Dynamic Lane Expansion for Autonomous Vehicle Obstacle Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating through environments with dynamic and static objects, as existing methods require costly computational or temporal adjustments, such as lane changes, to avoid obstacles, which can be inefficient and unsafe.
Innovation Solution
The system determines an expanded drivable region by analyzing map data and sensor information to identify candidate regions and their classifications, allowing the vehicle to expand its initial region and plan a trajectory that circumvents obstacles without lane changes, using expansion regions to safely traverse the environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the autonomous vehicle uses traditional lane-based navigation methods to avoid obstacles, then the vehicle can maintain safe distance from obstacles, but the computational cost and time required for lane change decisions increase significantly
Solution Approach 1:
The patent segments the drivable space into multiple candidate regions with different classifications (e.g., drivable, partially drivable, non-drivable). This segmentation allows the planning system to evaluate multiple potential paths simultaneously rather than performing sequential lane change decisions, reducing computational time while maintaining safety through systematic evaluation of each region's suitability for traversal
Solution Approach 2:
The patent performs preliminary classification and evaluation of candidate regions before final path selection. By pre-processing the environment map to identify and categorize different regions in advance, the system reduces the computational burden during real-time navigation, allowing faster decision-making while ensuring safety constraints are met
2Productivity
If the autonomous vehicle performs frequent lane changes to avoid obstacles, then the vehicle can maintain optimal path, but the operational smoothness and passenger comfort deteriorate
Solution Approach 1:
The patent implements dynamic region expansion where the drivable region is adaptively expanded based on real-time obstacle detection and candidate region evaluation. This dynamic approach allows the vehicle to smoothly transition into expanded regions when obstacles are detected, avoiding abrupt lane changes while maintaining path optimization through continuous adaptation of the drivable space boundaries
3Productivity
If the autonomous vehicle expands the drivable region to include candidate regions, then the vehicle can traverse obstacles more efficiently, but the complexity of region determination and classification increases
Solution Approach 1:
The patent changes the parameter of region classification by introducing multiple predefined region types (drivable, partially drivable, non-drivable) with specific characteristics. This parameter-based classification system simplifies the complexity of region determination by providing clear criteria for each region type, enabling efficient traversal decisions without requiring complex real-time analysis of every possible path
Data Source
AI summary
A vehicle can determine a drivable region of an environment and determine an expansion region to expand the drivable region. Candidate regions can be identified in the environment and portions of the candidate regions which may be used for planning can be determined. The width of such a portion can meet or exceed a threshold and an expansion region can be determined. The expansion region can be associated with the drivable region to determine an expanded drivable region. The vehicle can traverse the environment based on the expanded drivable region to avoid, for example, an object in the environment while maintaining a safe distance the object and/or other entities in the environment.