Vehicle Cabin Event Detection Using Multi-Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In shared vehicle services, enforcing rules and policies regarding vehicle usage is challenging due to the lack of direct interaction between customers and operators, leading to difficulties in detecting and addressing violations autonomously.
Innovation Solution
A system that includes gas sensors, particulate matter sensors, and image sensors connected to a controller, which generates time series datasets and uses machine learning models to detect and classify events within the vehicle cabin, allowing for autonomous enforcement of rules and policies by notifying operators.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous detection systems are implemented in shared vehicles, then enforcement of vehicle usage rules is improved, but device complexity increases
Solution Approach 1:
The patent combines multiple sensing modalities (gas sensors, PM sensors, image sensors) into a unified detection system that works together to identify vehicle policy violations. This merging approach improves detection reliability by cross-validating events across multiple sensor types while managing complexity through integrated processing in a single controller.
Solution Approach 2:
The detection system is designed to monitor multiple types of events simultaneously (smoking, vaping, eating, drinking, cleanliness issues) using a single multi-functional platform. The controller processes data from various sensors to detect different violation types, reducing the need for separate dedicated systems for each event type.
2Measurement precision
If multiple sensors and machine learning models are used for event detection, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The controller serves as an intermediary that receives raw data from multiple sensors (gas, PM, image), processes this data through machine learning models, and generates event determinations. This intermediary layer manages the complexity of coordinating multiple sensors while improving measurement precision through data fusion and pattern recognition.
Solution Approach 2:
The system pre-processes sensor data by generating time series datasets and applying machine learning models before final event determination. This preliminary processing of gas and PM sensor data helps filter noise and identify patterns, improving detection accuracy while managing complexity through structured data preparation.
3Reliability
If continuous monitoring of gas and PM sensors is performed, then event detection reliability is improved, but energy consumption increases
Solution Approach 1:
The controller performs periodic analysis of time series data from gas and PM sensors rather than continuous processing. By sampling sensor data at intervals and analyzing trends over time, the system maintains detection reliability while reducing energy consumption compared to continuous real-time processing of all sensor streams.
Data Source
AI summary
A system for determining an event in a cabin of a vehicle includes a gas sensor, a PM sensor, an image sensor configured to generate an image signal corresponding to a captured image of at least a portion of the cabin, and a controller operably connected to the gas, PM, and image sensors. The controller is configured to receive the sensor signals, generate time series datasets of the sensor signals, determine an event in the cabin of the vehicle by analyzing the time series datasets using a machine learning model, and in response to determining the event in the cabin, operate the image sensor to generate image data corresponding to a captured image of at least a portion of the cabin of the vehicle.