Building Transport Mode Detection via Mobility Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation and mapping applications face challenges in accurately determining transport modes within buildings, such as elevators, escalators, and stairs, which affects user experience and the accuracy of mobility patterns analysis.
Innovation Solution
A system utilizing a trained machine learning model to analyze sensor data from user equipment to determine transport modes by extracting mobility features and calculating confidence scores, allowing for the classification of transport modes like stairs, elevators, or escalators, and updating map data accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor data from mobile phones is collected and analyzed to determine transport modes, then mobility pattern analysis and navigation accuracy are improved, but data processing complexity and computational requirements increase
Solution Approach 1:
The system segments the data processing into distinct modules: sensor data collection from mobile phones, feature extraction (mobility features), machine learning classification, and map data updating. This segmentation allows each component to be optimized independently and reduces overall system complexity while maintaining high detection accuracy.
Solution Approach 2:
A trained machine learning model serves as an intermediary between raw sensor data and final transport mode classification. The model processes mobility features and outputs transport mode predictions, acting as a computational mediator that simplifies the decision-making process and improves detection accuracy without requiring complex real-time analysis.
2Measurement precision
If mobility features are extracted and classified using machine learning models, then transport mode determination accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The system performs preliminary feature extraction and mobility feature computation during the data collection phase, preparing the data in advance for classification. This preliminary processing reduces the computational burden during the critical classification step and speeds up the overall processing time while maintaining high accuracy through pre-computed mobility features.
Solution Approach 2:
The machine learning model transforms complex sensor data into simplified mobility feature parameters that capture essential transport mode characteristics. By changing the data representation from raw sensor values to processed mobility features (such as velocity, acceleration, direction changes), the system reduces processing time while maintaining classification accuracy.
3Measurement precision
If map data is updated based on determined transport modes, then navigation accuracy and user experience are improved, but system responsiveness and data freshness requirements increase
Solution Approach 1:
The system implements a feedback mechanism where determined transport modes are used to update map data, which in turn improves navigation accuracy for subsequent queries. The feedback loop continuously refines the map data based on observed transport mode usage patterns, creating an increasingly accurate navigation system that adapts to real-world conditions.
Solution Approach 2:
The map data structure is designed to be dynamic and adaptable, allowing incremental updates based on new transport mode determinations. Rather than requiring complete map recalculations, the system dynamically adjusts specific map elements (such as adding elevator or escalator information) based on the determined transport modes, maintaining responsiveness while improving accuracy.
Data Source
AI summary
A system, a method and a computer program product are provided for determining one or more transport modes for one or more buildings in a geographic region, using a machine learning model. The system may include at least one memory configured to store computer executable instructions and at least one processor configured to execute the computer executable instructions to obtain sensor data associated with the one or more buildings for time duration. The processor may be configured to determine mobility features associated with the one or more buildings, based on the sensor data. The processor may be further configured to determine, using a trained machine learning model, the one or more transport modes for the one or more buildings, based on the mobility features.