Trajectory Prediction Using Interaction Graphs for Ego Vehicle Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for trajectory prediction and ego vehicle control lack precision in accounting for interactions among road users and adherence to traffic rules, leading to inefficient and potentially unsafe trajectory planning.
Innovation Solution
A computer-implemented method using graph representations to combine trajectory data and map data, generating interaction graphs to predict future motion trajectories, and creating anchor paths that adhere to traffic rules, while classifying road users and considering roadway characteristics to refine predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional trajectory prediction methods are used, then the computational process is simpler, but the prediction precision and ability to account for road user interactions deteriorates
Solution Approach 1:
The patent segments the complex trajectory prediction problem into distinct components: trajectory data processing, map data processing, and interaction graph generation. Each component is handled separately through graph representations, allowing the system to manage complexity while maintaining high prediction precision by addressing each aspect independently rather than as a monolithic problem
Solution Approach 2:
The interaction graph serves as an intermediary data structure that bridges trajectory data and map data. This graph representation mediates between raw input data and final predictions, enabling the system to account for road user interactions without requiring direct complex processing of all raw data, thus improving precision while managing computational complexity
2Measurement precision
If detailed interaction data among road users is considered, then the trajectory prediction accuracy improves, but the data processing time and computational load increases
Solution Approach 1:
The system performs preliminary processing of trajectory data and map data into graph representations before actual prediction occurs. Interaction graphs are pre-computed to capture spatial and temporal relationships, so that during prediction, the system can quickly query these pre-processed structures rather than computing interactions from raw data in real-time, thereby improving accuracy without excessive processing time
Solution Approach 2:
The patent transforms raw trajectory and map data into graph representations with specific parameters (nodes for positions, edges for relationships). This parameter transformation allows the system to efficiently represent and process complex interaction data in a standardized format that optimizes both prediction accuracy and processing efficiency
Data Source
AI summary
A computer-implemented method for trajectory prediction. The method includes: receiving trajectory data of motion trajectories of road users arranged in a surrounding area of the ego vehicle by a prediction module, wherein the trajectory data are arranged in a graph representation; receiving map data of a map representation mapping the surrounding area of the ego vehicle by the prediction module; generating an interaction graph representation for the plurality of road users based on the trajectory data of the road users and roadway location information of the map representation by the prediction module; and predicting a future motion trajectory to be executed for at least one other road user based on the trajectory data, the map data, and the interaction graph representation of the road users by the prediction module.