Dynamic Lane Segmentation for Autonomous Vehicle Trajectory Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating through changing environments due to inaccurate prior maps and limited use of sensor data, leading to incorrect lane detection and potential safety issues, especially in conditions with narrowed lanes, construction, or missing lane markings.
Innovation Solution
A method and system for dynamically creating a trajectory by segmenting the driving surface using sensor data, assigning nodes to generate segmentation polylines, and updating the nominal path to navigate the vehicle safely through lane changes and intersections, incorporating real-time lane segmentation and curvature optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If autonomous vehicles rely on pre-generated trajectories and prior maps for navigation, then the navigation system is simple to implement, but lane detection accuracy deteriorates in changing road conditions such as construction zones, narrowed lanes, or missing lane markings
Solution Approach 1:
The patent segments the driving surface into multiple lanes by detecting lane markings and dividing the road into distinct navigable areas. This segmentation allows the vehicle to accurately identify and navigate within specific lanes even when prior maps are outdated or road conditions have changed, directly improving lane detection accuracy without requiring complete system redesign
Solution Approach 2:
The navigation system dynamically updates lane boundaries and trajectories in real-time based on current sensor data rather than relying solely on static pre-generated maps. This dynamic adaptation enables the system to handle changing road conditions such as construction zones and narrowed lanes, maintaining high detection accuracy while managing complexity through incremental updates
2Adaptability or versatility
If autonomous vehicles use static pre-generated trajectories for navigation, then computational resources are conserved, but adaptability to changing road conditions deteriorates
Solution Approach 1:
The system performs preliminary segmentation of the driving surface and identification of nominal paths in advance, creating a structured framework for navigation. This preliminary action reduces the computational burden during real-time operation by pre-organizing road data into manageable segments that can be quickly adapted to changing conditions without requiring full re-computation
Solution Approach 2:
The trajectory is updated dynamically based on current sensor data and detected lane changes, allowing the system to adapt to construction zones, narrowed lanes, and other changing conditions. The dynamic update mechanism focuses computational resources only on relevant changes rather than re-processing entire trajectories, managing energy consumption while maintaining adaptability
3Reliability
If autonomous vehicles rely on minimally delineated lane markings for navigation, then road infrastructure requirements are reduced, but navigation reliability deteriorates in difficult identification conditions
Solution Approach 1:
The system merges multiple sensor data sources including camera images, LIDAR point clouds, and inertial measurement unit data to create a comprehensive view of the driving surface. This multi-sensor fusion compensates for poorly defined or missing lane markings by cross-validating information from different modalities, significantly improving navigation reliability under difficult detection conditions
Solution Approach 2:
The patent introduces an intermediary processing layer that transforms raw sensor data into structured representations of the driving surface, including inferred lane boundaries and nominal paths. This intermediary representation acts as a mediator between ambiguous sensor inputs and navigation decisions, enhancing reliability by providing a stable intermediate model even when lane markings are difficult to detect
Data Source
AI summary
This disclosure provides methods and systems for dynamically creating a trajectory for navigating a vehicle. The method may include receiving sensor data from at least one sensor of the autonomous vehicle, the sensor data representative of a driving surface in a field of view of the autonomous vehicle; segmenting a portion of the driving surface in the field of view of the autonomous vehicle by determining nominal path based at least in part on the image data; assigning a plurality of nodes to at least a portion of the nominal path; associating the plurality of the nodes assigned to the nominal path with a line to generate at least one segmentation polyline; determining updated nominal path by fitting the each of the plurality of segmentation lines to the nominal path; generating a trajectory based on the updated nominal path; and navigating the autonomous vehicle according to the trajectory.