Exclusive Traffic AVP

Exclusive Traffic AVP

Automated Valet Parking in Controlled Environment

ViaLab, Pusan National University
AVP MPC Parking ROS

Developed autonomous parking system using linear MPC algorithm with structured parking space management. Built a controlled environment where only autonomous vehicles operate, enabling systematic development and testing. Tested on actual autonomous vehicle platform in ROS environment.

Key Technologies

Control System

  • Linear MPC: Model Predictive Control for precise path tracking and trajectory execution
  • Stanley Controller: Alternative path tracking for comparison
  • Pure Pursuit: Benchmark algorithm for performance evaluation

Parking Management

  • Structured Parking Management: System for organizing and allocating parking slots
  • Slot Selection: App-based selection of target parking spaces (empty spaces intentionally designated)
  • Localization: Google Cartographer-based positioning (external implementation)
  • Approach Planning: Optimized trajectories for parking entry

Implementation Details

  • C++, Python, and ROS integration
  • Gazebo simulation environment for controlled testing
  • Real vehicle platform validation
  • Comprehensive unit and integration testing

System Architecture

  • Perception: LiDAR-based localization and slot detection
  • Planning: Path planning for parking approach and execution
  • Control: MPC-based trajectory tracking
  • Management: Parking slot allocation and coordination

Video Demo