Parallel Parking Intention Prediction Using Vehicle Polygons
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in identifying and predicting parallel parking opportunities amidst static and dynamic objects in their environment, leading to inefficient or unsafe parking maneuvers.
Innovation Solution
An autonomous vehicle system uses sensor data to identify open parallel parking locations and vehicles with potential parking intentions by generating a polygon around the vehicle, determining if the polygon contains the parking space, and adjusting its movement to safely navigate around the intended parking spot.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the autonomous vehicle uses basic sensor data processing to identify objects, then the processing speed is maintained, but the ability to accurately predict parallel parking intentions is insufficient
Solution Approach 1:
The system segments the prediction task into multiple independent analysis dimensions: polygon generation from sensor data, parallel parking location identification, and intention prediction based on spatial relationships. Each segment processes specific features independently, then combines results to achieve accurate prediction without requiring a monolithic complex system.
Solution Approach 2:
The system performs preliminary actions by generating polygons around detected vehicles and pre-identifying potential parallel parking locations before making prediction decisions. This advance preparation of spatial data structures enables faster and more accurate intention prediction without increasing real-time processing complexity.
2Reliability
If the autonomous vehicle maintains a large safety distance from potential parking vehicles, then collision risk is reduced, but parking efficiency decreases
Solution Approach 1:
The system dynamically adjusts the safety distance based on the predicted parallel parking intention of detected vehicles. When a vehicle is identified as intending to parallel park, the autonomous vehicle adaptively modifies its following distance and lateral positioning, maintaining greater distance during the parking maneuver and reducing it when safe, thereby balancing safety and parking efficiency.
Solution Approach 2:
The system uses feedback from the prediction model about vehicle intentions to continuously adjust the autonomous vehicle's positioning and speed. The prediction output feeds back into the control system, enabling real-time modification of safety margins to optimize both collision avoidance and parking completion time.
3Productivity
If the autonomous vehicle uses simple object detection, then the processing speed is fast, but the ability to identify parallel parking opportunities is insufficient
Solution Approach 1:
The detection system is segmented into a two-stage process: first, simple and fast object detection identifies potential vehicles in the environment; second, a more sophisticated analysis stage processes spatial relationships and generates polygons to identify parallel parking opportunities. This segmentation allows the system to maintain high processing speed for basic detection while applying advanced analysis only where needed.
Solution Approach 2:
The system applies partial advanced analysis by generating polygons and analyzing spatial relationships only for objects that pass the initial simple detection filter. Rather than applying complex analysis to all detected objects, it selectively processes only those with potential parallel parking relevance, maintaining processing speed while improving detection capability for parking opportunities.
Data Source
AI summary
Disclosed herein are system, method, and computer program product embodiments for identification of intention and prediction for parallel parking vehicles. For example, the method includes: obtaining sensor data associated with an environment surrounding an autonomous vehicle; identifying, based on the sensor data, a plurality of static objects in the environment, an open parallel parking location between the plurality of static objects, and a vehicle in the environment that satisfies a parallel parking condition; generating a polygon that extends beyond a side and a rear of the vehicle that satisfies the parallel parking condition; and controlling movement of the autonomous vehicle based on a prediction that the vehicle is intending to parallel park in the open parking location, the prediction that the vehicle is intending to parallel park in the open parking location being determined based on an amount of the open parallel parking location that is contained within the polygon.