A Personalized Navigation Route Recommendation Strategy Based on Differential Perceptron Tracking User's Driving...

Area of Science:

• Autonomous Driving
• Artificial Intelligence
• Human-Computer Interaction

Background:

• Personalized path planning is crucial for autonomous driving.
• User path preferences are dynamic and influenced by various factors.
• Existing strategies may not effectively adapt to changing user needs.

Purpose of the Study:

• To propose a personalized path recommendation strategy that tracks and adapts to user path preference changes.
• To develop a system capable of dynamically updating path selection criteria.
• To ensure optimal path planning for autonomous vehicles under evolving user preferences.

Main Methods:

• Data collection and establishing relationships with user preference factors.
• Utilizing dichotomized K-means algorithm for initial preference weight vector determination.
• Implementing a threshold-based system to detect preference changes and redefining factors or using difference perception for updates.
• Quantifying the road network based on the user preference weight vector.
• Employing the Tabu search algorithm for optimal pathfinding.

Main Results:

• The proposed strategy successfully tracks and adapts to user path preference changes.
• Dynamic updating of preference weight vectors ensures continued relevance of path recommendations.
• Simulation results demonstrate the strategy's effectiveness in two distinct scenarios.
• The system meets the requirements of autonomous driving even when user preferences fluctuate.

Conclusions:

• The developed personalized path recommendation strategy effectively handles dynamic user preferences in autonomous driving.
• The adaptive nature of the system ensures robust and personalized route selection.
• This approach enhances the user experience and efficiency of autonomous vehicle navigation.