Introduction
Background
Technology | State/Action | Power (mW) |
---|---|---|
Wi-Fi | In connection | 868 |
Idle | 58 | |
3G | Downloading@1Mbps | 1400 |
2G | Downloading@44 kbps | 500 |
Bluetooth | On | 15 |
Connected and idle | 67 | |
Receiving/sending | 432 | |
CPU usage | 50% | 462 |
75% | 561 | |
100% | 612 | |
Display | Black light background - 60% | 98.65 |
White light background - 60% | 254.65 | |
Black light background - 100% | 259.65 | |
White light background - 100% | 527.05 |
Review
Operating systems
Software solutions
Energy-efficient wireless interfaces
Energy-efficient sensors
Energy-efficient computation off-loading
Battery management mobile applications
Hardware solutions
Performance evaluation
Energy management technique | Implementation | Evaluated YES/NO | Evaluation setup/ methodology | Energy consumption | |
---|---|---|---|---|---|
Operating Systems
| |||||
Cinder (Roy et al. 2011) | Abstractions are implemented in the Cinder kernel, which runs on AMD64, i386, SPARC, and ARM architectures. It is freely available on Internet | Yes | HTC Dream (Google G1), based on the Qualcomm MSM7201A chipset | 12.5% total system power reduction over the 20min experiment | |
ErdOS (Vallina-Rodriguez 2011) | Prototype is implemented as an extension of the Android OS | No | N/A | Authors claim that it improves the battery capacity of smart phones by managing resources proactively | |
CondOS (Zhao 2011) | N/A | No | N/A | Authors claim that it provides several opportunities for energy reduction, such as shared dataflow processing, dataflow-to-hardware mapping, and principled flow degradation | |
CONTEXTO (Schirmer and Bertel 2014) | Prototype is currently available as an iOS framework | Yes | Apple iPhone 4, 4S, and 5 | Provides energy-awareness to developers of context-aware applications | |
AURA (Pasricha et al. 2015) | Prototype is implemented as a middleware on two android based smart phones | Yes | HTC Dream and Nexus One | Can achieve up to 29% energy savings as compared to the baseline device manager & it’s 5 times more energy efficient then previous approaches | |
Software Solutions
| |||||
Energy-efficient Wireless Interfaces | Catnap (Dogar and Steenkiste 2010) | Prototyped in C for the Linux environment | Yes | Nokia N810 and IBM Thinkpad T60 both supporting 802.11 PSM | Allows the NIC to sleep for around 40% of the time for a 10MB transfer while 70% of the time for a 5MB transfer. Improves battery capacity up to 2-5x for real devices like Thinkpad T60 & Nokia N810 |
NAPman (Rozner et al., 2010) | Prototyped using the MadWifi v0.9.4 driver for Atheros-based WiFi cards on the Linux platform | Yes | HP iPAQ hw6945, iPhone 3GS, gPhone HTC Magic and HTC Tilt 8900 | Under varied settings of background traffic, it improves the energy savings on a smartphone by up to 70% while ensuring fairness | |
Bartendr (Schulman et al. 2010) | N/A | Yes | 4 cellular networks across 2 metropolitan areas, one in US & the other in India, and spans 3G networks based on both EVDO & HSDPA | Significant energy savings of up to 10% for email sync and up to 60% for on-demand streaming | |
SALSA (Ra et al. 2010) | Implemented SALSA algorithm in Urban Tomography system which runs on the Nokia N95 smartphone, having 802.11b/g WiFi interface, 3G/EDGE, a 2GB micro-SD card, & supports 640x480-resolution video recording capability | Yes | Nokia N95 and Android G1 | Closer to an empirically determined optimal than any other alternatives compared with it, and, can save 10-40% of battery for some workloads | |
PhoneJoule (Liu et al. 2013) | Prototype is implemented using java and Eclipse integrated with Android SDK & ADT. It can work on all smartphones which support Android OS 2.2 or later versions | Yes | ZTE v880 smartphone which supports Android OS 2.2 and SEMO to measure power consumption | Very effective for energy saving in smartphones and makes it very convenient for users to manage battery usage of their smartphones | |
PerES (Cui et al., 2013) | Implemented as a traffic management application by utilizing IPTABLES (a system tool in Android) | Yes | Google Nexus S and Monsoon Power Monitor device | Better than peer schemes, TailEnder & SALSA. Using 821 million traffic flows collected from commercial cellular carrier, it can achieve on average 32% to 56% energy savings with different levels of user experience | |
Energy-efficient Sensors | A-loc (Lin 2010) | Prototype is implemented on an Android G1 phone | Yes | Android G1 and AT&T Tilt phones, on paths that include indoor and outdoor locations, using war driving data from Microsoft & Google | Saves significant amount of energy and also improves the accuracy |
Adaptive location-sensing framework (Zhuang et al., 2010) | Design principles are implemented as a middleware on G1 Android Phone with OS version 1.5 Cupcake, by modifying the Application Framework | Yes | G1 Android Developer Phone (ADP) | Minimize the usage of the energy-consuming GPS up to 98% and improve battery life by up to 75% | |
RAPS (Paek et al., 2010) | Prototype is implemented in Symbian C++ for the Symbian S60 3rd FP1 devices | Yes | Nokia N95-3 smartphone, with GPS, accelerometer, Bluetooth, WiFi and 3G/EDGE interfaces, & 2GB micro-SD card | Can increase phone battery by more than a factor of 3.8 as compared to the approach where GPS is always on | |
Bayesian Networks (Yi and Cho 2012) | Proposed context-aware system for GPS has prototyped as an application in Android platform | Yes | LG SU-660 with Android OS 2.2 version | Active person and inactive person can save energy of about 5% and 3% per hour, respectively | |
Jigsaw (Lu et al. 2010) | Proposed continues sensing engine has implemented on two smartphone platforms, Nokia N95 & Apple iPhone, as background service and library, respectively | Yes | Nokia N95 and Apple iPhone | Authors claim that Jigsaw is capable of performing long-term energy efficient GPS tracking without sacrificing the accuracy. However, the paper lacks clear performance evaluation results | |
WheelLoc (Wang et al. 2013) | Implemented as a continuous background system service on NexusOne phones running Android 2.3 | Yes | NexusOne phones with Android OS 2.3 version | Can return a location estimate within 40ms with an accuracy about 40 meters, consumes only 240mW energy, & effectively strikes a better energy-accuracy tradeoff than GPS duty-cycling | |
Energy-efficient Computation-offloading | MAUI (Cuervo and Balasubramanian 2010) | Prototype is implemented on HTC Fuze smartphone running Windows Mobile 6.5 with .Net Compact Framework v3.5 | Yes | HTC Fuze smartphone running Windows Mobile 6.5 with .Net Compact Framework v3.5 and for MAUI server, dual-core desktop with 3GHZ & 4GB RAM running Win 7 | For 4 applications running on Windows Mobile phones, it can achieve energy conservation of up to one order of magnitude |
Cuckoo (Kemp et al., 2012) | Integrates with the popular open source Android framework and the Eclipse development tool | Yes | 2 real world apps that contain heavy weight computation, eyeDentify and PhotoShoot | With little effort computation off-loading can be enabled for object recognition and gaming app, using the Cuckoo framework | |
Synergy (Kharb et al., Kharbanda et al. 2012) | Prototype implementation is developed for the Android operating system | Yes | 2 compute intensive apps – image smoothing & video processing running on Google Nexus S with Android OS 2.2 version and PowerTutor | Can save up to 30.6% of the system battery with less than 5% latency penalty | |
MADNet (Ding et al., 2013) | Prototype is implemented on Nokia N900 smartphones | Yes | Nokia N900, Nokia E71 and Samsung Nexus S. Energy Profiler application and Monsoon Power Monitor | Can achieve more than 80% energy saving | |
Self learning off-loading scheme (Arora 2014) | N/A | No | N/A | Authors claim that enabling the off-loading system to self learn makes it more reliable, fast and energy efficient | |
Battery Management Mobile Applications | Android and iOS platform | Yes | Android and iOS | It allows users to quickly look up battery status as well as track down what applications are draining battery life. It also helps to charge the device healthily with 3 Stage Charging system and it can extend the battery life up to 50% | |
Android | Yes | Android and AV-TEST | Multifunctional application which works like a cleaner and as an antivirus at the same time. It can boosts mobile applications by almost 32% and it also protects the device from unwanted malware & spyware | ||
Hardware solutions
| |||||
Little Rock (Priyantha and Lymberopoulos 2011) | Integrated Little Rock into an actual prototyping phone | Yes | Pedometer app while running on the phone, on Little Rock as well as on a hybrid architecture that includes the phone with an embedded Little Rock board | For a pedometer application, the energy savings by running with Little Rock is three orders of magnitude compared to the normal approaches | |
MobileHub (Haichen et al., 2014) | Prototyped with a sensor hub comprised of an 8-bit AVR micro-controller attached to sensors, and by extending the Android OS to use this sensor hub | Yes | Galaxy Nexus phones with Android OS 4.2.2 version | For three applications downloaded from the Android marketplace, it can improve power consumption by up to 83%, with no effort from the developer |
Summary and findings
Operating systems | • An energy-efficient operating system must be user-centric and context-aware in order to anticipate future battery limitations. |
• To prioritize the tasks at runtime operating system should employ self organization based energy management techniques. | |
Wireless interfaces | • Modern networks can be modeled and simulated in a better way by treating them as artificial Complex Adaptive Systems (CAS), or generalizing as Complex Adaptive COmmunicatiOn Networks and environmentS (CACOONS). • Energy-efficient communication via wireless interfaces optimization usually requires some collaborative mechanism between operating system, applications, and network infrastructure. |
Sensors | • Energy-efficient continuous sensing has become a hot concern for both researchers and developers of mobile applications. |
• Several schemes have been proposed for energy-efficient sensing, which combine different sensors to minimize the energy consumption as well as the error. | |
• Most of the techniques are based on probabilistic models of user’s history and patterns to deduce future locations. | |
Computation off-loading | • Extensive research is being carried out on using cloud services for energy efficiency in mobile devices. However, it depends on different factors (e.g. the amount of data exchanged, the network state, etc.) which can affect the efficiency of computation off-loading. |
• Self organization based collaborative mechanism for resource sharing with co-located devices using low-energy wireless connectivity has shown to be very effective for energy saving and improving the user experience. | |
Battery management mobile applications | • Recently, increased energy-awareness at application level has also received interest and is likely to gain more attention in the future. |
Hardware solutions | • Some researchers have suggested that re-designing the mobile device’s hardware architecture is beneficial for energy-efficient sensing. |
• Experimental results revealed that the idea of additional low energy micro-controller or core for managing sensors is actually very effective in terms of saving energy. |