bdsim: Block Diagram Simulation for Python

A Pythonic block-diagram environment for the simulation and analysis of dynamic systems.

Block diagram thinking → Python coding

GitHub • Wiki • Changelog • Installation

Status & Project Health

Ecosystem & Dependencies

Synopsis

bdsim bridges the gap between mathematical block diagrams and executable Python code. Unlike traditional graphical-only tools, it treats modelling as code, allowing you to define, simulate, and analyze continuous-time, discrete-time, or hybrid systems within a modern software engineering workflow. Wires in bdsim aren't limited to scalars; they pass NumPy arrays, dictionaries, or even SpatialMath objects seamlessly through your system, integrating directly with the Robotics and Machine Vision toolboxes.

🚀 Key Features

• Python-First Workflow: Define your systems in pure Python code. Use your preferred IDE (like VS Code), manage versions with Git, and integrate with standard unit-testing frameworks.
• Rich Data Types: Wires in bdsim aren't limited to scalars. Seamlessly pass NumPy arrays, dictionaries, or complex objects like SE3 and SO3 from the spatialmath-python library.
• Modular & Extensible: Adding new functionality is straightforward. Create custom blocks by simply subclassing the Block class.
• Hybrid Ecosystem: Native integration with the Robotics Toolbox and Machine Vision Toolbox for Python.
• Powerful Editor: Includes bdedit, a PySide-based graphical editor for visual system design and discovery.
• Publication Ready: Export diagrams or simulation data to to high-quality formats such as PDF or SVG for use in publications.

💻 Quick start

You can install bdsim directly from PyPI:

pip install bdsim

To include the graphical editor (bdedit) and its dependencies:

pip install bdsim[editor]

Install locally with help from the detailed installation guide.

Or skip setup and run the browser-based JupyterLite examples.

Example

The power of bdsim lies in its conciseness. The step response of a simple first-order system can be defined and simulated in just a few lines of code:

import bdsim

sim = bdsim.BDSim()
bd = sim.blockdiagram()

# Define blocks
step = bd.STEP(T=1, pos=1)
plant = bd.LTI_SISO(1, [1, 1]) # 1/(s+1)
scope = bd.SCOPE()

# Connect blocks
bd.connect(step, plant)
bd.connect(plant, scope)

bd.compile()
out = sim.run(bd, T=5)

🔍 Why bdsim?

In contrast to traditional graphical simulation tools, bdsim treats modelling as code. This ensures:

• Better Version Control: No more opaque binary blobs. Your models are searchable, diffable text.

• Seamless Integration: Incorporate SciPy solvers, PyTorch models, or custom robotic vision pipelines directly into your simulation loops.

• Scalability: Programmatically generate massive block diagrams or run large-scale batch simulations on headless servers.

📚 Documentation

Full Documentation: https://petercorke.github.io/bdsim/

Wiki: Access the community wiki for deep dives into specific block behaviours and tutorials.