This is a public repository that invites collaborators who would like to analyze and develop the proposed idea — a memory-based localization method for controlled environments.
To achieve localization without sensor fusion, we adopt a camera-based approach using computer vision. Unlike traditional AI models, this method is designed to be efficient and lightweight, avoiding computationally intensive training. Instead of analyzing every pixel, this method shrinks an image to a single pixel, capturing its color and grayscale ratios as a compact memory vector.
Check out the ideation write-up for LVM and its process flowchart for clarification in the LVM_proposal.pdf attached in the main branch of this repository.
Interested collaborators are invited to join in:
- Idea analysis
- Development
- Simulation
- Testing
Kindly, collaborators are requested to check the CONTRIBUTING.md file for instructions on how to collaborate globally through this repository.
Regards,
Harishma P.
Chennai Institute of Technology
(Originator, 2025)