CRAWDAD dartmouth/cenceme

CenceMe is a sensing system based on standard and sensor-enabled mobile phones. CenceMe uses the output of the phones' sensors and external data (if such is available) to infer human presence and activity information. This dataset contains movements and inferred activities of participants using CenceMe on their mobile phones.

The CenceMeLite traces were collected from 2008-07-28 to 2008-08-11 by students and staff members at Dartmouth College.

last modified 2010-08-30

reason for most recent change the initial version

release date 2008-08-13

date/time of measurement start: 2008-07-28

date/time of measurement end: 2008-08-11

collection environment: CenceMe is a personal sensing system, which uses sensor data gathered using mobile devices (e.g. sensor-enabled cell phones) to learn about the activities of their carriers.

network configuration: The dataset was collected during the deployment of a modified version of the CenceMe application, CenceMeLite, that logged all the sensed information and high-level inferred activities on the phone's on-board flash memory.

data collection methodology: The phones recorded information about the system and raw data from accelerometer and GPS devices.

limitation: Some users did not use the phone much and thus did not collect useful data.

CenceMeLite traces collected by students and staff members at Dartmouth College.

The data were collected by means of 20 Nokia N95 phones carried by students and staff members from the departments of Computer Science and Biology at Dartmouth College.

network type: sensor network

methodology: The dataset includes the following information for each user: accelerometer raw data and GPS location coordinates.

format of trace data:  In the files there are both information lines (about system configuration) and data lines. Information lines are written when the system is started.

All lines (except for the preamble written at the start of the system) start with a timestamp followed either by the keyword INFO (for information lines) or DATA (for data lines).

The preamble written when the system starts consists of the following three lines:

------------- NEXT LOG ------------------

Data lines can be distinguished with the keywords: ACC, ACT, GPS.

ACC lines contain accelerometer raw data in the format: Timestamp DATA (0) - ACC: Xacc,Yacc,ZaccXacc,Yacc,Zacc... where: - Timestamp is the time when the line has been written into the log file. - Xacc, Yacc, Zacc are respectively the accelerations on the three axes. Every accelerometer sample is separated by the symbol *.

ACT lines contain information about inferred activities in the following format: Timestamp DATA (0) - ACT: AccSamplingStart,AccSamplingEnd, Fact, where: - Timestamp is the time when the line has been written into the log file. - AccSamplingStart is the time the accelerometer starts the sampling. - AccSamplingEnd is the time the accelerometer starts the sampling. The codes of the different facts ("Fact" field) are the following:

Sitting: 0 Running: 1 Walking: 2 Standing: 5

"AccSamplingStart" and "AccSamplingEnd" are the start and end times of the interval during which the accelerometer data used for classification of that particular activity were collected.

GPS lines are of 3 types: - No samples - N samples - the string "GPS-Skipped: user sitting".

------------------------------------