Vehicle Trajectory Prediction Using Vectorized Interaction Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting vehicle trajectories in autonomous driving struggle to fully utilize and consider interactions between a self-vehicle and its surrounding environment, due to the limited receptive field of convolutional neural networks.
Innovation Solution
A method that obtains vectorized features from perceived information of autonomous vehicles, including self-vehicle, surrounding vehicles, and road network information, and performs feature interaction and fusion to predict multimodal trajectories, using an encoder-decoder network model to generate accurate trajectory predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If convolutional neural network is used for trajectory prediction, then image features can be processed, but the receptive field is limited and cannot fully utilize interaction between self-vehicle and surrounding environment
Solution Approach 1:
The patent transforms the trajectory prediction approach from image-based convolutional processing to vectorized feature representation. By converting spatial image data into vectorized features that encode positional, velocity, and trajectory information, the system extends the effective receptive field beyond the limitations of convolutional kernels, enabling comprehensive utilization of interactions between self-vehicle and surrounding environment across the entire field of view.
2Measurement precision
If vectorized features of multiple target vehicles are obtained and feature interaction is performed, then trajectory prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments the trajectory prediction task by processing each target vehicle independently through separate encoder-decoder networks. Each network takes vectorized features of a specific target vehicle as input and predicts its trajectory separately. This segmentation allows for parallel processing of multiple vehicles, improving prediction accuracy through individualized analysis while managing computational complexity through modular architecture that can be scaled based on the number of detected vehicles.
Data Source
AI summary
A method for predicting a vehicle trajectory, a control device, a readable storage medium, and a vehicle are proved to solve the problem of effectively predicting a multimodal trajectory of a vehicle. According to the disclosure, perceived information of an autonomous vehicle is converted into vectorized features of target vehicles, where the target vehicles include the autonomous vehicle and a plurality of first surrounding vehicles; and trajectory prediction results of the target vehicles are obtained based on the vectorized features. The vectorized features in the disclosure are obtained based on the perceived information of the autonomous vehicle, which enables the vectorized features to include rich vehicle trajectory information and environment information of the target vehicle; and interaction and fusion are performed on the vectorized features of the target vehicle, so that a plurality of vehicle trajectories of each target vehicle can be effectively predicted.