Mobile Interaction Classification Using RF and Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for locating and classifying the interaction type of mobile communication devices within a defined volume, such as vehicles, do not effectively differentiate between user interactions, limiting the assessment of driving safety.
Innovation Solution
A sensor fusion approach that combines localization data from radio frequency signals with various sensor measurements, including inertial, proximity, and vital signs data, to classify the type of interaction between a human user and a mobile communication device, determining its position and usage patterns over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor fusion approach is implemented to classify interaction types, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent combines multiple sensor types (accelerometers, gyroscopes, magnetometers, barometers, proximity sensors, cameras, microphones) into a unified sensor fusion system that processes data from all sensors simultaneously to classify interaction types, thereby improving measurement precision through multi-source data integration
Solution Approach 2:
The sensor fusion system serves multiple functions: it determines device position within the vehicle, classifies interaction types (driving, passenger, distracted driving), and assesses driving safety, allowing a single system to perform diverse classification tasks with different sensor subsets
2Reliability
If multiple sensor measurements are collected and processed over time, then reliability of interaction classification is improved, but loss of time increases
Solution Approach 1:
The system performs preliminary classification using readily available sensor data (accelerometer, gyroscope, magnetometer) to determine basic interaction type, then selectively collects additional sensor measurements (proximity, camera, microphone) only when needed to confirm or refine the classification, reducing overall processing time while maintaining reliability
Solution Approach 2:
The system uses a tiered approach where not all sensors are activated or processed for every classification decision. Instead, it uses a subset of sensors appropriate for the specific classification task (e.g., using only motion sensors for basic interaction type, adding proximity and visual sensors only when distracted driving is suspected), optimizing the balance between reliability and processing time
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the accuracy of interaction classification, providing insights into user activities and improving driving safety assessments by distinguishing between different types of interactions, such as engagement levels and routine activities.
Implementation Method 1
determining a position of the at least one mobile communication device relative to a frame of reference of the defined volume, based on at least one of: angle of arrival, time of flight such as Round Trip Time (RTT) and Time Difference Of Arrival (TDOA)
Implementation Method 2
received intensity of radio frequency (RF) signals transmitted by the at least one mobile communication device and received by a phone location unit located within the defined volume and comprising a transceiver and antennas
Data Source
AI summary
A system and method for classifying a type of interaction between a human user and a mobile communication device within a defined volume, based on multiple sensors. The method may include: determining a position of the mobile communication device relative to a frame of reference of the defined volume, based on: angle of arrival, time of flight, or received intensity of radio frequency (RF) signals transmitted by the mobile communication device and received by a phone location unit located within the defined volume configured to wirelessly communicate with the mobile communication device; obtaining at least one sensor measurement related to the mobile communication device from various non-RF sensors; repeating the obtaining, to yield a time series of sensor readings; and using a computer processor to classify the type of interaction into one of many predefined types of interactions, based on the position and the time series of sensor readings.