Visibility Confidence Modeling for Multi-Sensor Occlusion Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous systems face challenges in accurately perceiving their environment due to sensor blockage, blur, and occlusion, which can compromise the reliability of sensor data, especially when multiple sensors are involved, leading to inefficiencies in control and decision-making processes.
Innovation Solution
A visibility confidence model is generated using neural networks to assess the confidence levels of sensor data across multiple sensors, providing a top-down view of the aggregate field of view and subdividing it into sections to determine the reliability of sensor data based on faults, errors, gross-level and fine-level degradations, and occlusions, allowing systems to make informed control decisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional confidence modelling is performed on individual sensors, then sensor degradation can be predicted, but other systems cannot easily query individual sensor models before making determinations
Solution Approach 1:
The patent combines individual sensor confidence models into a unified visibility confidence model that aggregates data from multiple sensors. This merged model allows any system to query overall visibility confidence without needing to access or understand individual sensor models, resolving the contradiction between maintaining reliable sensor degradation prediction and enabling easy system-wide queryability.
2Reliability
If multiple sensors are used to perceive the environment, then perception coverage is improved, but the complexity of determining sensor blindness and reliability increases
Solution Approach 1:
The patent segments the complex multi-sensor perception problem into manageable components by creating individual confidence models for each sensor, then combining them into an aggregated visibility confidence model. This segmentation approach maintains the reliability benefits of multiple sensors while simplifying the overall determination process through modular, queryable confidence assessments.
Solution Approach 2:
The visibility confidence model serves as an intermediary layer between multiple sensors and the systems that need to make decisions. Instead of requiring direct access to complex individual sensor models, systems can query the intermediary visibility confidence model to determine sensor blindness and reliability, reducing overall system complexity.
3Productivity
If sensor data is used without assessing confidence levels, then processing speed is maintained, but control decisions may be made based on unreliable sensor data
Solution Approach 1:
The patent implements preliminary action by pre-computing and maintaining confidence models for each sensor and an aggregated visibility confidence model before actual control decisions are needed. This allows systems to quickly query pre-assessed confidence levels without performing complex real-time analysis, thereby maintaining processing speed while ensuring control decisions are based on reliable sensor data assessments.
Data Source
AI summary
Embodiments of the present disclosure may include a method and system for performing one or more operations based on a visibility confidence model. In some embodiments, the method may include generating a visibility confidence model which may indicate a level of confidence in sensor data that may correspond to individual sub-sections of an aggregate field of view. In some embodiments, the levels of confidence may be determined based on one or more errors associated with an individual sensor, one or more gross-level degradations, one or more fine-level degradations, or one or more occlusions being present in the sensor data. In some embodiments, the method may additionally include performing one or more operations based on the visibility confidence model or the level of confidence corresponding to individual sub-areas of the aggregate field of view.