7V 锂电池 | 1 | 500mAh JST 接口，用于 Node B，实现极致便携。 |
| 3V 七号电池盒 | 1 | 两节 AAA 电池，为 Node C 屏幕端独立供电。 |
| 杜邦线 | 1 套 | 母对母跳线，用于连接 I2C 设备和舵机。 |
- 供电： 使用 USB 线连接至电脑或充电宝（舵机耗电较大，建议 USB 供电）。
- 舵机： X轴（俯仰）接扩展板 P0，Y轴（横滚）接扩展板 P1。
- 供电： 3.7V 锂电池直接插在 Micro:bit 背面的 JST 电池接口。
- 接线： 无需外部接线（使用主控板内置加速度计）。
- 供电： 3V 电池盒连接至扩展板的电源引脚。
- 屏幕： SCL 接 P19，SDA 接 P20，VCC 接 3V3，GND 接 GND。
Microbit-Distributed-Gimbal/
├── node_a_actuator.py # Node A: 云台端 (Silent PWM 与运动平滑)
├── node_b_sensor.py # Node B: 传感器端 (3D 向量计算与 Expo 混控)
├── node_c_display.py # Node C: 视觉端 (6 参数形态学渲染引擎)
├── ssd1306.py # 底层驱动: 纯 Python 版 OLED 驱动库
├── demo.gif # 项目演示动图
├── README.md # 项目说明文档与 BOM 清单
└── LICENSE # MIT 开源协议
将对应节点的代码分别烧录至三个 Micro:bit 主板中（注：Node C 还需要同时烧录 ssd1306.py 驱动文件）
- 统一频道： 确保三份代码中的无线电频道保持一致（默认
group=22）。 - 零点校准： 系统通电后，水平握住传感器 Node B，按下 Button B 进行一键零点校准。
- 开始互动： 倾斜手里的传感器，体验丝滑的仿生机器人联动！
- 熊珅黎 (XIONG Shenli) - 独立开发者
- 负责内容: 硬件架构搭建、全套平滑滤波算法开发、OLED 仿生渲染引擎设计。
If you find this project interesting, feel free to give it a ⭐ Star!