Data-Driven Activity Prediction With Sensor Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing activity prediction techniques in smart environments face challenges due to noisy sensor data and errors in activity labeling, which affect the accuracy of predicting future activity times, especially when considering spatial and temporal relationships.
Innovation Solution
The implementation of data-driven activity prediction techniques that utilize activity-labeled sensor events to learn an individual's routine, employing both independent and recurrent activity predictors to generate predictions for future activity occurrence times, and integrating these predictions into an activity prompting application to facilitate facility automation and user intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing activity prediction techniques are used, then activity prediction can be implemented, but prediction accuracy deteriorates due to noisy sensor data and labeling errors
Solution Approach 1:
The system implements feedback loops where predicted activities are compared with actual sensor data and user corrections. The activity prediction module continuously refines its predictions based on feedback from activity recognition results and user corrections, improving accuracy over time despite noisy input data.
Solution Approach 2:
The patent introduces an activity recognition module as an intermediary between raw sensor data and activity predictions. This intermediary processes and labels sensor events before they reach the prediction module, filtering out some noise and providing structured input that improves prediction reliability.
2Ease of operation
If manual input of user schedules is required, then activity prompting can be implemented, but system complexity and user burden increase
Solution Approach 1:
The system automatically learns user activity patterns and schedules from sensor data without requiring manual input. The activity prediction module autonomously generates prompting schedules by analyzing historical sensor events and recognized activities, eliminating the need for users to manually configure their daily routines.
Solution Approach 2:
The system performs preliminary analysis of sensor data to pre-establish activity patterns and schedules before prompting is needed. By continuously learning from incoming sensor events, the system prepares prediction models in advance, so when prompting is required, accurate predictions are already available without requiring user configuration.
Data Source
AI summary
A physical environment is equipped with a plurality of sensors (e.g., motion sensors). As individuals perform various activities within the physical environment, sensor readings are received from one or more of the sensors. Based on the sensor readings, activities being performed by the individuals are recognized and the sensor data is labeled based on the recognized activities. Future activity occurrences are predicted based on the labeled sensor data. Activity prompts may be generated and/or facility automation may be performed for one or more future activity occurrences.