Perception, Positioning and Decision-Making Algorithms Adaptation for an Autonomous Valet Parking System Based on...

Area of Science:

• Robotics
• Computer Vision
• Artificial Intelligence

Background:

• Autonomous parking valet systems enhance user convenience by automating parking searches and maneuvers.
• Existing systems often rely on complex infrastructure or multiple sensors.
• Challenges remain in achieving reliable autonomous parking with minimal sensing.

Purpose of the Study:

• To integrate perception, positioning, decision-making, and maneuvering algorithms for autonomous vehicle control in parking lots.
• To develop a system utilizing a single LiDAR sensor without additional infrastructure support.
• To enable autonomous vehicles to autonomously search for and execute parking maneuvers.

Main Methods:

• A simplified digital map was used to define the search range for parking spaces.
• The autonomous vehicle employed a single LiDAR sensor for navigation and space detection.
• Algorithms for perception, positioning, decision-making, and maneuvering were integrated.
• Two alternative parking strategies were implemented to optimize space utilization.

Main Results:

• The system successfully integrated perception, positioning, decision-making, and maneuvering algorithms.
• Autonomous vehicles accurately located and maneuvered into parking spaces using only LiDAR.
• The system demonstrated reliable performance in a real-world parking lot with canopies.

Conclusions:

• A single LiDAR sensor is sufficient for robust autonomous parking in complex environments.
• The developed algorithms enable efficient and accurate autonomous parking maneuvers.
• This approach reduces reliance on external infrastructure, making autonomous parking more feasible.