Salp Simulation

This repository contains MATLAB code for modeling, simulating, and analyzing the motion of a modular salp-inspired robot. The code supports symbolic modeling, ROS bag data extraction, motion reconstruction, and forward simulation.

Features

• Symbolic Model Construction: Uses CasADi for symbolic kinematics and dynamics.
• ROS Bag Data Processing: Extracts and processes experimental data from ROS bag files.
• Motion Reconstruction: Compares predicted and experimental velocities and shape changes.
• Forward Simulation: Simulates the robot's motion under specified control inputs.
• Visualization: Plots translational, angular, and shape velocities for analysis.

Directory Structure

.
├── main.m                      # Main entry point for configuration and running examples
├── salp_model.m                # Symbolic model construction
├── salp_bag.m                  # ROS bag data extraction and processing
├── salp_motion_reconstruction.m# Motion reconstruction and comparison
├── salp_forward_simulation.m   # Forward simulation and plotting
├── utils/                      # Utility functions (add to MATLAB path)
├── data/
│   ├── experiment/             # Folder for ROS bag files
│   ├── data_collection/        # Folder for ROS bag files
│   └── res.mat                 # Precomputed drag coefficients and control handles

Requirements

• MATLAB R2020a or newer (recommended)
• CasADi (symbolic computation library)
• Signal Processing Toolbox (for filtering)
• Robotics System Toolbox (for ROS bag support)

Getting Started

1. Clone the repository and open it in MATLAB.

2. Install CasADi and add its path:

   addpath('D:/casadi')

3. Add the utils directory to your MATLAB path:

   addpath('./utils')

4. Prepare your data:

   • Place your ROS bag files in ./data/experiment/ and ./data/data_collection/.
   • Ensure res.mat (with drag coefficients and control handles) is in ./data/.

5. Run the main script:

   main

   This will:

   • Build the symbolic model.
   • Run motion reconstruction (requires processed data).
   • Run a forward simulation.

Typical Workflow

1. Extract Data from ROS Bag:

   salp_bag(sys)

   This processes the ROS bag and saves a .mat file with experimental data.

2. Reconstruct and Compare Motion:

   salp_motion_reconstruction(sys)

   Loads processed data and compares predicted vs. experimental velocities.

3. Simulate Forward Dynamics:

   sys.control_handle = load('./data/res.mat').control;
   salp_forward_simulation(sys)

Notes

• Adjust the configuration in main.m as needed for your hardware or experiment.
• The code assumes a 3-link salp robot; modify dimensions in main.m for other configurations.
• The code assumes LandSalp configuration with jet velocity control and viscous drag on wheels and joints; modify the drag coefficients in main.m and use q_dot_thrust_func in salp_model.m to implement a system with thrust control and viscous drag on links and joints.
• The IMU data is in its own frames, which are calculated in salp_model.m.