Vehicle Visibility Detection Using Maps and Multi-Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional safety systems for vehicles are limited in detecting low visibility conditions beyond 200 meters, necessitating a need for determining low visibility closer to the vehicle to enable effective counter-measures.
Innovation Solution
An apparatus and method using a machine learning model to analyze images from vehicle-mounted sensors, coupled with a map database, to determine visibility information and control vehicle functions based on real-time environmental conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional safety systems are used to detect low visibility conditions, then the detection range is limited to 200 meters, but this limitation prevents effective counter-measures in dense fog or heavy rainfall conditions where visibility is less than 200 meters
Solution Approach 1:
The patent combines multiple sensor types (image capture sensors, laser sensors, radar sensors) to create a multi-sensor fusion system that overcomes the limitations of individual sensors. This merging allows the system to detect low visibility conditions at ranges less than 200 meters by compensating for the weaknesses of each individual sensor type through data fusion and cross-validation.
Solution Approach 2:
The safety system is designed to perform multiple detection functions using a single integrated apparatus that can operate in various weather conditions (fog, rain, snow). The system universally detects different types of hazardous conditions (low visibility, road surface conditions, obstacles) and provides counter-measures for multiple scenarios, making it adaptable to diverse environmental challenges beyond just the 200-meter visibility limitation.
2Adaptability or versatility
If the visibility detection range is extended beyond 200 meters, then more counter-measures can be implemented, but the system complexity and computational requirements increase
Solution Approach 1:
The patent segments the detection system into specialized sensor modules (image capture, laser, radar) that each handle specific aspects of environmental sensing. This segmentation allows the system to extend detection range and capabilities while managing complexity by dividing functions into manageable, specialized components rather than requiring a single complex sensor to handle all functions.
Solution Approach 2:
The system introduces intermediate processing layers including machine learning models and data fusion algorithms that mediate between raw sensor data and final detection decisions. These intermediaries simplify the overall system architecture by automatically processing and integrating data from multiple sensors, reducing the computational burden on the main control system while enabling extended detection capabilities.
3Measurement precision
If machine learning models are used to analyze sensor images for visibility determination, then accuracy in low visibility conditions improves, but processing time and computational energy consumption increase
Solution Approach 1:
The system performs preliminary processing of sensor images by extracting key features and pre-processing data before applying complex machine learning models. This preliminary action reduces the computational load during critical real-time operation, maintaining high accuracy in visibility determination while reducing processing time by preparing data in advance through feature extraction and filtering operations.
Solution Approach 2:
The system uses periodic action by implementing multi-stage processing where machine learning models are applied at different intervals and stages. Critical visibility assessments use rapid simplified models for immediate response, while more comprehensive analysis is performed periodically when time permits. This periodic application of different processing depths balances accuracy requirements with real-time response constraints.
Data Source
AI summary
An apparatus for determining low visibility of an environment around a vehicle is disclosed. The apparatus obtains location information indicating a location of the vehicle. The apparatus further detects, by using a map database, a road segment satisfying a road attribute requirement proximate to the location based on the location information. The apparatus further obtains, from the map database, at least one attribute associated with the road segment. The apparatus further determines a position of the vehicle with respect to the road segment based on the at least one road attribute. The apparatus further acquires at least one image via at least one image capture sensor equipped by the vehicle based on the position. The apparatus further determines visibility information indicative of visibility of the environment of the vehicle based on the at least one image. The apparatus further outputs the visibility information.