Radar-Centric Parking Spot Selection for Occluded Perpendicular Parking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous and automated parking systems rely on infrastructure sensors or vision-based systems that struggle to identify and confirm available parking spots, especially in crowded environments where spots are occluded, limiting their effectiveness.
Innovation Solution
The use of radar-centric occupancy grid (RCOG) maps by a processor on a vehicle to identify and select available parking spots using sensor data from radar, camera, and other systems, allowing for front-in perpendicular parking without relying on infrastructure sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If vision sensors (cameras) are used to identify parking spots, then the system can detect parking spot availability, but the system fails to reliably identify spots in crowded environments where spots are occluded by parked vehicles
Solution Approach 1:
The patent combines radar sensors with vision sensors to create a fused perception system. The radar occupancy grid map integrates radar-detected objects with camera-identified parking spots, allowing the system to reliably detect occluded spots by using radar's ability to see through obstacles while maintaining camera's spot identification capabilities.
Solution Approach 2:
The patent introduces a radar occupancy grid map as an intermediary layer between raw sensor data and parking spot identification. This intermediary structure merges radar occupancy information with vision-based spot detection, enabling reliable identification of occluded spots by filling in gaps where direct vision is blocked.
2Reliability
If infrastructure sensors are used to identify parking spots, then the system can detect available spots, but the system is limited to specific parking lots and parking garages
Solution Approach 1:
The patent enables the vehicle to perform its own parking spot identification using onboard radar and vision sensors, eliminating dependence on external infrastructure sensors. The vehicle independently creates a radar occupancy grid map and identifies parking spots, making the system universally applicable to any parking environment without requiring specialized infrastructure.
Solution Approach 2:
The patent replaces infrastructure-based sensor systems with a vehicle-mounted sensor fusion system. By substituting external infrastructure sensors with onboard radar and camera systems that work together, the system achieves both reliable detection and universal adaptability across different parking environments.
3Reliability
If back-in parking techniques are used, then the system can identify parking spots, but the system needs to pass each spot to confirm availability, increasing time and complexity
Solution Approach 1:
The patent performs preliminary identification of multiple parking spots simultaneously using the radar occupancy grid map before the vehicle needs to confirm availability. By pre-identifying and filtering suitable spots based on characteristics like width and occlusion level, the system reduces the time needed for confirmation during actual parking execution.
Solution Approach 2:
The patent identifies and evaluates more parking spots than ultimately needed by using the radar occupancy grid to pre-screen multiple candidates. This excessive identification action allows the system to have multiple confirmed options ready, reducing time pressure during the actual parking maneuver and enabling faster decision-making.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables the vehicle to autonomously identify, select, and navigate to a parking spot based on characteristics like width, turning radius, and distance, improving parking efficiency and practicality in crowded environments.
Implementation Method 1
identifying, using sensor data obtained from one or more sensors of a host vehicle, one or more available parking spots near the host vehicle in a parking environment, the one or more sensors including a radar system and the sensor data including a radar occupancy grid (ROG) map
Data Source
Figure 1
Figure 2
Figure 3
AI summary
This document describes techniques and systems for identifying and selecting a parking spot using radar-centric occupancy grid (RCOG) maps. An example system includes a processor that can identify available parking spots near a host vehicle using sensor data, including a radar occupancy grid (ROG) map. Parking-spot characteristics of each available parking spot are also determined using the sensor data. The processor can then determine a selected parking spot based on the parking-spot characteristics. The processor or another processor can then control operation of the host vehicle to park in the selected parking spot using an assisted-driving or autonomous-driving system and based on the parking-spot characteristics. In this way, the described system can identify, select, and navigate to a parking spot using a RCOG map without having to rely on infrastructure sensors or pass by open spots.