Grab Detection Calibration Using Sensor Derivatives
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile devices face challenges in accurately detecting user intent due to variations in hand size, orientation, temperature, and humidity, which affect the baseline for sensor data, making it difficult to differentiate between meaningful transitions like grabbing the device.
Innovation Solution
The use of multiple touch sensors to capture user actions and treat the detection space as the derivative of raw sensor data, combined with machine learning techniques like SVM models, to classify transitions as 'grab' or 'not-a-grab', and correlating with data from motion, proximity, pressure, and temperature sensors to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple touch sensors are used to capture user actions, then detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent combines data from multiple sensor types (touch sensors, motion sensors, proximity sensors, pressure sensors, temperature sensors) into a unified detection system. By merging these diverse sensor inputs and processing them together through a common framework, the system achieves improved detection accuracy while avoiding the complexity of treating each sensor independently.
Solution Approach 2:
The detection system is designed to handle multiple sensor types and multiple user actions through a single universal framework. The system can detect various actions (grabbing, holding, releasing) using the same processing logic applied to different sensor data streams, reducing overall system complexity while maintaining high detection accuracy.
2Reliability
If sensor data baseline is adjusted for environmental variations, then detection reliability is improved, but processing complexity increases
Solution Approach 1:
The system performs preliminary calibration by collecting sensor data during a calibration phase before actual use. During this phase, the system captures baseline sensor readings under various environmental conditions (different temperatures, humidities, hand orientations) and uses this pre-collected data to establish reference values. This preliminary action eliminates the need for complex real-time adjustments during actual detection, improving reliability while keeping processing simple.
Solution Approach 2:
The system adjusts detection parameters based on environmental conditions by comparing current sensor readings against calibrated baseline values collected under different conditions. Instead of implementing complex real-time environmental compensation, the system changes its detection thresholds and parameters based on pre-determined calibration data that accounts for temperature, humidity, and orientation variations.
3Speed
If real-time sensor data analysis is performed to detect user intent, then responsiveness is improved, but energy consumption increases
Solution Approach 1:
The system performs partial analysis by focusing only on the most relevant sensor data for detecting grab actions. Rather than continuously analyzing all sensor inputs at full processing depth, the system applies simplified detection logic for common actions like grabbing, using pre-established calibration data to make quick determinations without exhaustive real-time computation, thus reducing energy consumption while maintaining responsiveness.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In one embodiment, a method includes receiving real-time sensor data from N sensors on the computing device. The real-time sensor data corresponds to a transition in a physical state of the computing device caused by a user of the computing device. The method also includes applying a linear function to the real-time sensor data from each of the N sensors; determining a vector based on an N-tuple comprising the derivatives; comparing the vector with a pre-determined hyperplane with N-1 dimensions; and determining based on the comparison whether the transition is an event corresponding to any of one or more pre-determined imminent uses of the computing device by the user or a non-event not corresponding to any of the pre-determined imminent uses of the computing device by the user.