Open-Space Trajectory Planning With Relaxation QP for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion planning for autonomous driving vehicles (ADVs) is inadequate in open spaces without reference lanes, as it fails to accurately plan trajectories for scenarios like parking or U-turns and does not consider vehicle-specific features, leading to potential collisions.
Innovation Solution
A relaxation optimization model using quadratic programming (QP) and reinforcement learning (RL) to generate trajectories for ADVs in open spaces, incorporating obstacle constraints and map information, and an actor-critic RL model to determine optimal control actions based on perceived environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional motion planning based on curvature and speed is used, then the planning process is simple, but the trajectory accuracy and smoothness deteriorate for different vehicle types
Solution Approach 1:
The patent transforms the motion planning problem from conventional curvature-speed parameter space to a polynomial parameter space using Bernstein basis functions. This parameter transformation enables precise control of trajectory shape, position, and orientation while maintaining computational efficiency through the polynomial representation and its derivatives.
Solution Approach 2:
The patent introduces dynamic adjustment of trajectory parameters through iterative optimization. The polynomial coefficients are dynamically adjusted based on vehicle-specific constraints and environmental conditions, allowing the system to adapt to different vehicle types and scenarios while maintaining trajectory smoothness and accuracy.
2Manufacturing precision
If trajectories are planned based on pre-labeled traffic lanes in HD maps, then the navigation is accurate on structured roads, but the adaptability to open space scenarios deteriorates
Solution Approach 1:
The patent inverts the conventional approach by not relying on pre-labeled reference lines from HD maps. Instead, it generates trajectories directly in open space by defining polynomial curves that connect start and end points while satisfying vehicle constraints and avoiding obstacles, thereby enabling operation in unstructured environments.
Solution Approach 2:
The patent segments the trajectory planning into independent polynomial curve segments (Bernstein polynomials) that can be individually optimized and combined. This segmentation allows flexible construction of complex trajectories in open space by connecting multiple smooth segments, each controlled by its own set of polynomial coefficients.
3Device complexity
If same motion planning is applied to all vehicle types, then the system is simple to implement, but the trajectory smoothness and accuracy deteriorate for vehicle-specific characteristics
Solution Approach 1:
The patent applies local quality by incorporating vehicle-specific parameters (length, width, wheelbase, maximum steering angle, etc.) into the trajectory optimization for each vehicle type. The polynomial coefficients and constraints are locally adjusted based on individual vehicle characteristics, ensuring each vehicle receives optimized trajectories tailored to its specific dynamics and geometry.
Data Source
AI summary
In one embodiment, an open space model is generated for a system to plan trajectories for an ADV in an open space. The system perceives an environment surrounding an ADV including one or more obstacles. The system determines a target function for the open space model based on constraints for the one or more obstacles and map information. The system iteratively, performs a first quadratic programming (QP) optimization on the target function based on a first trajectory while fixing a first set of variables, and performs a second QP optimization on the target function based on a result of the first QP optimization while fixing a second set of variables. The system generates a second trajectory based on results of the first and the second QP optimizations to control the ADV autonomously according to the second trajectory.