A Computing Framework to Check Real-Time Requirements in Ambient Intelligent Systems.

Ambient Intelligent (AmI) systems are related to environments that usually demand real-time characteristics in their context-aware applications. These systems can require a timing response according to specific needs dealing with context changes and user interactions. The current work presents a computing framework able to check and measure those times required by ubiquitous systems, running AmI processes on mobile or networked devices. The proposed framework focuses on checking real-time requirements of applications which can monitor and recognize human activities. This framework is then applied to two examples of well-known datasets related to Human Activity Recognition (HAR) and it supports the implementation of a mobile application which allows researchers to measure those latency times involved in HAR processes. The computation cost of training and prediction processes for these datasets is measured using such application and the obtained results highlight the critical effect of smartphone performance when processing sensor-based data and, specially, during the elaboration of online training models in this context.