Scholarly article on topic 'Energy Efficiency on Location Based Applications in Mobile Cloud Computing: A Survey'

Energy Efficiency on Location Based Applications in Mobile Cloud Computing: A Survey Academic research paper on "Electrical engineering, electronic engineering, information engineering"

Share paper
{Mobile / "Cloud computing" / "Location based service" / Energy-efficiency}

Abstract of research paper on Electrical engineering, electronic engineering, information engineering, author of scientific article — Xiao Ma, Yong Cui, Ivan Stojmenovic

Abstract With the emergence of mobile cloud computing (MCC), an increasingly number of applications and services becomes available on mobile devices. In the meantime, the constrained battery power of mobile devices makes a serious impact on user experience. As one increasingly prevalent type of applications in mobile cloud environments, location based applications (LBAs) present some inherent limitations surrounding energy. For example, the GPS (Global Positioning System) based positioning mechanism is well-known to be extremely power-hungry. Due to the severity of the issue, considerable researches have been devoted to energy-efficient locating sensing mechanism in the last few years. These efforts toward enhancing energy efficiency have allowed us to provide a comprehensive survey of recent work on low-power design of LBAs.

Academic research paper on topic "Energy Efficiency on Location Based Applications in Mobile Cloud Computing: A Survey"

Available online at

SciVerse ScienceDirect

Procedía Computer Science 10 (2012) 577 - 584

The 9th International Conference on Mobile Web Information Systems (MobiWIS)

Energy Efficiency on Location Based Applications in Mobile

Cloud Computing: A Survey

Xiao Maa, Yong Cuia*, Ivan Stojmenovicb

aDepartment of Computer Science, Tsinghua University, Beijing, P.R. China b School of Information Technology and Engineering, University of Ottawa, Ottawa, Canada


With the emergence of mobile cloud computing (MCC), an increasingly number of applications and services becomes available on mobile devices. In the meantime, the constrained battery power of mobile devices makes a serious impact on user experience. As one increasingly prevalent type of applications in mobile cloud environments, location based applications (LBAs) present some inherent limitations surrounding energy. For example, the GPS (Global Positioning System) based positioning mechanism is well-known to be extremely power-hungry. Due to the severity of the issue, considerable researches have been devoted to energy-efficient locating sensing mechanism in the last few years. These efforts toward enhancing energy efficiency have allowed us to provide a comprehensive survey of recent work on low-power design of LBAs.

© 2011 Published by Elsevier Ltd.

Keywords: Mobile, Cloud computing, Location based service, Energy-efficiency

1. Introduction

Mobile cloud computing is born to leverage powerful computing and storage resources in the cloud to provide abundant services in mobile environment conveniently and ubiquitously. The features of MCC include no up-front investment, lower operating cost, highly scalable and easy access, etc.. However, with the characteristics of user mobility and wireless access pattern, many obstacles such as mobility management, quality of service (QoS) guarantee, energy management, security and privacy issues are brought to MCC. The most critical one among them is energy efficiency issue of mobile devices. Since the battery manufacturing industry moves forward slowly (battery capacity grows by only 5% annually [1]), and the demand of computing and storage capability is rapidly increasing, how to provide better user experience with constrained battery power supply is becoming more urgent in recent years. Plenty of research has been proposed during the course of the last five years as shown in Figure 1.

As one of the most typical services in MCC, Location based services (LBS) which make use of the geographical position of mobile device, have the advantages of both user mobility and cloud resources in MCC.

* Corresponding author. Tel.: +86 10 627 85822; fax: +86 10 627 71138. Email address: (Yong Cui)

1877-0509 © 2012 Published by Elsevier Ltd. doi: 10.1016/j.procs.2012.06.074

Fig. 1: Publications with mobile and energy in title or abstract

These services gain user's current position by utilizing GPS, and provide various location-related services. However, experiments show that GPS only allows continuously working for 9 hours on smartphones, which indicates that saving energy costs in mobile locating sensing is a significant issue.

The locating technologies used today mainly include GPS, WiFi, and GSM. Each of these technologies can vary widely in energy consumption and localization accuracy. Experiments have shown that GPS is able to run continuously for 9 hours only, while WiFi and GSM can be sustained for 40 and 60 hours respectively. At the same time, the corresponding localization accuracies are about 10m, 40m, and 400m [2]. Recently, most LBAs prefer GPS for its accuracy although it is also perceived as extremely power-hungry. What is worse, phones currently only offer a black box interface to the GPS for the request of location estimates and the lack of sensor control makes energy consumption more inefficient [3]. Additionally, many LBAs requires continuous localization over reasonably long time scales. Therefore, energy-efficient locating sensing methods must be adopted to obtain accurate position information while expending minimal energy.

To the best of our knowledge, this is the first survey that takes locating sensing energy efficiency issue into account in mobile cloud environment in detail, while previous works in the literature usually mitigate the problem by performing different optimizations without considering the additional characteristics brought in by remote computing scheme and user behavior pattern. We believe our survey could provide a better understanding of the design challenges of energy-efficient locating sensing in MCC, and pave the way for further research in this area.

The rest of the paper is organized as follows. Section II provides a layered model of MCC and the related commercial products of LBAs. Section III presents several sensing based technologies, mainly including dynamic tracking and other alternative positioning technologies. Additional optimizations used before or after locating are proposed in section IV, including the management of multiple LBAs and the simplifications of trajectory tracking. Finally, to address future trends, we summarize and conclude the survey in Section V.

2. MCC business model and commercial products

Since MCC is proposed as an epitome of cloud computing, a layered model of cloud computing is provided as the business model of MCC. In MCC, the application layer is more appropriate for mobile users (such as navigation service based on position sensing), while the platform layer may provide distributed storage/database software framework allowing the mobility characteristics to fit in mobile environments. We categorize multiple cloud services into three service groups: software as a Service (SaaS), platform as a Service (PaaS) and Infrastructure as a Service (IaaS) shown in Figure 2.

Infrastructure as a service (laaS). This layer creates a pool of storage and computing resources by partitioning the available physical resources by using virtualization technologies. The related commercial products of this layer includes Amazon EC2 [4], GoGrid [5] and Flexiscale [6].

Platform as a service (PaaS). The platform layer mainly refers to the software or storage framework which aims to minimize the burden involved with deploying applications directly into VM containers. Examples of

End Users

Wireless network

Platform asase (PaaS)

Business Applications & web services

Application Laye

Computation (virtual machine)


CPU, memory, Disk, Bandwidth

Hardware Laye

Software framework (java/.NET) & Storage

Commercial products

Apple Appstore, RIM mobile service, LBAs, android market

Amazon SimpleDB/S3, Google AppEngine, Microsoft Azure

Amazon EC2, GoGrid

Fig. 2: Layered model of mobile cloud computing

PaaS include the Google AppEngine [7], Microsoft Azure [8] and Amazon S3 [9]. In addition, some studies involving code migration also propose several code offloading architectures aimed to reduce the burden on application programmers ([10], [11], etc.).

Software as a service (SaaS). The software level is on the top of the infrastructure layer, with the services on this level providing on-demand applications over the Internet. Presently, there are many representative business products, such as mobile services provided by RIM Blackberry, Apple AppStore, Google Android Market and Location Based Services.

LBA is one of the most typical applications on Saas layer. It gains user's current position and provide various user position related services (e.g. social network, health care, mobile commerce, transportation and entertainment). Besides, many of these LBAs need continuous position updates, such as Real Time Traffic, health care applications that visualize daily patterns, habits of patients [12] and My Experience [13]. Since the energy consumption involved with location sensing is extremely tremendous. Therefore, energy saving involved with mobile location sensing in MCC is a vital issue that cannot be ignored.

Several methods of energy-efficient locating sensing have been proposed in recent years, which have been proved to be useful (Table 1). Intuitively, leveraging large intervals between contiguous position updates may minimize the power consumption. The next challenge involves maintaining position accuracy, which is the motivation of the most general solution called dynamic tracking. As GPS is still frequently used by dynamic tracking, several novel solutions are explored by other works. These works are sensing-oriented and independent of applications, we call them sensing based technologies. Beyond the action of positioning, additional methods used before or after locating are also proposed to include the management of multiple LBAs and simplification of the trajectories for data transmission. We refer to this as application related optimizations. Therefore, we classify the exist approaches into two main categories: 1) sensing based technologies and 2) application related optimizations.

3. Sensing based technologies

3.1. Dynamic tracking

The basic idea of dynamic tracking is to attempt to minimize the frequency of needed position updates by only sampling positions (generally with GPS) when the estimated uncertainty in position exceeds the accuracy threshold.

Leonhardi et al. [21] first studied time-based and distance-based tracking about ten years ago. Since this time, several works - [22] [23] focusing on both energy efficiency and GPS positioning - formally proposed dynamic tracking techniques. Farrell et al. [22] take into account a constant positioning delay and target speed, while You et al. [23] take into account a constant positioning accuracy and delay, target speed and acceleration to detect if the target is moving or not. They assume that the parameters mentioned are constant,

Table 1: Energy-efficient locating sensing

Paper Target Sensors Scheme Simplify History

[14] EnT racked position tracking GPS .accelerometer Dynamic prediction with less power-intensive sensors NO NO

[2] EnLoc position tracking GPS. wifi, GSM. compass and accelerometer Dynamic selection among alternative location-sensing mechanisms. Dynamic prediction with less power-intensive sensors and historical data NO YES

[15] LBAs position tracking GPS.GSM and accelerometer Dynamic selection among alternative location-sensing mechanisms; Dynamic prediction with less power-intensive sensors; management of LBAs ; battery level considering NO NO

[16] CAPS Trajectory Tracking GSM with GPS cell-ID sequence matching with historical sequences NO Yes

[17] EnTrackedr Trajectory Tracking GPS. compass. and accelerometer Dynamic prediction with less power-intensive sensors. Trajectory simplification YES NO

[18] CT rack Trajectory Mapping GSM compass, accelerometer with GPS Building a database with GPS.GSM when training and mapping in database using GSM when working NO YES

[19] a - Loc Position tracking GPS. wifi .Bluetooth, and cell-tower Dynamically select among alternative location-sensing mechanisms with prediction NO YES

[20] RAPS position tracking GPS. accelerometer .Bluetooth, and cell-tower Dynamic prediction with less power-intensive sensors NO YES

deeming their methods to be inefficient and unreliable. Dynamic tracking is further developed in [14] [20] [17]. EnTracked [14], RAPS [20] and EnTrackedT [17] share a similar system structure, while differing in some technologic details. These three works represent the most typical instances of dynamic tracking, and will be discussed in detail below.

Now we will briefly introduce the general steps of dynamic tracking. This process obtains a GPS position and then uses a certain method to determine the user state (i.e. whether the device is moving or not). If the device is not moving, the logic waits for movement. When it is moving, the speed of the device is determined. Then a scheduling plan of sensors and radio is calculated with some principles included to minimize power consumption. When the estimated uncertainty in position exceeds the accuracy threshold, the process restarts and samples the next GPS position. In this process, the methods of movement detection, velocity estimation and scheduling principles can be very different.

EnTracked [14] uses an accelerometer alone to detect movement. It proposes an energy model to dynamically estimate parameters such as the delays and consumption, which can describe the power consumption of a real phone with a much higher precision. Speed is estimated using the speed and accuracy provided by the GPS module. The error limit (accuracy threshold) is previously given to EnTracked. Then, the point at which to power features (mainly GPS and radio) on and off is calculated from the parameters estimated above and the device model.

However, this method has several limitations. First, the accelerometer would not be able to power off when EnTracked is running. The power used by the accelerometer may be higher than occasionally wake it up for a simple position update and to calculate a new sleeping period in some scenarios. Second, the movement detection algorithm is not clever and accurate enough. Not only can the algorithm be misled by handset activity, but it is also deemed to be suitable only for pedestrians with a speed less than 10m/s.

EnTrackedr [17] extends EnTracked system in several aspects. It proposes the idea of trajectory tracking corresponding to position tracking in EnTracked. The former refers to a sequence of continuous po-

sitions, while the latter focuses on a current position. Error thresholds for position and trajectory tracking are illustrated respectively in Figure 3(a). Firstly, EnTrackedT adopts a Heading-Aware Strategy, which employs the compass as a turn point sensor and significantly reduces power consumption of trajectory tracking. EnTrackedT calculates the accumulated distance traveled orthogonal to the initial heading given by the compass, and compares this to the prescribed trajectory error threshold. We can see clearly that intervals between GPS usage can be much larger than in EnTracked, as seen in Figure 3(b). Secondly, EnTrackedT uses adaptive duty cycling strategies for the accelerometer and compass sensors, which make the system more efficient. Thirdly, EnTrackedT uses a speed threshold based strategy together with an accelerometer based strategy for movement detection. This strategy enables the system to handle different transportation modes e.g., walking, running, biking or commuting by a car. Fourthly, it explored algorithms of a simplified motion trajectory to reduce data size and communication costs caused by sending motion information.

Fig. 3: (a) Error thresholds for position and trajectory tracking (b) Heading deviations will increase the orthogonal distance beyond the threshold and force the GPS position to be updated

The error percentage of the EnTrackedT system is relatively high when the requested error threshold is small while the power consumption is much lower at the same time. Although EnTrackedT claims to have joint trajectory and position tracking, it seems to work better for trajectory based applications.

RAPS [20] is based on the observation that GPS is generally less accurate in urban areas. It introduces the concept of activity ratio, which is the fraction of time that the user is in motion between two position updates. It uses an accelerometer to detect movement while measuring the activity ratio at the same time. It then uses this activity ratio along with the history of velocity information to estimate the current velocity of the user. RAPS duty-cycles the accelerometer carefully, using a duty-cycling parameter deduced empirically. A significant portion of the energy savings of RAPS come from avoiding GPS activation when it is likely to be unavailable to use celltower-RSS (the received signal strength) blacklisting. It records the current celltower ID and RSS information and associates with the success or failure of GPS. Additionally, RAPS utilizes Bluetooth to share the newly updated position information to save more energy.

RAPS uses a combination of spatiotemporal location history, user activity, and celltower-RSS blacklisting and it also proposes sharing position readings among nearby devices (which is a different approach from the former two options). However, it has limitations as well. First, RAPS is mainly designed for pedestrians in urban areas. Second, the user space-time history and the celltower-RSS blacklist must be populated for RAPS to work efficiently. Third, its velocity estimation based on activity ratio can be misled by handset activity not related to human motion. Fourth, accelerometers on smartphones may need a onetime, per-device calibration of the offset and scaling before running RAPS. Moreover, context sharing using Bluetooth raises privacy and security concerns.

The three systems described above are all validated in real-world deployments and have been proved to be useful. Though they have a similar system structurally, they are quite different from each other. They all have advantages and limitations. Considering individual components, more works are related. For example, EEMSS [24] and LBAs [15] employ the idea of using low power sensors (i.e. accelerometer) to detect user state and context, while triggering activation of high power sensors (i.e. GPS) only if necessary. EnLoc [2] proposes a Simple Linear Predictor, and so on.

3.2. Alternative positioning technologies

GPS is still used frequently although at intervals as large as possible for dynamic tracking. As periodic or adaptive duty-cycling of GPS may not achieve significant energy savings under all conditions, several works have explored schemes which would rarely use GPS for positioning. These strategies are based on the spatio-temporal consistency in user mobility, or the large population statistics on routes in an area. These strategies are also integrated with GPS-assisted training.

CAPS [16] presents a Cell-ID Aided Positioning System based on the consistency of traveled routes and consistent cell-ID transition points. It stores the history of cell-ID and GPS position sequences, and then senses the cell-ID sequences to estimate the current position using a cell-ID sequence matching technique. According to the observation, for mobile users with consistent routes, the cell-ID transition point for each user can often uniquely represent the current user position.

CAPS consists of three core components - sequence learning, sequence matching and selection, and position estimation. CAPS opportunistically learns and builds the history of a user's route for future usage. Using a small memory footprint, CAPS maintains the user's past routes and triggers GPS, if necessary. For each cell-ID in a sequence, CAPS maintains a list of tuples following < position, timestamp >, where position represents a GPS reading, and timestamp is the time at which that reading was taken. It uses modified Smith-Waterman algorithm for cell-ID sequence matching.

CAPS is designed for highly mobile users who travel long distances in a predictable fashion. It will not work in some cases where GPS is not available such as indoors and the size of the historical database may be very large if the user travels much. Also, it is evaluated only in urban areas where cell-tower density is high. CAPS does not make use of the underlying geography in this paper.

EnLoc [2] also explored how to make use of the spatio-temporal consistency in user mobility. When exploiting habitual mobility, EnLoc uses the logical mobility tree (LMT) to record the person's actual mobility paths showed in Figure 4. The vertices of the LMT are also referred to as uncertainty points. The basic idea is to sample the activity at a few uncertainty points, and EnLoc predicts the rest.

Fig. 4: Personal mobility profile: (a) A spatial logical mobility tree (LMT) (b) A spatio-temporal LMT

The scheme mentioned above highly relies on, as well as limits to the spatio-temporal consistency in user mobility. It cannot handle users' deviation from habits. So EnLoc further exploits mobility of large populations as a potential indicator of the individual's mobility.

EnLoc hypothesizes that a "probability map" can be generated for a given area from the statistical behavior of large populations. Then an individual's mobility in that area can be predicted. For example, considering a person approaching a traffic intersection of street A: since the person has never visited this street, it is difficult to predict how he/she will behave at the imminent intersection. However, if most people are used to take a left turn to Street B, the person's movement can be inferred accordingly.

EnLoc is evaluated using traces collected from a UIUC campus, which is not representative of EnLoc's actual service territory. Additionally, it does not describe the detailed implementation. These two issues suggest room for improvement. However, we can still see the potential for the heuristic prediction in energy saving.

4. Application Related Optimization

Except for the action of positioning, the optimizations performed before or after locating are also considerable. These methods include the management for multiple LBAs and simplifications of the trajectories for data transmission to reduce communication overhead.

4.1. Multiple LBAs Management

As more than one location based application may run on a single smartphone at the same time, the asynchronous invokes of GPS from different LBAs unnecessarily lead to higher energy cost. LBAs [15] presents a design principle called Sensing Piggybacking (SP) to overcome this limitation. It proposes a middleware to manage muitiple LBAs to avoid unnecessary GPS invoking events.

Applications mainly request and register location sensing in two ways. The first one is One-time Registration, in which statically registers a location listener and periodically notifies the listener of location updates based on the specified parameters such as time interval and distance interval. The other type of registration is Multi-time Registration in which explicitly registers/unregisters GPS requests to enable hardware sleeping. LBAs focuses on Multi-time Registration, as mobile platforms such as Android have already employed mechanisms to synchronize the location sensing actions for One-time Registration scenarios.

SP listens to the sensing requests of LBAs and forces the incoming registration request to synchronize with existing location-sensing registrations. LBAs uses a triple (G1,T1,D1) to describe the location sensing requirement of the joining LBA, where G1 is the granularity of sensing (e.g.,fine (or GPS) and coarse (or Net)), T1 is the minimum time interval and D1 is the minimum distance interval for location updating. It uses (Gf, T2,D2) to denote the finest existing GPS registration, where T2 and D2 are the finest sensing intervals. Similarly, it uses (Gc, T3,D3) to denote the finest Net registration. The incoming triple is compared with the existing registration, and SP determines whether to register a new request or simply use the current one according to the granularity and interval requirements. It can re-use the existing sensing registrations thus eliminating some location-sensing invocations.

Since more than one LBA may be running on one smartphone at the same time, a middleware of multiple LBA management is essential for energy-efficient sensing. This middleware should be redesigned when incorporated with other energy-efficient mechanisms, just as SP is used with other principles in LBAs.

4.2. Trajectory Simplification

Trajectory simplification has been proposed as a means to reduce data size and communication costs caused by sending motion information. It is used for applications which need trajectory information instead of a single position.

The basic idea of trajectory simplification is to use a smaller subset of obtained positions, one which is minimal in size while still reflecting the overall motion information. In EnTracked, trajectory simplification is viewed as a special case of line simplification (which has been thoroughly discussed in the computational geometry community).

Based on the observation that most services will enforce a more verbose data format for sending trajectory data (e.g. for reasons of cross-platform utilization and web-service compliance), we can enforce a considerably higher amount of data to be sent per time stamped position which results in higher energy savings achievable by simplification.

The main consideration of trajectory simplification is the trade-off between computation cost and simplification. EnTrackedT designed several algorithms and made comparisons. The power consumptions of different algorithms are measured to choose the suitable one and it may be relevant to different applications or mobile systems.

5. Conclusion

Appealing to the requirement of energy savings, many approaches of energy-efficient locating sensing have been explored. Methods beyond the action of locating are somehow auxiliary, and most of the attentions are focused on locating sensing based methods. A class of lightweight positioning systems has been

developed to explore a large part of the energy-accuracy tradeoff space. These systems either reduce accuracy requirements, or aggressively use other cues to determine when and where to turn on GPS. Implicitly or explicitly, these systems generally make several assumptions about the environment or about user activity. We envision a day when smartphones will implement different lightweight systems, each suited to different environments and/or user activities, and selectively triggered under the appropriate circumstances.

In this paper, we present an in-depth survey of energy-efficient locating sensing technologies within the environment of MCC. We hope our work will provide a better understanding of design principles and challenges surrounding location based applications in MCC.


This work is supported by the National Core-High-Base Major Project of China 2010ZX01045-001-005-4, NSFC Project 61120106008,60911130511, partially supported by NSERC Canada Discovery grant.


[1] S. Robinson, Cellphone energy gap: Desperately seeking solutions, Strategy Analytics Tech. Rep.

[2] I. Constandache, S. Gaonkar, M. Sayler, R. Choudhury, L. Cox, Enloc: Energy-efficient localization for mobile phones, in: INFOCOM 2009, IEEE, IEEE, 2009, pp. 2716-2720.

[3] N. Lane, E. Miluzzo, H. Lu, D. Peebles, T. Choudhury, A. Campbell, A survey of mobile phone sensing, Communications Magazine, IEEE 48 (9) (2010) 140-150.

[4] Amazon elastic computing, (10 2011).

[5] Cloud hosting, cloud computing and hybrid infrastructure from gogrid, (11 2011).

[6] Flexiscale, (11 2011).

[7] Google app engine, (11 2011).

[8] Windows azure, (11 2011).

[9] Amazon s3, (102011).

[10] C. Young, Y. Lakshman, T. Szymanski, J. Reppy, D. Presotto, R. Pike, G. Narlikar, S. Mullender, E. Grosse, Protium, an infrastructure for partitioned applications, in: Hot Topics in Operating Systems, 2001. Proceedings of the Eighth Workshop on, IEEE, 2001, pp. 47-52.

[11] E. Cuervo, A. Balasubramanian, D. Cho, A. Wolman, S. Saroiu, R. Chandra, P. Bahl, Maui: making smartphones last longer with code offload, in: Proceedings of the 8th international conference on Mobile systems, applications, and services, ACM, 2010, pp. 49-62.

[12] J. Ryder, B. Longstaff, S. Reddy, D. Estrin, Ambulation: A tool for monitoring mobility patterns over time using mobile phones, in: Computational Science and Engineering, 2009. CSE'09. International Conference on, Vol. 4, IEEE, 2009, pp. 927-931.

[13] J. Froehlich, M. Chen, S. Consolvo, B. Harrison, J. Landay, Myexperience: a system for in situ tracing and capturing of user feedback on mobile phones, in: Proceedings of the 5th international conference on Mobile systems, applications and services, ACM, 2007, pp. 57-70.

[14] M. Kjœrgaard, J. Langdal, T. Godsk, T. Toftkjœr, Entracked: energy-efficient robust position tracking for mobile devices, in: Proceedings of the 7th international conference on Mobile systems, applications, and services, ACM, 2009, pp. 221-234.

[15] Z. Zhuang, K. Kim, J. Singh, Improving energy efficiency of location sensing on smartphones, in: Proceedings of the 8th international conference on Mobile systems, applications, and services, ACM, 2010, pp. 315-330.

[16] J. Paek, K. Kim, J. Singh, R. Govindan, Energy-efficient positioning for smartphones using cell-id sequence matching, in: Proceedings of the 9th international conference on Mobile systems, applications, and services, ACM, 2011, pp. 293-306.

[17] M. Kjœrgaard, S. Bhattacharya, H. Blunck, P. Nurmi, Energy-efficient trajectory tracking for mobile devices, in: Proceedings of the 9th international conference on Mobile systems, applications, and services, ACM, 2011, pp. 307-320.

[18] A. Thiagarajan, L. Ravindranath, H. Balakrishnan, S. Madden, L. Girod, et al., Accurate, low-energy trajectory mapping for mobile devices, Proc. of 8th USENIX Symosium on Networked Systems Design and Implementation (NSDI 2011), ACM Press.

[19] K. Lin, A. Kansal, D. Lymberopoulos, F. Zhao, Energy-accuracy trade-off for continuous mobile device location, in: Proceedings of the 8th international conference on Mobile systems, applications, and services, ACM, 2010, pp. 285-298.

[20] J. Paek, J. Kim, R. Govindan, Energy-efficient rate-adaptive gps-based positioning for smartphones, in: Proceedings of the 8th international conference on Mobile systems, applications, and services, ACM, 2010, pp. 299-314.

[21] A. Leonhardi, K. Rothermel, A comparison of protocols for updating location information, Cluster Computing 4 (4) (2001) 355-367.

[22] T. Farrell, R. Cheng, K. Rothermel, Energy-efficient monitoring of mobile objects with uncertainty-aware tolerances, in: Database Engineering and Applications Symposium, 2007. IDEAS 2007. 11th International, IEEE, 2007, pp. 129-140.

[23] C. You, P. Huang, H. Chu, Y. Chen, J. Chiang, S. Lau, Impact of sensor-enhanced mobility prediction on the design of energy-efficient localization, Ad Hoc Networks 6 (8) (2008) 1221-1237.

[24] Y. Wang, J. Lin, M. Annavaram, Q. Jacobson, J. Hong, B. Krishnamachari, N. Sadeh, A framework of energy efficient mobile sensing for automatic user state recognition, in: Proceedings of the 7th international conference on Mobile systems, applications, and services, ACM, 2009, pp. 179-192.