Probabilistic Parking Space Selection for Front-In Autonomous Parking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous and automated parking systems typically rely on back-in parking techniques, which are inefficient in crowded parking environments, as they require vehicles to pass through parking spaces to confirm availability and dimensions, making front-in perpendicular parking more natural and practical but challenging to implement.
Innovation Solution
A probabilistic approach using sensor data from vehicles to determine available parking spaces and their characteristics, such as width, entry turning radius, and longitudinal distance, allowing the system to select and navigate to a suitable space using assisted or autonomous driving, and perform either a single-turn or two-turn maneuver based on these characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If back-in parking technique is used, then parking space availability can be confirmed, but parking efficiency deteriorates in crowded environments
Solution Approach 1:
The system performs preliminary detection of parking space characteristics (width, length, occupancy) before the vehicle attempts to park. Sensors scan and evaluate multiple parking spaces in advance, identifying suitable targets without requiring the vehicle to pass through them, thus resolving the contradiction between confirmation reliability and parking efficiency
2Productivity
If front-in perpendicular parking is implemented, then parking efficiency improves, but maneuverability requirements increase
Solution Approach 1:
The control system acts as an intermediary between the driver's parking intent and the vehicle's physical maneuvering. It automatically calculates optimal trajectories, determines whether single-turn or two-turn maneuvers are feasible based on space characteristics, and executes the sequence of steering and acceleration commands, thereby enabling front-in parking without increasing driver maneuverability burden
3Loss of time
If single-turn maneuver is used, then parking time is reduced, but space requirements increase
Solution Approach 1:
The system dynamically selects between single-turn and two-turn maneuvers based on real-time assessment of parking space dimensions and vehicle characteristics. If the space is sufficient, a single-turn maneuver is executed for faster parking; if space is constrained, a two-turn maneuver is chosen instead, optimizing the trade-off between time and space requirements
Data Source
AI summary
This document describes techniques and systems for selecting a parking space using a probabilistic approach. An example system includes a processor that can determine whether multiple parking spaces are available in proximity to a host vehicle using sensor data. Parking-space characteristics (e.g., a width, entry turning radius, and longitudinal distance to the parking space) of each available parking space are determined using the sensor data. The processor can then choose a selected parking space among the multiple parking spaces based on the parking-space characteristics. The processor or another processor can then control the host vehicle to park in the selected parking space using an assisted-driving or autonomous-driving system. In this way, the described system can select and navigate to a parking space using a probabilistic approach. The processor can choose the parking space and parking maneuver based on programmable and customizable parking-space characteristics in some implementations.