Companion Testing for Body-Aware Device Validation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The testing of body-aware software applications, such as those for smart watches, is complicated by the need to integrate human interaction data, making it challenging to specify and generate test cases and differentiate between human input errors and software bugs, as human variability affects sensor data input.
Innovation Solution
A method and apparatus that utilize movement data classification using a processor to predict wearable device output, comparing it with the output of a body-aware application, and providing an indication of any discrepancies, while training a three-dimensional model to identify specific actions and forming an API for a classifier to evaluate the performance of body-aware applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If human interaction data is integrated into body-aware software testing, then the testing can evaluate real-world performance, but the testing process becomes complicated and difficult to automate
Solution Approach 1:
A companion application serves as an intermediary between the human user and the testing system. This companion app captures movement data, classifies actions, and generates expected outputs, thereby mediating the complexity of human interaction and providing structured test data without requiring direct human involvement in the testing process
Solution Approach 2:
The system creates a virtual copy of human interaction through the companion application that records and classifies movements. This copy allows the testing system to reproduce and analyze human behavior patterns without actual human participants, enabling automated testing while maintaining testing accuracy
2Adaptability or versatility
If human variability is accounted for in sensor data, then the testing can reflect real-world conditions, but it becomes difficult to differentiate between human input errors and software bugs
Solution Approach 1:
The companion application provides feedback by classifying human movements and generating expected outputs that are compared against the body-aware application's outputs. This feedback mechanism enables the system to distinguish between variations caused by human variability and actual software errors, while maintaining adaptability to different user behaviors
Solution Approach 2:
The testing process is segmented into separate components: the companion application handles human interaction and data collection, while the main testing system handles comparison and error detection. This segmentation allows human variability to be captured separately from software performance evaluation, making error detection clearer
3Productivity
If automated testing is implemented for body-aware applications, then testing efficiency improves, but the ability to handle human interaction nuances is reduced
Solution Approach 1:
The companion application performs self-service by automatically capturing, classifying, and processing human movement data. It autonomously generates expected outputs without requiring manual intervention, thereby enabling automated testing while maintaining the ability to handle complex human interaction nuances through its intelligent classification algorithms
Data Source
AI summary
One embodiment provides a method, including: receiving movement data describing physical movement of a person performing a predetermined action; generating, using a processor, classification of the movement data using a test application that predicts output of a wearable device, wherein the test application has been formed using previously collected data that describe the movement of a person performing the predetermined action; determining, using the processor, whether the movement data match the predetermined action in view of the classification; receiving output of a body-aware application that detects and responds to human movement; comparing, using the processor, the output of the body-aware application with the classification; and providing, using the processor, an indication of the comparing of the output of the body-aware application and the classification.