Ubiquitous Computing (UbiComp) is a computational paradigm that enhances the use of computing devices by making them available to the user anywhere and anytime. From the energy perspective, it is often very important to compute the entire UbiComp task within a specific deadline with minimum energy. The literature on determining the energy consumption of the system for computing the task does not consider periodic tasks and different sampling rate of the sensors, which eliminates the deadline constraints in the analysis. Since the period of the tasks is not fixed, the estimated delay without considering the fixed period is lower than the actual value. In this paper, we assume that an Edge, Fog, and Cloud layers based UbiComp system computes the periodic task within the specific deadline. We derive the expressions of total delay and energy consumption of the UbiComp system. Using the derived expressions, we estimate fractions of the task that are computed at each layer to reduce the energy consumption such that the task is computed within a specific deadline. Our numerical and prototype results demonstrate the impact of the data size, network topologies, deadline, and characteristics of the sensors on the energy consumption, delay, and accuracy of the system.

中文翻译：

---

具有时限约束的雾基泛在系统中的节能数据转发方案

无处不在的计算 (UbiComp) 是一种计算范式，它通过让用户随时随地可用来增强计算设备的使用。从能量的角度来看，在特定期限内以最小能量计算整个 UbiComp 任务通常非常重要。确定用于计算任务的系统能耗的文献没有考虑周期性任务和传感器的不同采样率，从而消除了分析中的期限约束。由于任务周期不固定，不考虑固定周期的估计延迟低于实际值。在本文中，我们假设基于边缘、雾和云层的 UbiComp 系统计算特定期限内的周期性任务。我们推导出 UbiComp 系统的总延迟和能量消耗的表达式。使用派生的表达式，我们估计在每一层计算的任务的分数，以减少能源消耗，以便在特定的期限内计算任务。我们的数值和原型结果证明了数据大小、网络拓扑、截止时间和传感器特性对系统能耗、延迟和准确性的影响。