Autonomous Vehicle Path Planning via Station-Lateral Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in generating optimal paths that balance passenger comfort, obstacle avoidance, and adherence to lane centerlines, particularly in navigating through environments with static obstacles, as existing systems often prioritize speed over comfort and safety.
Innovation Solution
A method and system for generating optimal paths for autonomous driving vehicles using nonlinear optimization, which considers comfort, obstacle avoidance, and lane centerline adherence by employing a station-lateral coordinate system and a cost function that weighs cumulative lateral distances, first-order, and second-order lateral rates of change, while avoiding static obstacles and limiting lateral jerk.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing path planning systems prioritize speed, then navigation efficiency is improved, but passenger comfort and safety deteriorate due to excessive lateral deviations and jerky movements
Solution Approach 1:
The patent transforms the path planning problem by changing the parameter space from standard Cartesian coordinates to station-lateral coordinates. This parameter transformation allows the system to independently control longitudinal position (station) and lateral deviation (lateral distance), enabling optimization of both speed and comfort through separate cost function terms for longitudinal progress and lateral smoothness
Solution Approach 2:
The cost function applies different weighting factors to different segments of the path based on local conditions. By assigning higher weights to lateral smoothness terms in regions where passenger comfort is prioritized and to longitudinal progress terms where speed is prioritized, the system achieves locally optimized path segments that collectively resolve the contradiction between navigation efficiency and comfort
2Stability of the object's composition
If path planning emphasizes staying close to lane centerline, then lane adherence is improved, but obstacle avoidance capability deteriorates due to reduced flexibility in path deviation
Solution Approach 1:
The system dynamically adjusts the desired lateral offset from the lane centerline based on real-time obstacle detection and path optimization results. Rather than rigidly following the lane centerline, the optimized path allows controlled deviations when obstacles are present, while maintaining lane adherence in obstacle-free regions. This dynamic adaptation resolves the contradiction between stable lane following and flexible obstacle avoidance
3Reliability
If path optimization minimizes lateral deviations, then passenger comfort is improved, but the ability to effectively avoid obstacles deteriorates due to excessive conservatism
Solution Approach 1:
The transformation to station-lateral coordinates enables decoupled optimization of longitudinal and lateral path segments. The lateral cost function terms minimize unnecessary deviations for comfort while the longitudinal terms ensure progress toward the destination. This parameter separation allows the system to achieve smooth lateral transitions for comfort while maintaining effective obstacle avoidance through longitudinal path adjustments
Solution Approach 2:
The path is segmented into discrete stations along the longitudinal direction, with lateral offset optimized independently at each station. This segmentation allows the system to minimize lateral deviations between stations for comfort while permitting necessary lateral adjustments at critical stations where obstacle avoidance requires it, thus resolving the contradiction between comfort and effectiveness
Data Source
AI summary
A method, apparatus, and system for generating an optimal path for an autonomous driving vehicle (ADV) are disclosed. The method includes receiving optimization inputs comprising an ADV starting state, a maximal lateral jerk, and static obstacle boundaries with respect to a reference line; receiving optimization constraints comprising constraints relating to the maximal lateral jerk and avoidance of one or more static obstacles; receiving a cost function associated with an optimization objective, the cost function comprising a first term relating to cumulative lateral distances, a second term relating to cumulative first order lateral rates of change, and a third term relating to cumulative second order lateral rates of change; generating planned ADV states as optimization results with nonlinear optimization, by minimizing a value of the cost function; and generating control signals to control the ADV based on the plurality of planned ADV states.