Location History in a Low-cost Context Awareness Environment
Teddy Mantoro and Chris Johnson
Department of Computer Science, Australian National University and Smart Internet Technology Cooperative Research Centre Building !", North Road, Australian National University ACT # !$!!, Australia The proiblem This paper e%plores ho& usefull is the location a&areness history for an office 'ased lo&(cost conte%t(a&areness environment field establishment Conte%t(a&areness solution )e develop a conte%t agent prototype using ASR called Speech Conte%t Agent *SpeechCA+ in the distri'uted support architecture for I,- )e use the cross platform .ava Speech A/I *.SA/I+ that implements speech synthesis and voice recognition- The SpeechCA needs a grammar to control the recognition process 'y telling the spea0er &hat &ords they are e%pected to say and the patterns in &hich these &ords may occur- The advantage of a grammar file is that it ma0es the recognition faster and more accurate contribution 1ur goal is to develop interfaces that implicitly can understand &hat the user &ants to do in the office lo&(cost conte%t environment, &hich is only possi'le if a conte%t a&are application e%ists for the devices- It should 'e 'ased on the user2s current conte%t and relevant history of location, activity, identity and time- The conte%t a&are application could 'e designed to use the incoming conte%t information services to use the incoming conte%t services to encode some action- evaluation In this paper &e sho&ed that lo&(cost conte%t(a&areness can 'e implemented using e%isting devices and 'y providing an improved interface 'et&een user and devices- )e gave an overvie& of availa'le technology that can 'e used for ac3uiring information from the user and the environment- User Location and Mobility for Distributed Intelligent Environment Teddy Mantoro Department of Computer Science, Australian National University, ACT(!$!!, Australia 45($(5$6 7"8" teddy-mantoro9anu-edu-au The proiblem )e have e%plored precise, pro%imate and predicted users2 location using variety of sensors *i-e- )i:i and Bluetooth+ and investigate ho& the sensors fit in an A1 to have interopera'ility to detect users- The pro'lem is to com'ine these 0no&n location data to determine a user2s actual location &ith sufficient precision for office activity purposes- field establishment Conte%t(a&areness solution )e used the Bluetooth access point as a sensor for several rooms &ithin the range- :or e%ample, &hen a user is close to a certain access point, his location &ill 'e pro%imately close to the access point and it could represent user location from several rooms- )e used self organi;ing map *<ohonen map+ approach of artificial neural net&or0 to cluster the signal strength data- 1nce &e get the signal strength cluster allocation in the local I,, &e can directly get current user location- In our e%periment using )i:i, &e used access points to measure signal strength in t&o ad=acent 'uildings, already installed for )>AN access- contribution 1ur e%periment using )i:i and Bluetooth as pro%imate location in an A1 has sho&ed their good results in sensing user location- The result can 'e improved 'y developing interopera'ility 'et&een sensors to get aggregate sensor data- evaluation Aggregate of precise location has first priority and is follo&ed 'y pro%imate and predicted locations respectively- This means that &hen the A1 receives information from aggregate precise location, then the current user2s location is determined- If not, then &e chec0 using aggregate pro%imate location data- ser Mobility Model in an Active !ffice Teddy ?antoro, C-)- .ohnson Department of Computer Science, Australian National University, North Road, ACT !$!!, Australia {teddy,cwj}@cs.anu.edu.au The proiblem This paper descri'es user mo'ility 'y detecting their changing locations- Unfortunately, computers and communications systems today are underused 'ecause the range of control mechanisms and application interfaces are too diverse- Conte%t a&areness mechanisms could 'e the 'est &ay to implement computer applications in I, field establishment The Intelligent ,nvironment *I,+ solution )e developed a model to predict and pro%imate user location using &ireless sensors in the ?erino layering architecture, i-e- the architecture for scala'le conte%t processing in an Intelligent ,nvironment- contribution This paper proposed an implementation model of an Active 1ffice environment &hich is simple, efficient, scala'le, fault tolerant and applica'le to a range of heterogeneity computing platforms- This paper has descri'ed the re3uirement of recogni;ing user locations and activities in the ?erino service layer architecture, i-e- the architecture for scala'le conte%t processing in a u'i3uitous computing environment and am'ient intelligent *em'edded sensing devices+ environment- evaluation 1ur e%periment using )i:i and Bluetooth in determining pro%imate location in an Active 1ffice has sho&n good results in sensing user location- At present, )i:i rather than Bluetooth seems to 'e the 'etter &ay to fi% precise location using pro%imate sensors - The result can 'e improved 'y developing interopera'ility 'et&een sensors to yield aggregate sensor data