This project implements SLAM (Simultaneous Localization and Mapping) for a 2D robot world, as part of the Udacity Sensor Fusion / Robotics curriculum.
The goal is to track a robot’s position over time while simultaneously estimating the locations of landmarks, using only noisy motion and sensor measurements.
In this project, SLAM is implemented for a two-dimensional grid world.
The robot moves through the environment while observing landmarks and estimating both:
- Its own pose (x, y position)
- The positions of landmarks in the environment
The system combines:
- Robot motion updates
- Noisy landmark measurements
- Constraint-based estimation (Omega–Xi formulation)
SLAM enables the robot to build a map of the environment incrementally over time, which is a fundamental capability in autonomous robotics and navigation.
Below is an example of a 2D robot world with:
- Purple “×” → detected landmarks
- Red “o” → robot final estimated position
This example corresponds to a 50×50 grid world, but multiple map configurations can be generated.
The project is organized into three Python notebooks, but only the final notebook and robot implementation are graded.
-
Notebook 1 – Robot Moving and Sensing
Introduces robot motion and noisy sensing models. -
Notebook 2 – Omega and Xi Constraints
Implements SLAM constraints using information matrices. -
Notebook 3 – Landmark Detection and Tracking Full SLAM implementation combining motion and sensor constraints.
robot_class.pyNotebook 3 – Landmark Detection and Tracking.ipynb
- Simultaneous Localization and Mapping (SLAM)
- Probabilistic robotics
- Constraint-based optimization (Omega–Xi)
- Noisy motion and measurement models
- Landmark-based mapping
- Graph-based state estimation
- Open Notebook 3 in Jupyter:
jupyter notebook
## Author
**Bhagyath Badduri**
Master’s Student – Robotics