ISEP · Final Engineering Project · 2009
Vision-Based Mobile Robot Localization & Navigation
Computer Vision · Mobile Robotics (Academic)
Context
An academic project at ISEP building a mobile robotic platform able to localize itself and navigate using vision.
The challenge
Getting a mobile robot to recognize its environment and obstacles from image data alone, and to use that to localize itself and avoid obstacles — including exploring introductory neural networks for the task.
The approach
Built in MATLAB, the project used image filtering and morphological operations (including erosion), pattern and color recognition, and introductory neural networks to achieve robot localization and obstacle avoidance on a mobile robotic platform, alongside a bee-communication board.
System architecture
MATLAB image pipeline (filtering, morphological operations/erosion, pattern & color recognition) → introductory neural network → localization & obstacle avoidance on a mobile robotic platform, with a bee-communication board.
Technologies
- MATLAB
- Image filtering
- Morphological operations / erosion
- Pattern and color recognition
- Introductory neural networks
- Mobile robotic platform
- Bee communication board
Project journey
- Concept
- Prototype
- Validation
- Industrialization
- Deployment
- Scale
My contribution
- Designed
- Developed
- Validated
Deployment / industrialization
Not yet industrialized/deployed — see Project journey above for the stage actually reached.
Skills demonstrated
- Computer vision (MATLAB)
- Mobile robotics
- Introductory machine learning
- Image recognition
- Robot localization