Applied Programming - EMG Data Viewer

A PyQt application for visualization, processing and recording of a simulated multi-channel EMG Signal

Getting Started

• Create virtual environment (assuming uv as package manager):

uv venv .venv

• Activate the virtual environment:

uv activate .venv

• Install dependencies:

uv pip install -r requirements.txt

• Run the application from the root folder:

python main.py

This will launch the EMG Viewer window.

Usage

1. Connect to Signal Source

• Click the “Connect” button to start both the TCP server and client.
• The application begins generating and receiving EMG data in the background.
• To stop the stream, click “Disconnect”.

---

2. Start/Stop Live Plotting

Click the “Start” button to begin plotting live EMG data. While active:

• The signal of the current channel is shown in the upper window, with the selected transformation
• The full 32-channel data is added to the recording, seen below. Click “Stop” to pause live plotting and recording. The signal continues in the background.

---

3. Live View Controls

• Channel Selector: Choose one of 32 channels to view.

  Only one channel is shown, but all channels are being recorded.

• Signal Processing Modes:

  Mode	Description
  Raw	Unprocessed signal
  RMS	Root Mean Square
  Filter	Band-pass Butterworth filter
  Envelope	Signal envelope using Hilbert function

---

4. Recording Controls

• Channel Selector: Selects which recorded channel to display.
• Signal Processing Modes: Processing modes, same as in the live view.
• Export Recording: Save the processed recording for the selected channel as a .csv file with columns for Time and Value
• Clear Recording: Clears the current recording, starts a new one.

---

TCP Connection Specifications

The application uses a custom TCP-based client-server architecture to simulate a real-time EMG data stream.

Server (class EMGTCPServer)

• Host: localhost
• Port: 12345
• Behavior:
  • Loads EMG data from a recording.pkl file
  • Sends data packages to connected clients
  • Automatically loops the signal when it reaches the end

Client (class EMGTCPClient)

• Connects to the server at localhost:12345.
• Receives data to simulate real EMG device
• Decodes each data packet into a NumPy array of shape (32, 18) — 32 channels, 18 samples per packet.

---

Packet Format

Each packet contains:

• Data Shape: (32 channels × 18 samples)
• Data Type: float32
• Total Size: 32 * 18 * 4 = 2304 bytes per packet
• Transmission Rate: One packet every 9 ms (based on sleep_time = 18 / 2000)

---

Timing and Flow

• The server uses a timer-based loop to maintain precise timing between transmissions.
• The client expects exactly 2304 bytes per read and reshapes it into a (32, 18) matrix.
• Both server and client continue streaming until manually stopped.
• When the end of the data is reached, it is restarted from the beginning.

MVVM Architecture Implementation

The application closely follows a Model-View-ViewModel (MVVM) architecture to separate data handling, business logic, and user interface logic.

---

Model: SignalProcessor

• Located in signal_processor.py
• Manages core data flow:
  • Starts the TCP server and client
  • Receives live data from the EMG stream
  • Buffers and records incoming signals
• Exposes:
  • live_signal: current buffer (rolling window)
  • recorded_signal: full accumulation of all recorded data

---

ViewModel: MainViewModel

• Located in main_view_model.py
• Acts as the central state and logic controller:
  • Connects to SignalProcessor
  • Manages current channel and processing modes (raw, RMS, filter, envelope)
  • Applies real-time and historical signal processing
  • Emits Qt signals to update the UI:
    • live_data_updated
    • recorded_data_updated
• Handles:
  • Timer-driven updates
  • Channel switching
  • Recording state
  • CSV export of processed data

---

Views

MainView

• Root layout controller (main_view.py)
• Composes all widgets:
  • ConnectionWidget
  • LivePlotWidget
  • RecordingPlotWidget
• Connects UI events (button presses, toggles) to ViewModel methods

LivePlotWidget

• Displays real-time EMG signals using VisPy
• Allows selection of channel and processing mode

RecordingPlotWidget

• Displays accumulated signal using Matplotlib
• Offers export and clear functionality

ConnectionWidget

• UI widget to connect/disconnect the TCP stream
• Visually reflects connection status

---

Data Flow Summary

SignalProcessor  <--->  MainViewModel  <--->  Views (LivePlot, Recording, etc)
      ▲                                      ▲
      │                                      │
   TCP I/O                         User interaction (UI)

• Model: Handles the TCP connection and provides the raw data.
• ViewModel: Transforms the data, emits it to the plots or exports it to a file.
• View Layer: Handles the UI +. visualizations, connects the view model layer to the plots.