Driving Path Prediction Using Trust-Weighted Data Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting a driving path of a motor vehicle are not sufficiently accurate and robust, leading to potential errors in vehicle assistance functions and automated guidance.
Innovation Solution
A method that involves ascertaining the environmental scenario ahead of the vehicle, determining the trustworthiness of various data sources based on the scenario, and amalgamating data weighted by their reliability to predict the driving path, using a combination of sensor data, map data, and learned trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional running-dynamics data alone is used for predicting driving path, then the prediction can be made with simple methodology, but the accuracy and reliability of the prediction deteriorates leading to errors in assistance functions
Solution Approach 1:
The patent combines multiple data sources (sensor data, map data, learned trajectories) with different trustworthiness weights to predict the driving path. This merging of diverse data sources improves prediction accuracy by compensating for the limitations of individual data sources while maintaining a manageable system through standardized integration procedures.
Solution Approach 2:
The patent dynamically adjusts the trustworthiness parameters of different data sources based on the environmental scenario. By changing the weight parameters assigned to each data source according to situational context, the system achieves higher prediction accuracy without requiring completely different methodologies for each scenario.
2Reliability
If multiple data sources are amalgamated with equal weighting, then comprehensive data coverage is achieved, but the reliability of the prediction deteriorates due to inclusion of untrustworthy data in certain scenarios
Solution Approach 1:
The patent assigns different trustworthiness weights to different data sources based on the specific environmental scenario. Instead of uniform treatment, each data source receives localized quality assessment tailored to the current situation, improving reliability by emphasizing trustworthy data while downweighting or excluding untrustworthy data sources in specific contexts.
Solution Approach 2:
The system performs preliminary assessment of data source trustworthiness before amalgamating the data. By evaluating and weighting data sources in advance based on the environmental scenario, the system ensures that only reliable data contributes to the prediction, avoiding the need for complex post-processing to correct unreliable predictions.
3Adaptability or versatility
If data sources are selected based on environmental scenario, then the adaptability of the prediction system improves, but the complexity of scenario recognition and data selection increases
Solution Approach 1:
The patent implements dynamic selection and weighting of data sources based on the recognized environmental scenario. The system adapts its data fusion strategy in real-time according to the situation, improving environmental adaptability while using predefined scenario categories to manage the complexity of recognition and selection processes.
Data Source
AI summary
A method and an assistance system predicts a driving path of a motor vehicle. According to the method, a respective surroundings scenario lying ahead of the motor vehicle in the driving direction is ascertained. The trustworthiness of a plurality of different data sources and/or of the data which originates therefrom and on the basis of which the driving path can be predicted are ascertained for the surroundings scenario. A plurality of the data originating from the different data sources is then fused together in a weighted manner according to the ascertained trustworthiness and is used to predict the driving path of the motor vehicle.