Created by Alex Adamov
An interactive, 3D educational tool for understanding how robots plan movement around obstacles.
Live demo: alexadamov.com/rrtsim
I started learning about robotic algorithms with Gemini 3. I initially built the 3d simulation for my own understanding, but as it came together, I decided to create an interactive tutorial to share these concepts with others.
Imagine you have a robot arm and you want it to grab an object behind a wall. You can't just move straight there—the arm would hit the wall. You need to figure out a path that snakes "around" the obstacles.
This application visualizes how computers solve this problem. It takes you through a journey from naive approaches that fail, to advanced "tree-growing" algorithms used in real-world robotics (like self-driving cars and industrial automation) that can explore complex spaces to find a safe path.
The app creates a 3D simulation of a multi-jointed robot arm. You can define obstacles (walls, gates, corridors) and a target position. The app then visualizes three different strategies for reaching that target:
-
The Greedy Approach (Cyclic Coordinate Descent):
- Concept: "Move every joint to get closer to the target right now."
- Result: It works great in open space but fails miserably with obstacles. It gets stuck trying to go "through" a wall because it never thinks to go "around" it.
-
Standard RRT (Rapidly-exploring Random Tree):
- Concept: "Let's explore blindly." The robot randomly samples a point in space and tries to extend a branch toward it. Over time, it builds a massive "tree" of safe movements.
- Result: It eventually finds a path, but it's slow and explores the entire room just to move a few inches.
-
RRT-Connect (Bi-directional):
- Concept: "Let's meet in the middle." One tree grows from the start, and another grows backward from the goal. They aggressively try to connect to each other.
- Result: Much faster and more efficient. This is the industry-standard approach for many motion planning problems.
The project is built with:
- React: For the UI and state management.
- Three.js: For the 3D rendering and simulation environment.
- Web Workers: To run the heavy RRT calculations off the main thread, keeping the UI smooth even while calculating thousands of path nodes.
This is an open educational project. I invite comments, additions, and improvements! Whether it's adding new algorithms (like RRT* or PRM), improving the visualization, or fixing bugs, your contributions are welcome.
If you found this useful or interesting:
This project is licensed under the MIT License - see the LICENSE file for details.