Vehicle Trajectory Prediction Using Interaction Features and Map Context
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting the future traveling trajectory of a target vehicle in intelligent driving systems have relatively poor precision, particularly when considering the impact of surrounding vehicles.
Innovation Solution
A trajectory prediction method that involves obtaining historical trajectory information of a target vehicle and associated vehicles, predicting location distribution information for both vehicles in a future time period based on historical data and map information, determining interaction features between the vehicles, and using these features to predict the target vehicle's trajectory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods are used to predict future traveling trajectory of a target vehicle, then the prediction process is simple, but the prediction precision is poor
Solution Approach 1:
The prediction process is divided into multiple independent modules: historical trajectory analysis module, map information processing module, interaction feature extraction module, and trajectory prediction module. Each module processes specific aspects separately, improving overall precision while maintaining manageable complexity through modular design.
Solution Approach 2:
The method transitions from analyzing only historical trajectory data to incorporating map information and interaction features as additional dimensions. This multi-dimensional approach includes spatial dimensions (map elements like intersections, lane lines, travelable areas), temporal dimensions (historical trajectories), and interaction dimensions (vehicle-vehicle and vehicle-environment interactions), significantly improving prediction precision.
2Measurement precision
If the impact of surrounding vehicles is considered in trajectory prediction, then the prediction accuracy improves, but the computational complexity increases
Solution Approach 1:
The method extracts only the essential interaction features between vehicles and between vehicles and the environment, rather than processing all possible data. Key interaction features such as cutting-in behavior, giving way, accelerating past, and avoidance maneuvers are identified and extracted for analysis, reducing computational complexity while maintaining prediction accuracy.
Solution Approach 2:
Map information and historical trajectory data are pre-processed before the actual prediction. Map elements (intersections, lane lines, travelable areas) are pre-identified and structured. Historical trajectories are pre-analyzed to establish baseline patterns, so that during prediction, only the critical interaction features need to be computed, reducing real-time computational complexity.
3Measurement precision
If location distribution information and interaction features are used to predict trajectory, then prediction precision improves, but the data processing requirements increase
Solution Approach 1:
The method focuses computational resources on locally relevant data. Map information processing is localized to the specific traveling environment of the target vehicle and associated vehicles. Interaction features are extracted only for vehicle pairs that are spatially close and potentially interacting, rather than processing all vehicle combinations globally, reducing overall data processing volume.
Solution Approach 2:
The method processes sufficient (but not excessive) data to achieve high precision. Location distribution information is computed for the necessary time horizon and spatial range to capture meaningful interaction patterns, without computing unnecessary redundant data. This partial action approach achieves high precision while controlling data processing requirements.
Data Source
AI summary
This application provides a trajectory prediction method and an apparatus therefor, a medium, a program product, and an electronic device. An example trajectory prediction method includes: obtaining historical trajectory information of a target vehicle and an associated vehicle; predicting location distribution information of the target vehicle and the associated vehicle based on the historical trajectory information and map information; determining an interaction feature between the target vehicle and the associated vehicle based on the location distribution information; and determining a traveling trajectory of the target vehicle based on the interaction feature, the location distribution information of the target vehicle, and the map information.