Design and Control of Modular Magnetic Millirobots for Multimodal Locomotion, Shape Reconfiguration and Grasping

Overview

The project presents a modular magnetic millirobotic platform capable of:

• Single-module locomotion using Helmholtz and Maxwell coil inputs.
• Multimodal reconfiguration into chains, squares, and grippers.
• Closed-loop navigation with real-time vision feedback and A* path planning.
  All functions are achieved at low magnetic field strengths (<13 mT) using a compact 2-D electromagnetic setup.

The framework consists of three layers:

1. Arduino firmware for low-level current control of electromagnetic coils.
2. Python framework for experiment orchestration, computer vision, recording, and autonomous path planning.
3. MATLAB tools for post-processing and statistical analysis of recorded trajectories.

Key Features

• Magnetic actuation control via serial communication with external drivers.
• Vision-based module detection (moduleDetection.py) using OpenCV for real-time tracking.
• Finite State Machine (FSM) (FiniteStateMachine.py) for handling motion commands and transformations.
• Path planning mode with goal setting and automated trajectory visualization.
• Sequence execution (Sequences/) for predefined current input patterns.
• Automated recording and analysis with trajectory overlays, merge detection, and reconfiguration logging.

Repository Structure

MIP/
├── Arduino/
│   ├── EMS.ino # Arduino code for coil current control
│   └── originalPWM.ino # Legacy version
├── MATLAB/
│   ├── Trials/ # Raw trial data
│   ├── DataAnalysis.m # Post-processing and plotting
│   ├── *.png / *.xlsx # Analysis figures and datasets
├── Python/
│   ├── Images/ # Module configuration dataset
│   ├── Basic Interfacing/ # Camera, calibration, serial tools
│   ├── Sequences/  # Predefined coil actuation sequences
│   ├── main.py # Main entry point (control + recording)
│   ├── pathPlanning.py # Path planning and trajectory execution
│   ├── FiniteStateMachine.py # FSM for command-based motion simulation
│   ├── moduleDetection.py # Vision-based module detection and tracking
│   ├── README.md # Project documentation

Hardware Setup

• Modular robot platform with electromagnetic coil actuation.
• Arduino microcontroller controlling 4 coil channels.
• USB camera for live vision-based tracking.
• DC power supply with current-limiting enabled (safety-critical).
• Gaussmeter (optional) for magnetic field calibration.

Safety: Ensure coil currents do not exceed 10 A on HX/HY channels. Always enable current limiting on the power supply.

Software Requirements

• Arduino IDE (to flash EMS.ino)

• Python 3.9+ with:

  • opencv-python
  • numpy
  • matplotlib
  • pyserial

• MATLAB (for analysis & plotting)

Install Python dependencies with:

pip install -r requirements.txt

Workflow

1. Upload Firmware

Flash Arduino/EMS.ino to the Arduino board. This program receives serial commands to set coil currents.

2. Calibrate Fields

Use Python/Basic Interfacing/fieldCalibration.py to generate calibration curves with a gaussmeter. Alternatively, serial_interface.py provides direct manual current control.

3. Run Experiments

Launch:

python Python/main.py

Features include:

• Manual control: enter custom current values.
• Predefined sequences: run stored coil actuation sequences.
• Autonomous path planning: select goals in live workspace view; system computes A* path and executes movements.

4. Recording & Replay

• All experiments can be recorded (AVI format).
• Trajectories and merge/reconfiguration events are detected automatically.
• After each run, playback allows saving/discarding videos and generating trajectory overlays.

5. Data Analysis

Run MATLAB analysis with:

DataAnalysis.m

This script:

• Loads CSV files generated during experiments.
• Computes statistics across experiments.
• Generates publication-ready plots (e.g., displacement, trajectory variance, time to merge).

Example Experiment

1. Upload firmware to Arduino.

2. Start main control:

   python Python/main.py

3. Select a sequence (e.g., seq1) or enter manual currents.

4. Record and replay the experiment.

5. Save trajectory outputs (PNG overlays, merge/reconfiguration logs).

6. Analyze results in MATLAB with DataAnalysis.m.

Outputs

Each experiment can produce:

• Videos (AVI) – raw and annotated recordings.
• Trajectory plots (PNG) – full and interval-based overlays.
• Merge/reconfiguration logs (TXT) – timing of module interactions.
• CSV files – displacement, centroid coordinates, and experimental metadata.
• MATLAB figures – summary statistics and comparative plots.