Autonomous Vehicle Trajectory Selection With Two-Stage Path Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating safely and accurately on roadways due to the need to process and interpret various types of data, including visual information, radar, lidar, GPS, and sensor data, while identifying location, avoiding obstacles, and responding to traffic signals and signs.
Innovation Solution
The system employs cameras to capture images of the environment, which are analyzed by a processor to determine navigational state information, potential trajectories, and assign rankings, ultimately selecting a planned trajectory and implementing navigational actions through actuators, integrating GPS, sensor, and map data for informed decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the system performs comprehensive analysis of all potential trajectories to ensure safe and accurate navigation, then the navigation reliability is improved, but the processing time and computational complexity increase
Solution Approach 1:
The patent segments the trajectory analysis into two distinct phases: preliminary analysis that filters trajectories based on basic safety and feasibility criteria, and secondary analysis that performs detailed evaluation only on the reduced subset of promising trajectories. This segmentation resolves the contradiction by eliminating unnecessary comprehensive analysis of all trajectories while maintaining navigation reliability through the two-stage filtering approach.
Solution Approach 2:
The system performs preliminary analysis of all potential trajectories before the detailed secondary analysis. This preliminary action identifies and eliminates obviously unsafe or infeasible trajectories early in the process, reducing the computational burden of subsequent detailed analysis while ensuring that only viable options proceed to thorough evaluation, thus balancing reliability with processing efficiency.
2Measurement precision
If the system analyzes multiple types of data including visual information, radar, lidar, GPS, and sensor data to improve navigation accuracy, then the measurement precision is improved, but the device complexity increases
Solution Approach 1:
The patent merges multiple data sources including camera images, radar data, lidar data, GPS information, and sensor readings into a unified navigational state representation. This integration allows the system to achieve high navigation accuracy by combining complementary information from different sensors while managing complexity through unified processing architecture that handles all data types consistently.
Solution Approach 2:
The navigational state information serves as a universal data structure that encapsulates information from multiple sensor types in a standardized format. This multi-functional representation allows the same processing pipeline to handle diverse data sources, improving measurement precision through comprehensive data integration while reducing device complexity by avoiding separate processing paths for each sensor type.
3Measurement precision
If the system performs detailed secondary analysis on a subset of trajectories to select the optimal path, then the trajectory selection accuracy is improved, but the processing time for the subset increases
Solution Approach 1:
The patent applies different levels of analysis quality to different trajectories: preliminary analysis provides basic filtering for all trajectories, while secondary analysis provides detailed evaluation only for the subset of trajectories that passed the preliminary filter. This local quality approach ensures high selection accuracy for the final trajectory choice while minimizing total processing time by applying intensive analysis only where necessary.
Data Source
AI summary
Systems and methods are provided for navigating a host vehicle. A navigation system for the host vehicle may include at least one processor programmed to receive images representative of an environment of the host vehicle; analyze at least one of the images to identify navigational state information associated with the host vehicle; determine a plurality of first potential navigational actions for the host vehicle based on the navigational state information; determine respective future states for the plurality of first potential navigational actions; determine a plurality of second potential navigational actions for the host vehicle based on the determined respective future states; select, based on the plurality of second potential navigational actions, one of the plurality of first potential navigational actions; and cause an adjustment of a navigational actuator of the host vehicle to implement the selected one of the plurality of first potential navigational actions.