Machine learning (ML) methods has recently contributed very well in the advancement of the prediction models used for energy consumption. Such models highly improve the accuracy, robustness, and precision and the generalization ability of the conventional time series forecasting tools. This article reviews the state of the art of machine learning models used in the general application of energy consumption. Through a novel search and taxonomy the most relevant literature in the field are classified according to the ML modeling technique, energy type, perdition type, and the application area. A comprehensive review of the literature identifies the major ML methods, their application and a discussion on the evaluation of their effectiveness in energy consumption prediction. This paper further makes a conclusion on the trend and the effectiveness of the ML models. As the result, this research reports an outstanding rise in the accuracy and an ever increasing performance of the prediction technologies using the novel hybrid and ensemble prediction models.

/pdf/energy-consumption-prediction-using-machine-learning-a-37fdysr0tu.pdf

Energy consumption prediction using machine learning: A review

Smartphones and smartphone apps have undergone an explosive growth in the past decade. However, smartphone battery technology hasn't been able to keep pace with the rapid growth of the capacity and the functionality of smartphones and apps. As a result, battery has always been a bottleneck of a user's daily experience of smartphones. An accurate estimation of the remaining battery life could tremendously help the user to schedule their activities and use their smartphones more efficiently. Existing studies on battery life prediction have been primitive due to the lack of real-world smartphone usage data at scale. This paper presents a novel method that uses the state-of-the-art machine learning models for battery life prediction, based on comprehensive and real-time usage traces collected from smartphones. The proposed method is the first that identifies and addresses the severe data missing problem in this context, using a principled statistical metric called the concordance index. The method is evaluated using a dataset collected from 51 users for 21 months, which covers comprehensive and fine-grained smartphone usage traces including system status, sensor indicators, system events, and app status. We find that the remaining battery life of a smartphone can be accurately predicted based on how the user uses the device at the real-time, in the current session, and in history. The machine learning models successfully identify predictive features for battery life and their applicable scenarios.

Predicting Smartphone Battery Life based on Comprehensive and Real-time Usage Data

Audio signal processing is the most challenging field in the current era for an analysis of an audio signal. Audio signal classification (ASC) comprises of generating appropriate features from a sound and utilizing these features to distinguish the class the sound is most likely to fit. Based on the application’s classification domain, the characteristics extraction and classification/clustering algorithms used may be quite diverse. The paper provides the survey of the state-of art for understanding ASC’s general research scope, including different types of audio; representation of audio like acoustic, spectrogram; audio feature extraction techniques like physical, perceptual, static, dynamic; audio pattern matching approaches like pattern matching, acoustic phonetic, artificial intelligence; classification, and clustering techniques. The aim of this state-of-art paper is to produce a summary and guidelines for using the broadly used methods, to identify the challenges as well as future research directions of acoustic signal processing.

Pattern analysis based acoustic signal processing: a survey of the state-of-art

Today, a wide range of developments and paradigms require the use of embedded systems characterized by restrictions on their computing capacity, consumption, cost, and network connection. The evolution of the Internet of Things (IoT) towards Industrial IoT (IIoT) or the Internet of Multimedia Things (IoMT), its impact within the 4.0 industry, the evolution of cloud computing towards edge or fog computing, also called near-sensor computing, or the increase in the use of embedded vision, are current examples of this trend. One of the most common methods of reducing energy consumption is the use of processor frequency scaling, based on a particular policy. The algorithms to define this policy are intended to obtain good responses to the workloads that occur in smarthphones. There has been no study that allows a correct definition of these algorithms for workloads such as those expected in the above scenarios. This paper presents a method to determine the operating parameters of the dynamic governor algorithm called Interactive, which offers significant improvements in power consumption, without reducing the performance of the application. These improvements depend on the load that the system has to support, so the results are evaluated against three different loads, from higher to lower, showing improvements ranging from 62% to 26%.

Smart Sensor Architectures for Multimedia Sensing in IoMT.

Heterogeneous multi-processor system-on-chip (MPSoC) smartphones are required to offer increasing performance and user quality-of-experience (QoE), despite comparatively slow advances in battery technology. Approaches to balance instantaneous power consumption, performance and QoE have been reported, but little research has considered how to perform longer-term budgeting of resources across a complete battery discharge cycle. Approaches that have considered this are oblivious to the daily variability in the user’s desired charging time-of-day (plug-in time), resulting in a failure to meet the user’s battery life expectations, or else an unnecessarily over-constrained QoE. This paper proposes QUAREM, an adaptive resource management approach in mobile MPSoC platforms that maximises QoE while meeting battery life expectations. The proposed approach utilises a model that learns and then predicts the dynamics of the energy usage pattern and plug-in times. Unlike state-of-the-art approaches, we maximise the QoE through the adaptive balancing of the battery life and the quality of service (QoS) for the duration of the battery discharge. Our model achieves a good degree of accuracy with a mean absolute percentage error of 3.47% and 2.48% for the energy demand and plug-in times, respectively. Experimental evaluation on an off-the-shelf commercial smartphone shows that QUAREM achieves the expected battery life of the user within 20–25% energy demand variation with little or no QoE degradation.

/pdf/quarem-maximising-qoe-through-adaptive-resource-management-rwtm5jl1.pdf

QUAREM: Maximising QoE Through Adaptive Resource Management in Mobile MPSoC Platforms

Optimizing energy consumption in modern mobile handheld devices plays a very important role as lowering energy consumption impacts battery life and system reliability. With next-generation smartphones and tablets, the number of sensors and communication tools will increase and more and more communication interfaces and protocols such as Wi-Fi, Bluetooth, GPRS, UMTS, and LTE will be incorporated. Consequently, the fraction of energy consumed by these components will be larger. Nevertheless, the use of the large amount of data from the different sensors can be beneficial to detect the changing user context, to understand habits, and to detect running application needs. All these information, when used properly, may lead to an efficient energy consumption control. This paper proposes a tool to analyze user/application interaction to understand how the different hardware components are used at run-time and optimize them. The idea here is to use machine learning methods to identify and classify user behaviors and habit information. Using this tool, a software has been developed to control at run-time system component activities that have high impacts on the energy consumption. The tool allows also to predict future applications usages. By this way, screen brightness, CPU frequency, Wi-Fi connectivity, and playback sound level can be optimized while meeting the applications and the user requirements. Our experimental results show that the proposed solution can lower the energy consumption by up to 30 % versus the out-of-the-box power governor, while maintaining a negligible system overhead.

/pdf/sensing-user-context-and-habits-for-run-time-energy-3f2qf2emuo.pdf

Sensing user context and habits for run-time energy optimization

https://hal.archives-ouvertes.fr/hal-01879172/document

ENOrMOUS: ENergy Optimization for MObile plateform using User needS

The success of mobile devices as measured by market penetration is undeniable. Tablets and Ultrabooks are enjoying similar growing enthusiasm among the worldwide population. By offering an ever increasing number of functionalities, these devices turned to be even more subject to energy constraints than the previous generation. Limited battery autonomy becomes a major user concern and cause of dissatisfactions. Therefore, managing efficiently the energy consumption can improve mobile systems reliability, battery life as well as user experience. In this paper we propose a new approach to optimize mobile devices energy efficiency based on use patterns detection. By identifying and classifying users behaviors, we can significantly improve over the platforms stock power managers. To do so, a run-time service is proposed to collect usage data, which in turn can be mined using machine learning techniques. Our apprach allows us to predict future applications usages, so the CPU frequency, Wi-Fi connectivity and the playback sound-levels can be optimized while meeting the applications and the users requirements. Our experimental results show that the proposed solution can lower the energy consumption by up to 20% vs. the out-of-the-box power governor, while maintaining a negligible system overhead.

Application Sequence Prediction for Energy Consumption Reduction in Mobile Systems

Optimizing energy consumption in modern mobile handheld devices plays a crucial role as lowering power consumption impacts system's autonomy and system reliability. Recent mobile platforms have an increasing number of sensors and processing components. Added to the increasing popularity of power-hungry applications, battery life in mobile devices is an important issue. However, we think that the utilization of the large amount of data from the various sensors can be beneficial to detect the changing device context, the user needs or the running application requirements in terms of resources. When these information are used properly, an efficient control of power knobs can be implemented thus reducing the energy consumption. This paper presents URBOC, for User Request Based Optimization Component. This component is an extension of our previous framework [7] ENOrMOUS. This framework was able to identify and classify the user contexts in order to understand user habits, preferences and needs which allow to improve the operating system power scheme. In this paper, we extend the use of ENOrMOUS by allowing the user to send requests in order to extend the battery life up to a specific time. Machine Learning (ML) and data mining algorithms have been used to obtain an efficient trade-off between power consumption reduction opportunities and user requests satisfaction. The proposed solution increases battery life by up to seven hours depending on user requests vs. the out-of-the-box operating system power manager with a negligible overhead.

Machine learning for improving mobile user satisfaction

The optimization of the energy consumption in mobile devices can be performed on hardware and software components For example, reducing the screen brightness or switching off GPS The energy control must take into account both the current context and user habits, on the base of usage knowledge acquired from sensors and OS data records The whole process of energy management then includes data collection, usage learning and analysis, decision-making and control of device components To integrate these activities, we propose to use a software agent whose goal is to save the energy of the mobile device with the lowest effect on QoS

Ismat Chaib Draa

Papers

Sensing user context and habits for run-time energy optimization

ENOrMOUS: ENergy Optimization for MObile plateform using User needS

Application Sequence Prediction for Energy Consumption Reduction in Mobile Systems

Machine learning for improving mobile user satisfaction

An Energy-Aware Learning Agent for Power Management in Mobile Devices