Parking Path Search With Intermediate Node Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous or automated parking systems face computational inefficiencies due to large search spaces and non-holonomic constraints, leading to prolonged computational times in finding optimal parking paths, especially in corner cases like parallel parking.
Innovation Solution
A modified parking path algorithm that uses a Dijkstra-based approach, such as Hybrid A*, determines initial and secondary waypoints by selecting an intermediate node close to the goal node after a threshold is reached, balancing computational efficiency with near-optimal path generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a traditional parking path algorithm is used to search for an optimal parking path, then the path quality is high (close to globally optimal), but the computational time is too long for practical implementations
Solution Approach 1:
The patent divides the parking path search into two segments: a first parking path from the initial pose to an intermediate pose, and a second parking path from the intermediate pose to the goal pose. This segmentation allows the system to find satisfactory paths faster by breaking down the complex search space into manageable portions, resolving the contradiction between path quality and computational time.
Solution Approach 2:
The system performs preliminary action by generating the first parking path to an intermediate pose before finalizing the complete path to the goal pose. This preliminary path provides a good starting point that reduces the search space for the second path, enabling faster computation while maintaining path quality.
2Productivity
If the search space is reduced to decrease computational time, then the computational efficiency is improved, but the ability to find the globally optimal path is compromised
Solution Approach 1:
By segmenting the search into two phases (initial pose to intermediate pose, then intermediate pose to goal pose), the system reduces the effective search space in each phase while still exploring sufficient variations to find high-quality paths. This maintains computational efficiency without completely sacrificing optimality.
Solution Approach 2:
The system applies different search strategies to different portions of the path: the first path uses a broader search from the initial pose, while the second path uses a more focused search from the intermediate pose. This local differentiation optimizes computational resources while maintaining overall path quality.
Data Source
AI summary
This document describes techniques and systems for performing a parking path search using a modified parking path algorithm. An example parking system can obtain an initial pose and a goal pose of a host vehicle and an obstacle map for a parking environment. The parking system can then use a parking path algorithm to determine first waypoints for a parking path. In response to a number of run-time iterations being greater than a threshold, the parking system can select an intermediate source node closest to the goal pose. The parking path algorithm can then be used to determine second waypoints from the intermediate source node to the goal pose. The host vehicle can then be controlled to park using the first waypoints and the second waypoints. In this way, a parking path search is performed more efficiently, while still generating a path close to the globally-optimal solution.