Geospatial data for peer-to-peer communication among autonomous vehicles using optimized machine learning algorithm.

Area of Science:

• Computer Science
• Artificial Intelligence
• Transportation Engineering

Background:

• Autonomous vehicles are crucial for future transportation, but driver error causes over 60% of accidents globally.
• Developing autonomous vehicle systems faces challenges in natural disaster identification and inter-vehicle data exchange.
• Current advancements in AI, ML, Data Science, and Big Data support autonomous systems but require specialized solutions for network communication.

Purpose of the Study:

• To address challenges in autonomous vehicle communication by focusing on data cleansing and seamless interaction.
• To develop a novel Support Vector Machine (SVM) kernel tailored for peer-to-peer (P2P) networks in autonomous vehicles.
• To create an optimized hybrid approach for weight selection in SVM to improve performance.

Main Methods:

• Proposed a new Weighted-Support Vector Machine (W-SVM) kernel to meet Mercer's theorem constraints.
• Developed a hybrid optimization strategy combining Grey Wolf Optimizer (GWO) and Elephant Herding Optimisation (EHO) for weight vector derivation.
• Implemented a data cleansing approach for enhanced vehicle-to-vehicle communication.

Main Results:

• The proposed W-SVM kernel, optimized via hybrid GWO-EHO, demonstrated superior performance in handling complex data issues.
• The hybrid optimization enhanced convergence speed, exploitation, and exploration capabilities compared to individual algorithms.
• The novel approach showed significant improvements in data cleansing and facilitated seamless interaction among autonomous vehicles.

Conclusions:

• The developed hybrid optimization approach for SVM weight selection is effective for autonomous vehicle networks.
• The novel W-SVM kernel offers a robust solution for improving data exchange and safety in autonomous driving.
• This research contributes to overcoming key hurdles in the widespread adoption of autonomous vehicle technology.