Adaptive Existence Probability Estimation for Vehicle Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current detection systems lack the ability to adaptively model existence probabilities using information from multiple sensor sources over time, leading to inaccuracies due to changing environmental conditions, and fail to effectively distinguish between false negative and false positive cases.
Innovation Solution
A method that involves each sensor or sensor group estimating object status and existence probabilities, merging these into a fusion list, identifying and storing false negative and positive cases, and using feedback to adapt detection and clutter probability models, thereby improving the accuracy of existence probability estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional detection systems use a priori knowledge for modeling detection and clutter probabilities, then the system can operate with simple initial models, but the accuracy of existence probability estimation deteriorates under changing environmental conditions
Solution Approach 1:
The patent implements feedback loops where detection results from multiple sensors are fed back to update detection and clutter probability models. The system continuously refines these models by comparing sensor measurements with fused results, identifying false negative and false positive cases, and adapting the probability models accordingly. This feedback mechanism enables the system to adapt to changing environmental conditions while maintaining high accuracy in existence probability estimation.
Solution Approach 2:
The system performs preliminary actions by establishing initial detection and clutter probability models using a priori knowledge before actual detection begins. These preliminary models serve as starting points that are subsequently refined through feedback from real sensor data, allowing the system to begin operations immediately while continuing to improve accuracy over time.
2Measurement precision
If multiple sensors monitor a region, then the system can identify false negative and false positive cases through comparison, but the device complexity increases
Solution Approach 1:
The patent merges data from multiple sensor sources into a unified fusion list, combining detection results from different sensors to identify false negative and false positive cases. By merging sensor information and processing them through a common framework, the system achieves improved measurement precision without proportionally increasing device complexity, as the merging process itself provides the cross-validation needed to identify errors.
Solution Approach 2:
The system implements a universal fusion framework that handles multiple sensor types and processing functions within a single system architecture. This multi-functional approach allows the same infrastructure to perform detection, fusion, cross-validation, and model adaptation, reducing overall system complexity compared to having separate specialized systems for each function.
3Measurement precision
If the system sums false negative and false positive cases over time to adapt detection probabilities, then the accuracy of existence probability estimation improves, but the loss of time for processing and storing historical data increases
Solution Approach 1:
The system performs periodic updates of detection and clutter probability models by summing false negative and false positive cases over time intervals. Rather than continuous processing, the system accumulates error cases and performs model adaptation at periodic intervals, reducing computational overhead while still achieving accurate probability estimation through cumulative learning from historical data.
Data Source
AI summary
A method for improving the estimation of an existence probability of objects. The objects are detected using sensors installed in a vehicle and/or an infrastructure component. Each tracker of a sensor and/or a sensor group estimates a status of an object and its existence probability using a detection probability model. The detected objects are merged in a fusion list, and each object is assigned a state and an existence probability. Each object of the fusion list is assigned existence probabilities. Each object of the fusion list is assigned additional information indicating which sensor and/or which sensor group has/have detected the respective object in the last measuring cycle. At least sensor-specific and/or sensor-group-specific existence probabilities of fused existence probabilities and the sensor detection probability are compared in a crosscheck, and false negative cases and false positive cases are ascertained for each sensor and/or sensor groups.