ub-clean-test 0.26.0

机器人控制框架

# ub-clean-test

[中文版](#chinese) | [English](#english)

---

## <a id="chinese"></a>中文版

一个模块化的机器人控制框架。

### 功能
- **基础控制**：PID、LQR、自适应控制
- **运动学**：逆运动学（速度级、位置级）
- **轨迹规划**：五次多项式平滑运动
- **路径规划**：RRT 快速扩展随机树
- **状态估计**：卡尔曼滤波
- **强化学习**：PPO 近端策略优化
- **自适应控制可视化**：（0.11.0 新增） 动态展示参数调整过程
- **线性倒立摆 (LIPM)**：双足机器人步态规划基础 (0.12.0 新增)
- **零力矩点 (ZMP)**：双足机器人稳定性判断 (0.13.0 新增)
- **模型预测控制 (MPC)**：步态预测与优化控制 (0.14.0 新增)
- **全身控制 (WBC)**：多任务协调（走路 + 上肢动作）(0.15.0 新增)
- **视觉-语言-动作 (VLA)**：听懂简单指令 (0.16.0 新增)
- **视觉感知 (Vision)**：检测物体颜色和位置 (0.17.0 新增)
- **任务规划 (Task Planning)**：听懂→看懂→行动 (0.18.1 新增)
- **经验记忆 (Experience Memory)**：记住成功经验，下次秒执行 (0.19.0 新增)
- **自动调参 (Auto-Tuning)**：自己找最优控制参数 (0.20.0 新增)
- **状态预测 (State Prediction)**：预判未来轨迹，提前避险 (0.21.0 新增)
- **安全规则库 (Safety Rules)**：定义机器人不能做的事 (0.22.0 新增)
- **自检系统 (Self-Diagnosis)**：检查自身状态，知道能不能干活 (0.23.0 新增)
- **故障恢复 (Fault Recovery)**：检测故障并尝试自动修复 (0.24.0 新增)
- **因果推理 (Causal Reasoning)**：理解“为什么”，做出更聪明的决策 (0.25.0 新增)
- **长期记忆 (Long-Term Memory)**：记住重要事件，重启也不丢 (0.26.0 新增)

### 安装

pip install ub-clean-test

### 快速开始
```python
from ultra_balance import create_robot
robot = create_robot('two_wheel')
robot.start()

示例

• demo_trajectory.py：轨迹规划演示
• demo_rrt.py：RRT 路径规划演示
• demo_ppo.py：PPO 强化学习演示
• demo_kalman.py：卡尔曼滤波演示
• demo_adaptive.py：自适应控制演示（0.11.0 新增）
• demo_lipm.py：线性倒立摆演示 (0.12.0 新增)
• demo_zmp.py：ZMP 稳定性演示 (0.13.0 新增)
• demo_mpc.py：MPC 控制演示 (0.14.0 新增)
• demo_wbc.py：全身控制演示 (0.15.0 新增)
• demo_vla.py：VLA 演示 (0.16.0 新增)
• demo_vision.py：视觉感知演示 (0.17.0 新增)
• demo_task.py：任务规划演示 (0.18.1 新增)
• demo_experience.py：经验记忆演示 (0.19.0 新增)
• demo_autotune.py：自动调参演示 (0.20.0 新增)
• demo_predict.py：状态预测演示 (0.21.0 新增)
• demo_safety.py：安全规则演示 (0.22.0 新增)
• demo_self_diagnosis.py：自检系统演示 (0.23.0 新增)
• demo_fault_recovery.py：故障恢复演示 (0.24.0 新增)
• demo_causal.py：因果推理演示 (0.25.0 新增)
• demo_long_term.py：长期记忆演示 (0.26.0 新增)

---

English Version

A modular robot control framework.

Features

• Control: PID, LQR, adaptive control
• Kinematics: Forward/Inverse kinematics
• Trajectory Planning: Quintic polynomial
• Path Planning: RRT (Rapidly-exploring Random Tree)
• State Estimation: Kalman filter
• Reinforcement Learning: PPO (Proximal Policy Optimization)
• Adaptive Control Visualization: (new in 0.11.0) Dynamic visualization of gain adaptation
• Linear Inverted Pendulum Model (LIPM): Foundation for bipedal robot gait planning (new in 0.12.0)
• Zero Moment Point (ZMP): Bipedal robot stability criterion (new in 0.13.0)
• Model Predictive Control (MPC): Gait prediction and optimization (new in 0.14.0)
• Whole Body Control (WBC): Multi-task coordination (walk + arm) (new in 0.15.0)
• Visual-Language-Action (VLA): Understand simple commands (new in 0.16.0)
• Vision Perception: Detect object color and position (new in 0.17.0)
• Task Planning: Understand → See → Act (new in 0.18.1)
• Experience Memory: Remember successful tasks, execute instantly next time (new in 0.19.0)
• Auto-Tuning: Find optimal control parameters automatically (new in 0.20.0)
• State Prediction: Predict future trajectory, avoid danger in advance (new in 0.21.0)
• Safety Rules: Define what the robot cannot do (new in 0.22.0)
• Self-Diagnosis: Check own status, know if ready to work (new in 0.23.0)
• Fault Recovery: Detect faults and try to recover automatically (new in 0.24.0)
• Causal Reasoning: Understand "why", make smarter decisions (new in 0.25.0)
• Long-Term Memory: Remember important events, survive restart (new in 0.26.0)

Installation

pip install ub-clean-test

Quick Start

from ultra_balance import create_robot
robot = create_robot('two_wheel')
robot.start()

Modules

· core/: Core algorithms (Kalman filter, PID, etc.) · planning/: Path planning (A*, RRT) · learning/: Reinforcement learning (PPO)

Examples

• demo_trajectory.py: Trajectory planning demo
• demo_rrt.py: RRT path planning demo
• demo_ppo.py: PPO reinforcement learning demo
• demo_kalman.py: Kalman filter demo
• demo_adaptive.py: Adaptive control demo (new in 0.11.0)
• demo_lipm.py: Linear Inverted Pendulum demo (new in 0.12.0)
• demo_zmp.py: ZMP stability demo (new in 0.13.0)
• demo_mpc.py: MPC control demo (new in 0.14.0)
• demo_wbc.py: Whole Body Control demo (new in 0.15.0)
• demo_vla.py: VLA demo (new in 0.16.0)
• demo_vision.py: Vision perception demo (new in 0.17.0)
• demo_task.py: Task planning demo (new in 0.18.1)
• demo_experience.py: Experience memory demo (new in 0.19.0)
• demo_autotune.py: Auto-tuning demo (new in 0.20.0)
• demo_predict.py: State prediction demo (new in 0.21.0)
• demo_safety.py: Safety rules demo (new in 0.22.0)
• demo_self_diagnosis.py: Self-diagnosis demo (new in 0.23.0)
• demo_fault_recovery.py: Fault recovery demo (new in 0.24.0)
• demo_causal.py: Causal reasoning demo (new in 0.25.0)
• demo_long_term.py: Long-term memory demo (new in 0.26.0)