Mobile Device Travel Classification Sampling Rate Dynamics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for classifying travel modes and activity types on mobile devices are inefficient, consuming excessive battery power and lacking in real-time accuracy, while also failing to effectively utilize user feedback for improving classification models.
Innovation Solution
A mobile device system that collects position, speed, and acceleration data to classify travel modes and activity types in real-time, reduces sampling rates based on identified trip patterns, and allows users to modify classifications, thereby enhancing the classification methods using user interfaces and feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the positioning module samples position data at a high sampling rate to improve classification accuracy, then measurement precision is improved, but use of energy increases
Solution Approach 1:
The system dynamically adjusts the sampling rate of the positioning module based on detected travel patterns. When a pattern is recognized, the sampling rate is reduced from the maximum rate, thereby reducing energy consumption while maintaining sufficient classification accuracy through the use of previously collected high-rate data for pattern matching.
Solution Approach 2:
The system performs preliminary classification of travel patterns using collected position data before reducing sampling. This preliminary action allows the system to identify recognizable patterns and then reduce sampling rate accordingly, ensuring that classification accuracy is maintained for novel patterns while saving energy for repetitive patterns.
2Measurement precision
If the system continuously collects position data at high sampling rates to improve classification accuracy, then measurement precision is improved, but duration of action decreases
Solution Approach 1:
The system dynamically adjusts the sampling rate of the positioning module based on detected travel patterns. When a pattern is recognized, the sampling rate is reduced from the maximum rate, thereby extending battery life and device usage time while maintaining sufficient classification accuracy through the use of previously collected high-rate data for pattern matching.
3Measurement precision
If the system uses processor-intensive techniques to classify dwelling episodes to improve classification accuracy, then measurement precision is improved, but use of energy increases
Solution Approach 1:
The system segments the classification process into two stages: a first technique that uses less processor power for initial classification of dwelling episodes, and a second, more processor-intensive technique reserved for cases where the first technique is insufficient. This segmentation reduces overall processor power consumption while maintaining classification accuracy.
Solution Approach 2:
The system applies the more processor-intensive classification technique only partially - specifically, only when the first classification technique is insufficient or when confidence is low. For most dwelling episodes that can be classified by the simpler technique, the expensive processing is avoided, reducing overall energy consumption while maintaining accuracy where needed.
Data Source
AI summary
A mobile device includes a positioning module sampling at least a position of the mobile device at a sampling rate when active and a processor capable of determining whether a mobile device is travelling or dwelling based on at least the sampled position of the mobile device. The processor further identifies a travel mode for a trip segment for the mobile device based on at least the sampled position of the mobile device when the mobile device is travelling and identifies an activity when then the mobile device is dwelling.