UR5e MuJoCo Simulation

A comprehensive simulation setup for the Universal Robots UR5e robot using MuJoCo physics engine.

Project Structure

ur5e_simulation/
├── ur5e_main.xml              # Main simulation XML file
├── assets/                    # Asset files
│   ├── ur5e/                 # UR5e robot assets
│   │   ├── ur5e_assets.xml   # Materials and mesh definitions
│   │   ├── ur5e_body.xml     # Robot body structure
│   │   ├── ur5e_actuators.xml # Actuator definitions
│   │   └── meshes/           # STL mesh files
│   └── common_arena/         # Environment assets
│       └── simple_plane.xml  # Ground plane
├── src/                      # Source code
│   ├── __init__.py          # Package initialization
│   ├── ur5e_simulator.py    # Main simulator class
│   ├── util.py              # Utility functions
│   ├── PID.py               # PID controller
│   └── ik_module.py         # Inverse kinematics utilities
├── demo_ur5e.py             # Demonstration script
├── test_setup.py            # Setup validation script
└── README.md                # This file

Features

Core Functionality

• Complete UR5e Model: Accurate 6-DOF UR5e robot with proper joint limits and dynamics
• MuJoCo Integration: Full integration with MuJoCo physics engine
• Interactive Viewer: Real-time 3D visualization with camera controls
• Forward Kinematics: Compute end-effector pose from joint angles
• Inverse Kinematics: Solve for joint angles given target end-effector pose
• Joint Control: Position and velocity control of individual joints

Simulation Components

• Physics Simulation: High-fidelity physics simulation with collision detection
• Visual Rendering: Real-time rendering with customizable materials and lighting
• Trajectory Planning: Tools for planning and executing robot trajectories
• PID Control: Built-in PID controller for joint position control

Installation

Prerequisites

1. Python 3.7+
2. MuJoCo: Install MuJoCo physics engine

   pip install mujoco

3. NumPy: For numerical computations

   pip install numpy

Setup

1. Clone or download this project
2. Ensure all mesh files are in the correct location (assets/ur5e/meshes/)
3. Run the setup test:

   python test_setup.py

Usage

Basic Usage

from src.ur5e_simulator import UR5eSimulator
import numpy as np

# Initialize simulator
ur5e = UR5eSimulator(
    xml_path='ur5e_main.xml',
    USE_VIEWER=True
)

# Set joint positions (6 joints in radians)
home_position = np.array([-1.5708, -1.5708, 1.5708, -1.5708, -1.5708, 0])
ur5e.set_joint_positions(home_position)

# Get end-effector pose
position, rotation = ur5e.get_end_effector_pose()

# Render simulation
for i in range(1000):
    ur5e.render()

Inverse Kinematics

# Define target position
target_pos = np.array([0.3, 0.3, 0.5])

# Solve IK
joint_solution = ur5e.inverse_kinematics(target_pos)

if joint_solution is not None:
    ur5e.set_joint_positions(joint_solution)

Running Demos

1. Setup Test: Validate installation

   python test_setup.py

2. Basic Demo: Run comprehensive demonstration

   python demo_ur5e.py

XML Configuration

Main Configuration (ur5e_main.xml)

• Simulation parameters (timestep, solver settings)
• Include statements for sub-components
• Default joint and material properties
• Keyframe definitions for common poses

Robot Definition (assets/ur5e/)

• ur5e_assets.xml: Material properties and mesh references
• ur5e_body.xml: Robot link structure and joint definitions
• ur5e_actuators.xml: Motor and control definitions

Environment (assets/common_arena/)

• Ground plane and environmental objects
• Lighting and camera definitions

Customization

Adding New End-Effectors

1. Add mesh files to assets/ur5e/meshes/
2. Update ur5e_body.xml with new link definitions
3. Modify end_effector_name in simulator class

Custom Control Strategies

# Example: Custom PID control
from src.PID import PID_ControllerClass

pid_controller = PID_ControllerClass(
    k_p=10.0, k_i=0.1, k_d=1.0, dim=6
)

# In control loop
target_joints = np.array([...])  # Target joint positions
current_joints = ur5e.get_joint_positions()
control_output = pid_controller.get_output(
    x_trgt=target_joints, 
    x_curr=current_joints, 
    t_curr=ur5e.get_simulation_time()
)
ur5e.step(ctrl=control_output)

Adding Sensors

Modify XML files to include:

• Force/torque sensors
• IMU sensors
• Camera sensors
• Proximity sensors

API Reference

UR5eSimulator Class

Core Methods

• __init__(xml_path, USE_VIEWER=True): Initialize simulator
• reset(): Reset simulation to initial state
• step(ctrl=None, nstep=1): Step simulation forward
• render(): Update visualization
• close_viewer(): Close simulation window

Kinematics

• get_joint_positions(): Get current joint angles
• set_joint_positions(positions): Set joint angles
• get_end_effector_pose(): Get end-effector position and orientation
• get_jacobian(body_name): Get Jacobian matrix
• inverse_kinematics(target_pos, target_rot=None): Solve IK

Utilities

• add_visual_marker(position, size, color): Add visual markers
• get_simulation_time(): Get current simulation time
• is_viewer_alive(): Check if viewer is active

Troubleshooting

Common Issues

1. ImportError: No module named 'mujoco'

   • Install MuJoCo: pip install mujoco

2. XML parsing errors

   • Check that all mesh files exist in assets/ur5e/meshes/
   • Verify XML file paths and syntax

3. Viewer not opening

   • Check display settings
   • Try running with USE_VIEWER=False for headless mode

4. IK not converging

   • Target position may be outside workspace
   • Adjust IK parameters (tolerance, max_iterations, step_size)
   • Check for joint limit violations

Performance Tips

• Use smaller timesteps for more accurate simulation
• Disable viewer for faster computation
• Use appropriate solver settings for your application

Contributing

Feel free to contribute by:

• Adding new robot models
• Implementing advanced control algorithms
• Improving IK solvers
• Adding more comprehensive examples

License

This project is based on components from Simple-MuJoCo-PickNPlace and MuJoCo tutorials.

References

• MuJoCo Documentation
• Universal Robots UR5e
• Simple-MuJoCo-PickNPlace