Overview

This research paper, published in the proceedings of SciTePress, introduces a novel consensus protocol designed for decentralized robotic systems operating in GPS-denied environments. The work focuses on improving coordination, communication efficiency, and reliability in multi-robot navigation.

Publication

Published by SciTePress (Science and Technology Publications), this paper contributes to research in robotics, blockchain-based consensus systems, and computer vision.

View Paper

Abstract

This paper presents Witness Byzantine Fault Tolerance (WBFT), a consensus protocol for Wide Area Visual Navigation (WAVN) systems where robots share visual data and agree on navigation decisions via blockchain. WBFT reduces communication overhead compared to traditional methods like PBFT by using a lightweight, signature-based approach. It integrates Proof-of-Navigation (PoN) for leader election and uses Ed25519 signatures with Merkle trees to improve efficiency, achieving O(n) communication complexity and O(log t) verification. The protocol tolerates Byzantine faults and demonstrates strong scalability, resilience, and performance in decentralized robotic networks.

My Contribution

I contributed to the written research, focusing on the concept of landmark detection and its importance in robotic perception and navigation.

Landmark Detection in Robotics

Landmarks are distinct environmental features that robots can recognize through sensory input such as images. These features typically have fixed positions and are visually distinguishable through characteristics like color or contrast.

When robots can determine their relative positions, they can narrow their search space and more efficiently identify shared landmarks. This reduces ambiguity caused by similar-looking features in different locations and improves coordination across multiple robots.

The research highlights current limitations in detecting natural landmarks (features not specifically designed for robotic navigation) despite advances in computer vision. Reliable and fast landmark recognition remains a key challenge in real-world applications.

Key Insight

Efficient and unambiguous landmark detection is essential for enabling robust, real-time multi-robot navigation in complex environments.

Topics & Technologies

  • Computer Vision
  • Robotics & Autonomous Systems
  • Blockchain & Consensus Algorithms
  • Multi-Robot Coordination