Temporal Occupancy Flow Graphs for Accurate Vehicle Motion Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for vehicle motion forecasting using graph neural networks often suffer from bias in interaction information due to the order of operations, leading to decreased accuracy in prediction outcomes, as they typically require separate attention layers for vehicle-to-vehicle and vehicle-to-lane interactions.
Innovation Solution
A method and system that generates a lane graph structure from raw map data, establishes occupancy flow graphs for consecutive vehicle frames, and constructs a temporal occupancy flow graph with temporal edges, allowing for simultaneous feature aggregation of vehicle-to-vehicle and vehicle-to-lane interactions to improve prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If separate attention layers are used to extract vehicle-to-vehicle interaction and vehicle-to-lane interaction, then the interactions can be processed separately, but information bias occurs due to the order of operations leading to decreased prediction accuracy
Solution Approach 1:
The patent merges the extraction of vehicle-to-vehicle interaction and vehicle-to-lane interaction into a single unified attention layer. This is achieved by constructing a graph where both vehicle nodes and lane nodes coexist, and performing a single attention operation that simultaneously captures both types of interactions, thereby eliminating the information bias caused by sequential processing in separate layers.
Solution Approach 2:
The patent introduces a new dimensional structure by elevating lane segments to the same node level as vehicles in the graph representation. This dimensional change allows both vehicle-to-vehicle and vehicle-to-lane interactions to be processed in the same attention layer, transforming the problem from sequential processing to simultaneous processing through graph neural network operations.
2Adaptability or versatility
If multiple separate attention layers are designed for different interaction types, then each interaction can be extracted independently, but the device complexity increases and processing efficiency decreases
Solution Approach 1:
The patent creates a universal attention layer that serves multiple functions simultaneously. By representing both vehicles and lanes as nodes in the same graph structure, a single attention operation can extract both vehicle-to-vehicle interactions and vehicle-to-lane interactions, making the attention layer multi-functional and eliminating the need for multiple specialized layers.
Solution Approach 2:
The patent combines multiple interaction extraction functions into a single unified attention mechanism. Instead of having separate attention layers for vehicle-to-vehicle and vehicle-to-lane interactions, the invention merges these functions by constructing a joint graph representation where both interaction types are captured in one attention operation, thereby reducing device complexity.
Data Source
AI summary
A method for vehicle motion forecasting includes the following steps. A lane graph structure is generated according to a raw map data. Multiple occupancy flow graphs which are homogeneous to data format of the lane graph structure are established according to trajectory data of a plurality of vehicles in multiple consecutive frames and the lane graph structure. Multiple temporal edges between the occupancy flow graphs are established according to the trajectory data of the vehicles in the consecutive frames to construct a temporal occupancy flow graph. Feature aggregation is performed on the temporal occupancy flow graph to generate multiple updated node features, and a motion forecasting of an ego-vehicle is generated according to the updated node features.