Predominant Activity Detection for Context-Specific Map Suggestions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital map applications on portable devices fail to provide relevant and context-specific suggestions based on the device's activity, such as mode of transport, leading to incongruent search results and user experience.
Innovation Solution
A software system that determines the predominant activity of a portable device over a period of time using sensor data and confidence metrics, then generates and provides relevant geospatial data and suggestions tailored to that activity, such as walking, driving, or biking, by analyzing activity detection results and confidence scores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the digital map application provides general suggestions without considering device activity, then the system complexity is reduced, but the relevance and context-specificity of suggestions deteriorates
Solution Approach 1:
The system performs preliminary activity detection and classification before generating suggestions. Sensor data is collected and analyzed in advance to determine the device's activity state (walking, driving, stationary), which then informs the generation of context-relevant suggestions. This preliminary action ensures suggestions are tailored to the actual usage context without adding complex real-time processing during suggestion generation.
2Loss of information
If the application uses sensor data to determine device activity, then the context-specificity of suggestions is improved, but the device complexity increases
Solution Approach 1:
The system uses a multi-functional approach where the same sensor data processing infrastructure serves multiple purposes: activity detection, context determination, and suggestion generation. The activity detection module is designed to handle various sensor inputs (accelerometer, gyroscope, GPS) and produce a unified activity state that can be applied across different suggestion scenarios, reducing overall system complexity while maintaining context-specificity.
Solution Approach 2:
The system changes parameters by transforming raw sensor data into standardized activity states (walking, driving, stationary). This parameter transformation simplifies the complexity by converting multiple sensor readings into a few discrete activity categories that are easier to process and use for generating context-relevant suggestions without requiring complex real-time analysis.
3Measurement precision
If the application analyzes activity detection results over time intervals, then the accuracy of predominant activity determination is improved, but the processing time increases
Solution Approach 1:
The system implements periodic action by analyzing activity detection results at regular time intervals rather than continuously. It collects sensor data over defined time windows, determines the predominant activity for each interval, and uses these periodic assessments to generate suggestions. This approach balances accuracy (by analyzing sufficient data over time) with efficiency (by not requiring continuous real-time processing).
Solution Approach 2:
The system applies partial action by focusing analysis only on the most recent time intervals relevant to current usage context, rather than processing the entire history of sensor data. It selectively analyzes a subset of time intervals that are most indicative of current device activity, reducing processing time while maintaining sufficient accuracy for generating relevant suggestions.
Data Source
AI summary
Relevant geospatial data is provided to a portable device. In particular, several activity detection results collected over a period of time, spanning several time intervals, are received, where each activity detection result provides an estimate of the portable device activity, and where the activity detection results indicate at least two different activities. A single predominant activity of the portable device is determined using the activity detection results. Based on the predominant activity, geospatial data is selected and then provided to the portable device.