IEEE Communications Magazine - June 2017 - page 122

IEEE Communications Magazine • June 2017
0163-6804/17/$25.00 © 2017 IEEE
Piggyback crowdsensing (PCS) is a novel ener-
gy-efficient mobile crowdsensing paradigm that
reduces the energy consumption of crowdsensing
tasks by leveraging smartphone app opportuni-
ties (SAOs). This article, based on several funda-
mental assumptions of
incentive payment
for PCS
task participation and
spatial-temporal coverage
assessment for collected sensor data, first pro-
poses two alternating data collection goals. Goal
1 is
maximizing overall spatial-temporal coverage
under a predefined incentive budget constraint
goal 2 is
minimizing total incentive payment while
ensuring predefined spatial-temporal coverage for
collected sensor data
, all on top of the PCS task
model. With all of the above assumptions, set-
tings, and models, we introduce
— a
generic incentive allocation framework for the
two optimal data collection goals, on top of the
PCS model. We evaluated CrowdMind extensive-
ly using a large-scale real-world SAO dataset for
the two incentive allocation problems. The results
demonstrate that compared to baseline algo-
rithms, CrowdMind achieves better spatial-tem-
poral coverage under the same incentive budget
constraint, while costing less in total incentive
payments and ensuring the same spatial-tempo-
ral coverage, under various coverage/incentive
settings. Further, a short theoretical analysis is
presented to analyze the performance of Crowd-
Mind in terms of the optimization with total incen-
tive cost and overall spatial-temporal coverage
With the rapid proliferation of sensor-equipped
smartphones, mobile crowdsensing (MCS) [1]
has become an efficient way to sense and collect
environmental data of urban areas in real time
(e.g., air quality, temperature, and noise level).
Instead of deploying static and expensive sensor
networks in urban areas, MCS leverages the sen-
sors embedded in mobile phones and the mobil-
ity of mobile users to sense their surroundings,
and utilizes the existing communication infrastruc-
ture (3G, WiFi, etc.) to collect data from mobile
phones scattered in an urban area. By collecting
sensor readings from mobile users, a “big pic-
ture” of the environment in the target area can be
obtained using MCS without significant cost.
Our earlier work [2] demonstrated that there
are two main players in MCS: the
, who
is the person or organization coordinating the
sensing task, and the
who are the
mobile users involved in the sensing task. An MCS
task usually requires the organizer to recruit par-
ticipants with incentive payment, to allocate sens-
ing tasks to selected participants, and to collect
sensor readings from these participants’ mobile
devices that will represent the characteristics of
the target sensing region, often with a predefined
budget for participant incentives.
Specifically, the MCS organizer needs to spec-
ify the target sensing area, which often consists of
a set of subareas, and further specify the sensing
duration (e.g., one week), which is usually divided
into equal-length sensing cycles (e.g., each cycle
lasts for an hour). Once the settings of subareas
and sensing cycles are determined, the MCS appli-
cation usually needs to collect a number of sensor
readings from each subarea of the target region
in each sensing cycle. Taking a one-week urban
air quality monitoring MCS task as an example,
the MCS organizer first divides the whole area
into 1 km
grid cells, then splits the one-week
MCS process into a sequence of one-hour sensing
cycles [3], and further requests at least one MCS
participant in each grid to upload the air quali-
ty sensor reading in each sensing cycle. Besides
full spatial-temporal coverage
[4], the orga-
nizer frequently uses the
partial spatial-temporal
metrics for MCS data collection, where
the fraction of subareas being covered by at least
one sensor reading in each sensing cycle is used
to represent the coverage [5]. Usually, the use
of full spatial-temporal coverage is to ensure the
collected sensor readings representing each sub-
area in each sensing cycle, while the use of partial
coverage aims to collect data that could represent
the majority part (e.g., 80 percent) of subareas in
each cycle.
In addition to organizers’ efforts in the pro-
cess of participant recruitment, task assignment,
and data collection, MCS also requires the par-
ticipants’ mobile devices to sustainably perform
sensing tasks and upload sensor data during the
MCS process. In order to prolong the battery
life of mobile devices engaged in MCS, various
solutions have been proposed to reduce energy
consumption of individual mobile devices, rang-
ing from adapting sensing frequency to inferring
part of data rather than sensing and uploading
all data [6]. One of the effective solutions is
Near-Optimal Incentive Allocation for
Piggyback Crowdsensing
Haoyi Xiong, Daqing Zhang, Zhishan Guo, Guanling Chen, and Laura E. Barnes
The authors propose two
alternating data collection
goals. Goal 1 is maximiz-
ing overall spatial-tem-
poral coverage under a
predefined incentive bud-
get constraint; goal 2 is
minimizing total incentive
payment while ensuring
predefined spatial-tempo-
ral coverage for collected
sensor data, all on top of
the PCS task model.
Haoyi Xiong and Zhishan Guo are with Missouri University of Science and Technology; Daqing Zhang is with Peking University;
Guanling Chen is with the University of Massachusetts Lowell;; Laura E. Barnes is with the University of Virginia.
Digital Object Identifier:
1...,112,113,114,115,116,117,118,119,120,121 123,124,125,126,127,128,129,130,131,132,...228
Powered by FlippingBook