EasyTracker: An Android application for capturing
Contents
1. N Pelekis A Gkoulalas Divanis M Vodas D Kopanaki and Y Theodoridis Privacy Aware Querying over Sensitive Trajectory Data Proceedings of the ACM Conference on Information and Knowledge Management CIKM 2011 C Parent S Spaccapietra C Renso G Andrienko N Andrienko V Bogorny M L Damiani A Gkoulalas Divanis J Macedo N Pelekis Y Theodoridis and Z Yan Semantic Trajectories Modeling and Analysis ACM Computing Surveys to appear N Pelekis E Frentzos N Giatrakos and Y Theodoridis HERMES Aggregative LBS via a Trajectory DB Engine Proceedings of the ACM SIGMOD Conference 2008 C Panagiotakis N Pelekis I Kopanakis E Ramasso and Y Theodoridis Segmentation and Sampling of Moving Object Trajectories based on Representativeness JEEE Transactions on Knowledge and Data Engineering 2011 Jose Vasquez OruxMaps URL http www oruxmaps com accessed 1 Jun 2012 Runtastic URL http www runtastic com accessed 1 Jun 2012 Alienman Tech Wheres My Droid URL http wheresmydroid com accessed 1 Jun 2012 GPS Tracking Pro URL https play google com store apps details id com fsp android c accessed 1 Jun 2012 Android developers Location Manager URL http developer android com reference android location Locat ionManager html accessed 1 Jun 2012 N Meratnia and R A de By Spatiotemporal Compression Techniques for Moving Point Objects Proceedings of the Extending Dat2. yes the point will be saved Otherwise it is omitted This method is very sensitive to the threshold values chosen by the user To demonstrate this we recorded the test path twice by varying the threshold values For the first track it is 3 sec and 10 meters while for the second it is 10 sec and 80 meters Fig 5 illustrates the results It is obvious that the larger the values the less the coordinates stored which affects the quality of the stored track On the other hand if we choose small values to avoid this problem we store many perhaps redundant points The optimal values of course depend on the mode of movement walking running driving etc and may be difficult to be known in advance by the user L 4 Figure 5 Compressed tracks with Minimal time distance method Left Minimal Time 3 sec Minimal Distance 10 m Right Minimal Time 10 sec Minimal Distance 80 m 2 Da eN Figure 6 Recorded tracks with the BOPW algorithm Left ED 8 m Right ED 15 m VS Vb Figure 7 Recorded tracks with the Threshold method Left threshold 8 m Right threshold 15 m The two methods proposed in 17 Before Opening Window BOPW and Normal Opening Window NOPW achieve a very good compression rate The user gives as parameter a distance threshold Euclidean Distance ED which is a bound on how much the original path can deviate from the compressed one However a high ED value may lead to loss of data whi
3. saved track in a geographical file following one of many well known data formats like GPX KML and CSV GPS Uses 6 of 6 available sats gooni h a f 7 4 Z i ios Openstreethlap Seon b TT 00 02 37 g 1 Ofertas Figure 2 Screenshots of EasyTracker Left recording a new track Right displaying a recorded track 72 EasyTracker Location 1 7 984784 Figure 3 Privacy filter of EasyTracker Left Showing privacy area while recording a track Right Setting privacy regions and respective levels of privacy An important feature of EasyTracker is its ability to choose among various maps Google Maps Open Street Map and Microsoft Bing Maps also the possibility to use offline maps The latter is very important when the user is currently in an area that has no network or a bad connection Fig 2 illustrates two screenshots of the application Departing from the features that are more or less available in many other similar applications to protect users privacy the application allows registering user defined areas inside where recording is not permitted Such an area is defined by a point either set manually by providing exact coordinates or by touching on the map along with a desired radius around this point and a characteristic name e g home work mam s house The user can specify up to five privacy locations using a radius which is limited from 15 meter till 300 kilometers Fig 3 depicts screenshots
4. EasyTracker An Android application for capturing mobility behavior Alexandros Doulamis Dept of Informatics University of Piraeus Piraeus Greece alex doulamis gmail com Abstract This paper presents EasyTracker a mobile application developed for the Android O S that enable the storage analysis and map visualization of routes of mobile users Furthermore it enable users to manually annotate part of their routes with labels describing their activity and behavior e g home having breakfast travelling by car to work etc Of equal importance the application encapsulates several state of the art line simplification algorithms for compressing the trajectories drawn from collected GPS records as well as segmenting trajectories into homogeneous parts in order to facilitate automatic auditing of the user s manual annotation Keywords EasyTracker GPS data trajectory path track privacy compression segmentation Android OS I INTRODUCTION Nowadays the number of users possessing and using smartphones equipped by a GPS receiver increases rapidly while their interest for mobile applications that enable the storage analysis and visualization of the collected space time information is apparent There are already plenty of such general purpose applications including My Tracks 1 AndAndo 2 GPS Tracker 3 and EveryTrail 4 Apart from these there are a great number of applications d
5. abase Technology Conference EDBT 2004 M Potamias K Patroumpas and T Sellis Sampling Trajectory Streams with Spatiotemporal Criteria Proceedings of the Scientific and Statistical Database Management Conference SSDBM 2006 Android developers SQLite Database URL http developer android com reference android database sqlit e SQLiteDatabase html accessed 1 Jun 2012 Android developers Supporting Multiple Screens URL http developer android com guide practices screens_support html accessed 1 Jun 2012
6. ch are crucial to represent well the sketch of the movement so the tradeoff between compression and quality is still under question In Fig 6 we record the test path using the BOPW algorithm with different ED values In the first path ED 8 meters the tracker stored only a few geographical coordinates without losing any remarkable information On the contrary in the nght path ED 15 meters the path is actually destroyed since only two points start and end point are stored Different from the previous ones the Threshold Method 18 relies on the speed of the user to achieve good results the speed has to be constant In Fig 7 we recorded the sample track twice in the first the threshold was set to 8 meters while in the second it was 15 meters and in both cases the speed was constant at 30 km h As with the previous techniques we notice that also this cannot avoid a loss in representativeness when the threshold is set to a high value C Manual annotation vs automatic segmentation EasyTracker allows the user to annotate parts of the track with labels in order to describe her current activity e g stopped at Acropolis museum walking towards Parthenon 9 This way we allow the automatic annotation of user s activities including her stops and movements In detail the user has the option to select an annotation tag label which is attached to a specific part of the track or to type a smart text of her choice F
7. ch feature can turn out to be very useful for automatic fill in of surveys performed in transportation science or for researchers working on activity recognition who are usually restricted to use manually processed surveys Third EasyTracker encapsulates state of the art algorithms that process in an online fashion the received stream of GPS recordings and transforms it into meaningful tracks ready to be stored into a trajectory database 10 for further analysis Specifically EasyTracker includes line simplification methods that compress the incoming stream of timestamped locations thus reducing the storage cost which are then partitioned into homogeneous portions according to some spatio temporal criteria using a state of the art segmentation method 11 According to this segmentation the track is split into portions i e sub tracks which can be labeled by tags that describe the corresponding spatio temporal behavior of the user e g STOPPED when the speed is very low This is important as it facilitates the user or some auditing algorithm to compare her manual annotations with the classified sub tracks as provided by the segmentation algorithm The big picture that illustrates these novel features of EasyTracker is depicted in Fig 1 The rest of the paper is organized as follows Section II presents related work In section II the various features of EasyTracker are explained in more detail Subsequently we provide a genera
8. ernal or external sensors RunKeeper 5 Endomondo Sports Tracker 6 and runtastic 13 to name a few Third there exist applications designed to track lost devices or devices that belong to another person assuming permission is granted For instance tracking a smartphone is feasible nowadays with the assistance of GPS and wireless networks Applications of this type are for example Wheres My Droid 14 GPS Tracking Pro 15 and GPS Tracker 3 We classify EasyTracker in the first category since it does not focus to a specific target group or application We also argue that the supported functionality is by far richer than the functionality provided by current commercial apps IHI EASYTRACKER IN ACTION A Basic functionality enhanced with privacy filters Common to all similar applications the main functionality of EasyTracker is to record a track and visualize it in real time on a user selected map Thereby it stores the geographical coordinates on a local database and it is able to calculate useful statistics total distance travelled average speed etc Using the device orientation sensor the real heading can be calculated and displayed on a map and on a built in compass The user has also the options to take photos or set places point of interests POI and associate them to geographical coordinates create notations to each part of a track view a stored track on a map or export a
9. eveloped for specific groups of users A few examples include Runkeeper 5 Endomondo Sports Tracker 6 and Sports Tracker 7 that support functionality mainly focusing on sports fitness All of these have as main function to store and display a route on a map usually Google Maps also extracting simple statistics such as average speed distance travelled etc In this paper we present EasyTracker for Android O S The core of EasyTracker is threefold At first EasyTracker collects space time points of user s movement using a GPS receiver thus creating and visualizing a path or trajectory followed by the user In contrast to similar applications the collected path to be stored in a server s database is sanitized by filtering out locations that lie inside user defined areas of interest sensitive areas where the user does not allow to be tracked e g around home a hospital etc This way Nikos Pelekis Dept of Statistics amp Insurance Sci University of Piraeus Piraeus Greece npelekis unipi gr Yannis Theodoridis Dept of Informatics University of Piraeus Piraeus Greece ytheod unipi gr EasyTracker provides end users with personalized privacy functionality 8 The second key functionality is that EasyTracker allows a user to annotate parts of her track with labels and therefore to describe her current activity e g stopped at caf A driving towards office 9 Su
10. ferent devices and screen resolutions But it is exactly this that increases the degree of complexity for the software developer On the following two sub sections we discuss this challenge A Devices with different displays As already mentioned there are several devices that support Android Except the various kind of sensors which differs from device to device and the different hardware performances one of the biggest variances has to do with the display screen which is not only the size but also the screen density Therefore Android uses dpi dots per inch instead of pixels 20 Nevertheless the developer should make extra effort especialy if graphics are used Fig 11 BME 12 01 pm EasyTracker EasyTracker Figure 11 EasyTracker s main menu in different displays Left 800x480 high dpi Right 240x380 low dpi B Orientation depentent devices To make the work as user friendly as possible Android has decided to support different device orientations This means that the application adapts to the current orientation of the device so the user can choose the desired orientation depending on the circumstances Frequently this change takes effect automatically by the O S but sometimes the developer has to do it manually To solve this several layouts for each orientation mostly horizontal or vertical should be developed see e g Fig 12 EasyTracker DISPLAY TRACH EasyTracker DISPLAY TRACH Goo
11. gie maps wp oone T7 Witnoue mars v Googe maps w werowwans gt orange gt Figure 12 Display Track pre menu in different orientations Left vertical portrait right horizontal landscape VI CONCLUSION AND FUTURE WORK In this paper we presented EasyTracker a mobile application developed for the Android O S that enables the storage analysis and map visualization of routes of mobile users In comparison with related applications EasyTracker provides novel functionality at three levels 1 it enables users to manually annotate part of their routes with labels describing their activity and behavior 11 it encapsulates several state of the art trajectory compression algorithms for a tradeoff between storage cost and quality of movement s representation and 111 it automatically segments tracks according to a state of the art trajectory segmentation algorithm in order to facilitate automatic auditing of the user s manual annotation As a fourth preliminary contribution it enables users to protect their privacy by defining sensitive areas where recording is not permitted This type of applications can be used in several fields from route planning and resources administration e g carpooling to entertainment and social networking e g next generation location based social networks Challenges for future work include the extension of TSA algorithm in order to recognize for each track segment the activities
12. ig 8 EasvtTracker Type of Movement Car Bus Train On foot i4 Motorbike ol of Rirvrle Figure 8 Selecting movement mode An important feature of EasyTracker is the usage of the Trajectory Segmentation Algorithm TSA proposed in 11 TSA partitions tracks into homogeneous portions according to their speed pattern This segmentation splits the track into portions i e sub tracks where each of them can be labeled by a tag that describes the corresponding speed range e g WALKING assuming the speed pattern is in a predefined range This is important as it facilitates the user to compare her manual annotations with the classified sub tracks as given by the segmentation algorithm IV THE ARCHITECTURE OF EASYTRACKER As already mentioned the main features of EasyTracker include the recording and storage of a track and the visualization of a stored track on a map A Recording a track Only a few steps are required to record a track As it is illustrated in Fig 9 the application at first verifies the GPS signal If it is available the tracker starts storing the coordinates Whether a location is stored is the decision of the compression method enabled by the user Each time a set of coordinates are stored into the database the application uses the TSA algorithm to segment if necessary the track and calculates various useful statistics total distance of the track average speed etc Compress
13. ion Method save coordinates to Data Base check availability of GPS Signal not available Figure 9 Recording and storing coordinates B Storing track data We decided to store information in a local database in order to avoid requiring Internet connection For this purpose we make use of the Sqlite data base which is embedded on the Android SDK 19 The database is a simple schema that contains the main information of a track the GPS coordinates it derives from as well as the statistics described above Accompanying places POIs and pictures are also stored in the database EasyTracker EasyTracker ba g ya S gt S 2 OpfegstreetMap Figure 10 Visualizing information of a point left calculating distance of two points right C Visualizing and exporting a track EasyTracker has the function to display a recorded track on a map so that the user can analyze her stored track In particular the user has the option to study each stored point closer calculate the distance between two different geographical points get more information about a stored place or view a picture taken at this point Fig 10 It is also feasible to export a stored track to a file format appropriate for further processing e g the general purpose CSV KML for visualizing the track in Google Earth GPX for GPS use V A DEVELOPER S DISCUSSION One of the major advantages of Android O S is the compatibility to many dif
14. l overview of EasyTracker s architecture section IV and describe the challenges to develop for different devices section V Section VI concludes and proposes improvements for future work Il RELATED WORK The number of existing applications that make use of the information captured by a GPS receiver is significant To obtain a better overview we present the most successful applications by classifying them in different categories In the general purpose category which targets to people who simply like to track and visualize their routes Google s MyTracks 1 is the most famous application for Android O S We have to note that this application is a relatively simple tracker with no advanced features however it has more than 5 million downloads Other similar applications in the same category include AndAndo 2 EveryTrail 4 which has more advanced features by allowing users to add pictures or points of interest and create guides and OruxMaps 12 which allows utilizing various kinds of maps incoming GPS 27x feed 1 I home l sw privacy filter compressed GPS movement mode On the Car compatibility check store on DB Figure 1 The big picture of EasyTracker Another category is about applications focusing on sports fitness Here we find many specialized features like the computation of calories consumption or the measurement of heart rate utilizing int
15. of the user in such a way as to enable the automatic validation and auditing of the user s annotation an energy saving policy since our application is used in devices with limited energy resources a possibility to detect user s position in non GPS available areas e g in indoors environments or in subsurface regions without loosing in accuracy and a hybrid local in the cloud storage schema that would make it ready for social networking applications ACKNOWLEDGMENT This work was partially supported by the European FP7 Open FET project DATASIM Data Science for Simulating the Era of Electric Vehicles URL http www datasim fp7 eu 3 4 5 7 8 9 10 11 12 13 14 15 16 17 18 19 20 REFERENCES Google MyTrack URL http www google com mobile mytracks accessed 1 Jun 2012 Javi Pacheco AndAndo URL https play google com store apps details id com javielinux andando accessed 1 Jun 2012 InstaMapper LLC GPS Tracker URL http www instamapper com accessed 1 Jun 2012 GlobalMotion Media Inc EveryTrail URL http www everytrail com accessed 1 Jun 2012 FitnessKeeper Inc RunKeeper URL http runkeeper com accessed 1 Jun 2012 Endomondo Sports Tracker URL http www endomondo com accessed 1 Jun 2012 Sports Tracking Technology Ltd Sports Tracker URL http www sports tracker com accessed 1 Jun 2012
16. of this privacy filter B Compressing incoming GPS feed One of the differences between conventional GPS Trackers and EasyTracker is the exploitation of several state of art algorithms which are responsible to decide which GPS recordings to be saved in the stored trajectory so as to minimize the mobile device s storage cost EasyTracker incorporates four spatiotemporal compression techniques for this purpose algorithms to save a track to a local data base Minimal Time Minimal Distance Method which is a component part of the Android SDK 16 Before Opening Window BOPW 17 Normal Opening Window NOPW 17 and Threshold Algorithm 18 The detailed presentation of the above methods is omitted due to space limitations To demonstrate the results of compression achieved by the above methods we use the path illustrated in Fig 4 Figure 4 A test path for the comparison of compression techniques The Minimal Time Minimal Distance Method 16 1s the simplest method to save coordinates The user sets two parameters a value for the minimal time and a value for the minimal distance The first value determines when to check the second value If the second value is exceeded then the incoming coordinate will be saved For example if minimal time is set to 5 sec and minimal distance is set to 12 meters the application checks every 5 sec whether the user is 12 meters away from the previously stored point If
Download Pdf Manuals
Related Search