Third Order Polynomial Vehicle Trajectory Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle systems using second order polynomial-based course predictions often overestimate curves, leading to inaccurate trajectory modeling and potential safety issues during maneuvers like exiting curves or lane changes.
Innovation Solution
Implementing a method that uses third order polynomials to predict vehicle trajectories when steering towards a drive-straight state, transitioning to non-third-order models after exiting curves, with a controller determining the appropriate polynomial based on vehicle dynamics and navigation data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If second order polynomial-based course prediction is used, then the system complexity is low, but the trajectory prediction accuracy deteriorates due to overestimation of curves
Solution Approach 1:
The patent changes the mathematical model parameter from second order polynomial to third order polynomial to improve trajectory prediction accuracy. This parameter change allows the system to better capture the curvature characteristics of vehicle trajectories during maneuvers like exiting curves or lane changes, reducing the overestimation problem while maintaining computational efficiency.
2Measurement precision
If third order polynomial is used for all steering situations, then the trajectory prediction accuracy is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent segments the steering situations into different categories: situations where the vehicle is steering toward a drive-straight state (using third order polynomial) and other situations (using non-third-order functions). This segmentation allows the system to apply the computationally intensive third order model only when necessary for accuracy, while using simpler models for other cases, thus balancing accuracy with processing efficiency.
Solution Approach 2:
The patent implements a dynamic model selection mechanism that adapts the polynomial order based on the current steering state of the vehicle. The controller dynamically determines whether to use third order or non-third-order functions based on real-time vehicle dynamics and navigation data, optimizing computational resources while maintaining prediction accuracy when needed.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and systems for performing vehicle driver assistance. One method includes determining, at a controller, whether the vehicle is steering toward a drive-straight state and when the vehicle is steering toward the drive-straight state, calculating, at the controller, a predicted course trajectory using a third order polynomial. The method also includes calculating, at the controller, the predicted course trajectory using a non-third-order function when the vehicle is not steering toward the drive-straight state. In addition, the method includes performing the driver assistance based on the predicted course trajectory.