Map-Fused Lane Localization for Target Vehicle Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated driving systems face challenges in accurately detecting and tracking target vehicles, especially in complex environments, due to limitations in sensor accuracy and field-of-view, which can lead to delayed or inaccurate responses in scenarios like lane changes and intersections.
Innovation Solution
The implementation of a target acquisition system with real-time lane localization using a controller-executable algorithm that employs map-based absolute lane assignment, combining sensor data with map data to calculate precise lane assignments for target vehicles, allowing for dynamic reclassification and improved mission planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor-based target detection is used, then target vehicle detection is achieved, but detection accuracy and field-of-view are limited
Solution Approach 1:
The patent introduces map data as an intermediary element that mediates between sensor detections and lane assignment decisions. By fusing sensor-based target detections with pre-stored map data containing lane geometry and road structure information, the system overcomes the limitations of sensor field-of-view and accuracy. The map data serves as a reference framework that extends the effective detection range and provides contextual information about lane configurations, intersections, and road topology that sensors alone cannot capture.
Solution Approach 2:
The patent merges sensor-based target detection with map-based lane assignment to create a hybrid system. The controller fuses real-time sensor data with pre-stored map data to determine target lane assignments. This combination allows the system to leverage the real-time responsiveness of sensors with the comprehensive spatial coverage and accuracy of map data, achieving both wide field-of-view coverage and high detection accuracy simultaneously.
2Measurement precision
If map-based absolute lane assignment is implemented, then lane assignment accuracy is improved, but system complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-storing map data containing lane geometry, road structure, and lane connectivity information before the driving scenario occurs. This pre-processing of environmental information allows the system to quickly perform lane assignments during real-time operation without complex calculations. The map data is prepared in advance and organized for efficient querying, reducing the computational burden during actual target tracking and lane assignment operations.
Solution Approach 2:
The patent uses map data as a simplified copy or representation of the actual road environment. Instead of processing complex real-time sensor data from all directions to determine lane assignments, the system references the pre-stored map copy of the road layout. This copy contains essential lane geometry and topology information that enables accurate lane assignments without requiring complex real-time environmental modeling.
3Speed
If real-time lane localization is performed, then responsiveness is improved, but computational load increases
Solution Approach 1:
The patent segments the lane assignment problem into discrete, manageable components based on pre-defined map data structures. The controller processes target detections and determines lane assignments by referencing segmented map information about lane geometry and road features. This segmentation allows for efficient, localized computations rather than processing the entire road network, reducing computational load while maintaining real-time responsiveness.
Solution Approach 2:
By pre-processing and storing map data in an optimized format during preliminary actions (map creation and updates), the system eliminates the need for complex real-time computations during target tracking. The preliminary organization of map data into lane-specific structures enables quick lookups and comparisons during real-time operation, achieving high responsiveness with minimal computational load during critical driving moments.
Data Source
AI summary
A method for controlling operation of a motor vehicle includes an electronic controller receiving, e.g., from a vehicle-mounted sensor array, sensor data with dynamics information for a target vehicle and, using the received sensor data, predicting a lane assignment for the target vehicle on a road segment proximate the host vehicle. The electronic controller also receives map data with roadway information for the road segment; the controller fuses the sensor and map data to construct a polynomial overlay for a host lane of the road segment across which travels the host vehicle. A piecewise linearized road map of the host lane is constructed and combined with the predicted lane assignment and polynomial overlay to calculate a lane assignment for the target vehicle. The controller then transmits one or more command signals to a resident vehicle system to execute one or more control operations using the target vehicle's calculated lane assignment.