BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Talk1: EmotionSense: A Mobile Phones based Adaptive Platform for E
 xperimental Social Psychology Research\, Talk2: Energy-Accuracy Trade-offs
  in Querying Sensor Data for Continuous Sensing Mobile Systems - Kiran Rac
 huri (University of Cambridge)
DTSTART:20100921T123000Z
DTEND:20100921T133000Z
UID:TALK26058@talks.cam.ac.uk
CONTACT:Eiko Yoneki
DESCRIPTION:Talk1-Abstract: Today's mobile phones represent a rich and pow
 erful computing platform\, given their sensing\, processing and communicat
 ion capabilities. Phones are also part of the everyday life of billions of
  people\, and therefore represent an exceptionally suitable tool  for cond
 ucting social and psychological experiments in an unobtrusive way. \nIn th
 is paper we illustrate EmotionSense\, a mobile sensing platform for social
  psychology studies based on mobile phones. Key characteristics include th
 e ability of sensing individual emotions as well as activities\, verbal an
 d proximity interactions among members of social groups. Moreover\, the sy
 stem is programmable by means of a declarative language that can be used t
 o express adaptive rules to improve power saving. We evaluate a system pro
 totype on Nokia Symbian phones by means of several small-scale experiments
  aimed at testing performance in terms of accuracy and power consumption. 
 Finally\, we present the results of real deployment where we study partici
 pants emotions and interactions. We cross-validate our measurements with t
 he results obtained through questionnaires filled by the users\, and the r
 esults presented in social psychological studies using traditional methods
 . In particular\, we show how speakers and participants' emotions can be a
 utomatically detected  by means of classifiers running locally on off-the-
 shelf mobile phones\, and how speaking and interactions can be correlated 
 with activity and location measures.\n\nTalk1 is a dry run for UbiComp 201
 0.\n\nTalk 2 - Abstract: A large number of context-inference applications 
 run on off-the-shelf smart-phones and infer context from the data acquired
  by means of the sensors embedded in these devices. The use of efficient a
 nd effective sampling technique is of key importance for these application
 s. Aggressive sampling can ensure a more fine-grained and accurate reconst
 ruction of context information but\, at the same time\, continuous queryin
 g of sensor data might lead to rapid battery depletion. \nIn this paper\, 
 we present an adaptive sensor sampling methodology which relies on dynamic
  selection of sampling functions depending on history of context events. W
 e also report on the experimental evaluation of a set of functions that co
 ntrol the rate at which the data are sensed from the Bluetooth device\, ac
 celerometer\, and microphone sensors and we show \n that a dynamic adaptat
 ion mechanism provides a better trade-offs compared to simpler function ba
 sed rate control methods. Furthermore\, we show that the suitability of th
 ese mechanisms varies for each of the sensors\, and the accuracy and energ
 y consumption values stabilize after reaching a certain level.\n\nTalk2 is
  a dry run for Workshop on Mobile context awareness workshop at UbiComp 20
 10.\n
LOCATION:FW11\, Computer Laboratory\, William Gates Builiding
END:VEVENT
END:VCALENDAR