[Dataset] wisc/airshark (v. 2011-12-14)

top

the initial version

@MISC{wisc-airshark-2011-12-14,
  author = {Shravan Rayanchu and Ashish Patro and Suman Banerjee},
  title = {{CRAWDAD} data set wisc/airshark (v. 2011-12-14)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/wisc/airshark},
  month = dec,  
  year = 2011
}
					

This is a dataset of RF device usage measurements collected at University of 
Wisconsin-Madison for demonstration of functionality of Airshark.

We collected the RF device usage measurements using a signal analyzer at 21 
locations. 

We broadly categorize these locations into three categories: 

(i) cafes (L1-L7): these included coffee shops, malls, book-stores,
(ii) enterprises (L8-L14): offices, university departments, libraries, and 
(iii) homes (L15-L21): these included apartments and independent houses.

We use AirMaestro RF signal analyzer to determine the ground truth about the 
prevalence of RF devices.

We collected the RF device usage measurements using the signal analyzer at 21
locations for a total of 640 hours.

At some locations, we could collect data for 24 hours (e.g., enterprises, 
homes), but for others we could collect measurements only during the day times
for a few hours (e.g., coffee shops, malls).

The data contains no personally identifiable information. Thus it has not
been sanitized.

Before using AirMaestro to understand the ground truth about the prevalence of
non-WiFi devices, we benchmarked its performance in terms of (i) device 
detection accuracy and (ii) false positives. The few cases where AirMaestro 
failed to detect the devices occurred when the devices were operating at very 
low signal strengths (less than -90 dBm).

[Traceset] wisc/airshark/rf (v. 2011-12-14)

top

the initial version.

@MISC{wisc-airshark-rf-2011-12-14,
  author = {Shravan Rayanchu and Ashish Patro and Suman Banerjee},
  title = {{CRAWDAD} trace set wisc/airshark/rf (v. 2011-12-14)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/wisc/airshark/rf},
  month = dec,  
  year = 2011
}
					

This is a traceset of RF device usage measurements collected at University of 
Wisconsin-Madison for demonstration of functionality of Airshark.

We use AirMaestro RF signal analyzer to determine the ground truth about the 
prevalence of RF devices. This device uses a specialized hardware (BSP2500RF 
signal analyzer IC), which generates spectral samples (FFTs) at a very high 
resolution (every 6 microseconds, with a resolution bandwidth of 156 kHz) and 
performs signal processing to detect and classify RF interferers accurately.

The data contains no personally identifiable information. Thus it has not
been sanitized.

Before using AirMaestro to understand the ground truth about the prevalence of
non-WiFi devices, we benchmarked its performance in terms of (i) device 
detection accuracy and (ii) false positives. The few cases where AirMaestro 
failed to detect the devices occurred when the devices were operating at very 
low signal strengths (less than -90 dBm).

[Trace] wisc/airshark/rf/2011 (v. 2011-12-14)

top

the initial version

@MISC{wisc-airshark-rf-2011-2011-12-14,
  author = {Shravan Rayanchu and Ashish Patro and Suman Banerjee},
  title = {{CRAWDAD} trace wisc/airshark/rf/2011 (v. 2011-12-14)}, 
  howpublished = {Downloaded from http://crawdad.cs.dartmouth.edu/wisc/airshark/rf/2011},
  month = dec,  
  year = 2011
}
					

This is a trace of RF device usage measurements collected in 2011 at 
University of Wisconsin-Madison for demonstration of functionality of Airshark.

The AirMaestro Embedded Signal Analyzer scans all IEEE 802.11 channels in the
2.4 GHz and 5 GHz Wi-Fi frequency bands. It generates spectral samples every
6 microseconds, with a resolution bandwidth of 156 kHz.

Measurement data corresponding to each location is in a separate file (21 files in total), and are labeled as <environment>_Location<index>.csv : here <environment> is Cafe/Enterprise/Home and <index> goes from 1 to 21. The data is the form of comma separated values. Each row consists of 29 fields. Below, we explain the fields and their description.

Field 1: RecordId
A unique value that identifies this interference event. It corresponds to a row id.

Field 2: Ignore

Field 3: Interferer Type
Represents a class of non-WiFi device (e.g., Bluetooth, Analog Cordless Phone etc.)

Field 4: Interferer Start Time
The time the interference started. The time is the number of seconds since the UNIX time epoch, i.e. seconds since midnight Jan 1, 1970 UTC. We use the standard C/C++ library functions time() to get the value and strftime() to convert the value into human readable time and date. Please refer to time.h for function prototypes.

Field 5: Interferer Stop Time
The time the interference stopped. The time is number of seconds since the Unix time epoch.

Field 6: Ignore
Field 7: Ignore
Field 8: Ignore

Field 9: Min RSSI
The minimum Receive Signal Strength Indicator (RSSI) value sampled for the interference event.

Field 10: Max RSSI
The maximum RSSI value sampled for the interference event.

Field 11: Avg RSSI
The average of all RSSI samples taken for the interference event.

Field 12: Center frequency
Center frequency of the interference event. The frequency value is in units of KHz.

Field 13: Number of Impacted WiFi channels
DThe number of WiFi channels impacted by the interference from this device.

Fields 14 - 29: Impacted channels
Sixteen fields that list the WiFi channel numbers impacted by the interference. Field 13 indicates how many of the fields are actually have channel values. For example, when Field 13=4, then Fields 14, 15, 16, and 17 contain the channel values impacted by this device. Unused fields are set to zero.

The data contains no personally identifiable information. Thus it has not
been sanitized.

Before using AirMaestro to understand the ground truth about the prevalence of
non-WiFi devices, we benchmarked its performance in terms of (i) device 
detection accuracy and (ii) false positives. The few cases where AirMaestro 
failed to detect the devices occurred when the devices were operating at very 
low signal strengths (less than -90 dBm).

[Author] Shravan Rayanchu

top

[Author] Ashish Patro

top

[Author] Suman Banerjee

top

[Paper] rayanchu-airshark

top