Dynamic Sensor Upsampling for Fitness Activity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile and wearable computing devices face challenges in accurately detecting user activities while managing power consumption, as using multiple sensors for activity identification increases power usage, and using fewer sensors may lead to inaccurate activity detection.
Innovation Solution
The computing device initially samples a limited set of sensors at a lower rate to detect activity changes, then upsamples a larger set of sensors at a higher rate when a fitness activity is detected, confirming the activity with increased data accuracy and reducing latency, and adjusts sampling rates based on activity recognition thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensor components are used to identify user activities, then activity detection accuracy is improved, but power consumption increases
Solution Approach 1:
The patent implements dynamic sampling rates where the system transitions between different operational states (normal, upsampling, active) based on detected activity patterns. The sensor sampling rate is adjusted from a lower first rate to a higher second rate when fitness activities are detected, allowing the system to maintain high accuracy when needed while conserving power during normal operations.
Solution Approach 2:
The system changes the sampling rate parameter dynamically based on the operational state. In normal state, sensors are sampled at a first rate; in upsampling state, the sampling rate increases to a second rate greater than the first rate. This parameter change allows the system to balance between power consumption and detection accuracy based on contextual needs.
2Use of energy by moving object
If a single or few sensor components are used to identify activities, then power consumption is reduced, but activity detection accuracy deteriorates
Solution Approach 1:
The system performs preliminary sampling of sensor data at a lower rate to detect potential fitness activities. When an activity initiation is detected, the system then transitions to upsampling mode to confirm the activity with higher sampling rates. This preliminary action allows the system to reduce power consumption during normal operations while maintaining the capability to accurately detect activities when they occur.
Solution Approach 2:
The system implements periodic sampling at different rates depending on the operational state. During normal operations, sensors are sampled periodically at a lower rate. When fitness activity is detected, the sampling frequency increases periodically to confirm the activity. This periodic action pattern allows the system to balance power consumption with detection accuracy.
3Speed
If sensor sampling rate is increased to detect activities faster, then detection speed is improved, but power consumption increases
Solution Approach 1:
The system dynamically adjusts the sensor sampling rate based on the operational state and detected activity patterns. The sampling rate transitions from a lower first rate during normal operations to a higher second rate when fitness activities are detected, enabling fast detection speed when needed while maintaining lower power consumption during normal operations.
Solution Approach 2:
The sampling rate parameter is changed dynamically based on system state. The system samples sensors at a first rate during normal state and increases to a second rate greater than the first rate when fitness activities are detected. This parameter change enables the system to achieve high detection speed only when necessary, thereby managing power consumption effectively.
Data Source
AI summary
In some examples, a method includes detecting, based at least in part on sampling a first set (122) of a plurality of sensors of a computing device (110) at a first rate, an indication that the user has initiated a fitness activity (112B), wherein the computing device (110) stores pre-defined identifiers of non-fitness and fitness activities in a set of pre-defined indications of activities; responsive to detecting the indication that the user has initiated the fitness activity, sampling, at a second rate that is greater than the first rate, a second set (120) of the plurality of sensors to determine a probability that the user is engaged in the fitness activity (112B); and responsive to determining that the probability satisfies a threshold, collecting, sensor data for the fitness activity using a particular set of the plurality of sensors that corresponds to a pre-defined identifier for the fitness activity (112B).