Vehicle Path Planning With Forward-Reverse Parking Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous vehicle parking systems face challenges in generating feasible paths and velocity profiles that account for non-holonomic kinematics and dynamics constraints, particularly in tight environments, leading to inefficient and error-prone parking maneuvers.
Innovation Solution
A four-phase path planning system that uses a virtual vehicle to generate forward and reverse path segments connected along a target line, incorporating switching control laws to ensure collision avoidance and optimal trajectory generation, while considering kinematics and dynamics constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional path planning methods are used for autonomous parking, then the vehicle can complete parking maneuvers, but the path smoothness and positioning accuracy deteriorate due to non-holonomic constraints and computational errors
Solution Approach 1:
The patent segments the continuous path planning problem into discrete grid-based cells, where the environment is divided into manageable units. This segmentation allows the vehicle to navigate through a simplified representation of space, improving positioning accuracy while reducing the computational complexity of handling continuous coordinates and constraints.
Solution Approach 2:
The patent introduces a virtual dimension by creating a virtual vehicle model that operates independently from the physical vehicle. This virtual model exists in a simplified computational space where path planning can be performed more accurately, and then the results are mapped back to the physical vehicle, enhancing positioning precision without proportionally increasing physical system complexity.
2Reliability
If complex path planning algorithms are used to account for kinematics and dynamics constraints, then path feasibility improves, but computational time and power consumption increase
Solution Approach 1:
The patent performs preliminary path planning in a virtual environment before executing maneuvers with the physical vehicle. The virtual vehicle tests and validates paths under simulated kinematics and dynamics constraints, so that when the physical vehicle executes the pre-planned path, feasibility is ensured without requiring complex real-time computations, thus reducing computational time.
Solution Approach 2:
The patent creates a virtual copy of the vehicle and its environment to perform path planning simulations. This copying approach allows complex constraint validation to be performed in the virtual model, ensuring path feasibility while the physical vehicle simply follows pre-determined commands, significantly reducing the computational burden on the actual vehicle system.
3Adaptability or versatility
If human intervention is used to guide parking maneuvers, then adaptability to unexpected obstacles improves, but automation level and efficiency decrease
Solution Approach 1:
The patent implements a feedback mechanism where sensors continuously monitor the physical vehicle's position and the virtual vehicle's simulated position. This feedback loop allows the system to detect deviations and correct them automatically, maintaining high automation levels while adapting to unexpected obstacles through real-time comparison and adjustment between virtual and physical states.
Solution Approach 2:
The virtual vehicle model autonomously handles obstacle detection and path adjustment without requiring human intervention. The system serves itself by using the virtual model to predict and respond to obstacles, automatically generating corrected paths when deviations are detected, thereby maintaining full automation while improving adaptability to unexpected situations.
Data Source
AI summary
Systems, methods, and computer program products for guiding vehicles. A vehicle control system detects obstacles and avoids contact between the vehicle, and the obstacles by solving a path planning problem using a four-phase system. In a parking scenario, switching control laws are used to drive the vehicle to a target line, and a forward path segment and a reverse path segment defined. The two path segments are connected along the target line to define a path for entering or exiting the parking space. Objects in a driving environment may be avoided by identifying an obstacle corner, defining an avoidance circle (62), (185) around the obstacle corner, and determining a path that allows the vehicle to avoid penetrating the avoidance circle. A line-of-sight guidance method may be used to follow the path by defining target points on the path at a lookout distance l, and steering the vehicle using the target points.