Trajectory Variance Estimation for Map-Based Vehicle Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inaccurate scene understanding in autonomous vehicles can lead to sub-optimal or unsafe trajectory planning due to discrepancies between map data and actual road user behaviors, resulting in potential collisions.
Innovation Solution
The implementation of map-based prediction with variance estimation, which considers both map-based predictions and variances to improve the prediction of future trajectories of other road users by utilizing driveline mean and variance data to generate a map-based variance input for vehicle control systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If map-based prediction is used to plan vehicle trajectory, then trajectory planning efficiency is improved, but accuracy deteriorates due to discrepancies between map data and actual road user behaviors
Solution Approach 1:
The system collects actual trajectory data from multiple road users and uses it to update and refine the map-based prediction model. This feedback loop allows the system to learn from real-world deviations and improve prediction accuracy over time while maintaining the efficiency of map-based planning.
Solution Approach 2:
The system performs preliminary trajectory planning based on map data to establish an initial efficient path, then applies corrections based on observed road user behaviors. This allows the system to benefit from both the efficiency of pre-planned routes and the accuracy of real-time behavioral adjustments.
2Measurement precision
If variance estimation is added to account for road user behavior variations, then prediction accuracy is improved, but system complexity increases
Solution Approach 1:
The system introduces variance as an additional parameter to characterize road user behavior deviations. By modeling behavior variations through statistical parameters (mean and variance) rather than complex behavioral rules, the system achieves improved accuracy while maintaining relatively simple system architecture.
Solution Approach 2:
The system uses statistical variance as an intermediary to bridge the gap between simple map-based predictions and complex real-world behaviors. This intermediary parameter allows the system to account for behavioral variations without directly implementing complex behavioral models.
3Measurement precision
If multiple road user trajectories are tracked and analyzed, then scene understanding accuracy is improved, but computational load increases
Solution Approach 1:
The system extracts only the essential features from multiple trajectory data sets - specifically the mean and variance parameters - rather than processing complete trajectory histories. This extraction approach maintains scene understanding accuracy while significantly reducing computational requirements.
Solution Approach 2:
The system processes data from multiple road users but focuses computation only on deriving statistical parameters (mean and variance) rather than analyzing every aspect of each trajectory. This partial processing approach achieves sufficient scene understanding with reduced computational energy consumption.
Data Source
AI summary
Observed driveline mean and variance data are used for determining the variance of a trajectory of tracked objects for use by a host vehicle. A map of a portion of a vehicle transportation network are determined, wherein the map is comprised of observed driveline mean and variance data for one or more map points. At least one trajectory of a tracked object is predicted, wherein a trajectory includes a series of location each corresponding to a respective predicted position of the tracked object at a future time. A map-based variance is generated for the location of the trajectory using a smoothed curvature of the trajectory within the map. A control system of the vehicle operates the vehicle using the map-based variance as input.