Autonomous Vehicle Parking Path Planning with Dynamic Obstacle Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current parking logic in autonomous driving vehicles (ADVs) does not account for the movement, speed, and direction of obstacles when reversing during the parking process, leading to potential collisions as obstacles are only considered when within a predetermined boundary.
Innovation Solution
A computer-implemented method for ADVs that generates a parking path considering obstacles' location, speed, and direction by switching between forward and reverse driving directions, using perception, prediction, and planning modules that invert their orientation to account for changes in driving direction, allowing for smooth navigation and obstacle avoidance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If parking logic treats obstacles as static objects and only detects them when within a predetermined boundary, then the parking system is simple to implement, but collisions with moving obstacles cannot be avoided
Solution Approach 1:
The system performs preliminary detection of moving obstacles and their trajectory prediction before the vehicle begins reverse parking maneuvers. The perception module identifies obstacles with movement characteristics, and the prediction module forecasts their future positions, allowing the planning module to generate collision-free parking paths in advance.
Solution Approach 2:
The system transitions from treating obstacles as static to dynamically tracking moving obstacles throughout the parking process. The perception module continuously updates obstacle positions and movement characteristics, the prediction module dynamically forecasts future positions based on current speed and direction, and the planning module continuously generates updated parking paths that account for changing obstacle positions during reverse driving.
2Measurement precision
If the ADV uses standard forward-driving perception and planning modules for reverse parking, then the system architecture remains simple, but the modules cannot accurately predict obstacle positions when driving in reverse
Solution Approach 1:
When the vehicle enters reverse driving mode, the system inverts the orientation of the perception and planning modules. This inversion allows the modules, originally designed for forward driving, to correctly interpret sensor data and generate appropriate parking paths for reverse maneuvers by flipping their coordinate systems and detection directions.
Solution Approach 2:
The system determines the driving direction (forward or reverse) before executing parking maneuvers and pre-configures the orientation of perception and planning modules accordingly. This preliminary orientation setup ensures that obstacle detection and path planning are performed with the correct spatial reference frame from the start of each parking attempt.
Data Source
AI summary
A parking system for autonomous driving vehicles (ADV) is disclosed that utilizes the perception, planning, and prediction modules of ADV driving logic to more safely and accurately park an ADV. An ADV scans a parking lot for an available space, then determines a sequence of portions or segments of a parking path from the ADV's location to a selected parking space. The sequence of segments involves one or more forward driving segments and one or more reverse driving segments. During the forward driving segments, the ADV logic uses the perception, planning, and prediction modules to identify one or more obstacles to the ADV parking path, and speed and direction of those obstacles. During a reverse driving segment, the ADV logically inverts the orientation of the perception, planning, and prediction modules to continue to track the one or more obstacles and their direction and speed while the ADV is driving in a reverse direction. For each parking path portion, the planning module generates a smooth reference line for the portion, taking into account the one or more obstacles, and their speed and direction.