Accelerometer-Based Transportation Mode
Detection on Smartphones
Samuli Hemminki, Petteri Nurmi, Sasu Tarkoma
Helsinki Insitute for Information Technology HIIT
PO Box 68, Department of Computer Science
FI-00014, University of Helsinki, Finland
firstname.lastname@cs.helsinki.fi
ABSTRACT
We present novel accelerometer-based techniques for accu-
rate and fine-grained detection of transportation modes on
smartphones. The primary contributions of our work are an
improved algorithm for estimating the gravity component
of accelerometer measurements, a novel set of accelerometer
features that are able to capture key characteristics of vehic-
ular movement patterns, and a hierarchical decomposition
of the detection task. We evaluate our approach using over
150 hours of transportation data, which has been collected
from 4 different countries and 16 individuals. Results of
the evaluation demonstrate that our approach is able to im-
prove transportation mode detection by over 20% compared
to current accelerometer-based systems, while at the same
time improving generalization and robustness of the detec-
tion. The main performance improvements are obtained for
motorised transportation modalities, which currently repre-
sent the main challenge for smartphone-based transporta-
tion mode detection.
Categories and Subject Descriptors
I.5.2 [Pattern Recognition]: Design Methodology: Fea-
ture evaluation and selection; I.5.4 [Pattern Recognition]:
Applications: Signal processing; H.4.m [Information Sys-
tems]: Information Systems Applications: Miscellaneous
General Terms
Algorithms, Experimentation
Keywords
Mobile Sensing, Activity Recognition, Transportation Mode
Detection
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. To copy otherwise, to
republish, to post on servers or to redistribute to lists, requires prior specific
permission and/or a fee. Request permissions from Permissions@acm.org.
SenSys’13, November 11 - 15 2013, Roma, Italy.
Copyright 2013 ACM 978-1-4503-2027-6/13/11 ...$15.00
http://dx.doi.org/10.1145/2517351.2517367.
1. INTRODUCTION
The increased sensing capabilities of contemporary smart-
phones combined with their easy programmability, large mar-
ket penetration rate, and effective distribution channels for
third party applications, have resulted in smartphones ma-
turing into an effective tool for unobtrusive monitoring of
human behavior [21]. This paper focuses on a specific as-
pect of human behavior, the transportation behavior of indi-
viduals. The capability to capture transportation behavior
accurately on smartphones would have a positive impact on
many research fields. For example, human mobility track-
ing would directly benefit from an ability to automatically
monitor the transportation behavior of individuals [18, 28].
This in turn would enable improving urban planning [38],
monitoring and addressing the spread of diseases and other
potential hazards, as well as providing emergency respon-
ders information of the fastest route to aid the lost or in-
jured [29]. Localization and positioning algorithms could be
improved by constructing more elaborated motion models
with the help of information of the user’s current trans-
portation mode [25] or by constricting the plausible loca-
tions of the user to the route of the detected transportation
mode. Persuasive applications could use the transportation
behavior monitoring to automatically calculate, for example,
CO
2
-footprint or level of physical activity [9]. Finally, trans-
portation monitoring could be used as part of user profiling,
e.g., for real-time journey planning and guidance systems,
or targeted advertising.
While the idea to use smartphones for monitoring trans-
portation behavior itself is not new (see Sec. 2), previous
work has primarily focused on elaborate use of the phone’s
integrated GPS receiver. While GPS-based systems can be
very efficient when GPS signals are available, they suffer
from some important limitations. First, integrated GPS re-
ceivers are well-know to suffer from high power consumption,
which means these approaches rapidly deplete the battery of
the mobile device, especially when the user is mobile. Sec-
ond, the GPS receiver’s dependency on unobstructed view
to satellites presents problems in many common situations
of urban transportation, e.g., while the user is moving un-
derground, inside a station, along urban canyons, or while
the user is traveling in a vehicle but is unable to stay suf-
ficiently close to a window. Third, current GPS-based so-
lutions provide only modest accuracy when a fine-grained
distinction of motorised transportation modes is required.
Distinguishing reliably between different motorised trans-
portation modes would provide more detailed information
about human transportation behavior, e.g., enabling to au-
tomatically estimate the carbon footprint of individuals or
to obtain a detailed understanding of the commuting pat-
terns of urban citizens.
In this paper we present novel accelerometer-based tech-
niques which can be used individually, or in conjunction
with other sensors for transportation mode detection on
smartphones. We focus on accelerometers as they are well-
suited to overcome the above mention limitations. First,
accelerometers have very low power consumption, enabling
continuous transportation behavior monitoring. Second, ac-
celerometers measure user’s movements directly and there-
fore do not depend on any external signal sources. Third,
accelerometers contain highly detailed information about
phone movement, enabling fine-grained distinction of dif-
ferent motorised transportation modalities.
A central challenge in accelerometer-based transportation
mode detection is to distinguish information pertaining to
movement behavior from other factors that affect the ac-
celerometer signals. In particular, gravity, user interactions
and other sources of noise can mask the relevant information.
As our first technical contribution, we describe a novel grav-
ity estimation technique for accelerometer measurements.
Our gravity estimation technique provides more accurate
and robust gravity component estimation during motorised
transportation, which in turn translates to more accurate
horizontal accelerometer representation. The horizontal ac-
celeration is a key factor for identifying motorised modali-
ties, as the acceleration/deceleration periods are typically
similar within the same modality, but also distinct from
other modalities. The real-world analogy is that different
types of vehicles can be identified from their acceleration
and breaking periods. As our second technical contribution,
we introduce a new set of accelerometer features, so-called
peak features, which characterize acceleration and decelera-
tion patterns during motorized modalities. These features
are a key enabler for improving the detection performance of
transportation mode detection approaches. Moreover, these
features pave way for new types of sensing applications that
analyze vehicular movement patterns, e.g., driving style or
fuel consumption estimation.
We evaluate our approach using over 150 hours of trans-
portation data collected from 16 individuals and 4 different
countries. The results of our evaluation demonstrate that
our approach is able to improve detection accuracy by over
20% compared to current accelerometer-based solutions, and
even exceed the accuracy of the current state-of-art hybrid
GPS and accelerometer system by over 10%. Moreover, our
proposed techniques improve the robustness of the detection
and generalize well across users and geographical locations.
2. RELATED WORK
Transportation mode detection can be considered a special
subfied of activity recognition, a widely studied field within
the wearable and ubiquitous computing communities [17].
The first approaches to transportation mode detection re-
lied on custom sensing platforms [1, 5], whereas recent work
has predominantly considered smartphones as the sensing
platform. In the following we focus exclusively on previous
work on smartphone-based transportation mode detection.
We refer to [26] for information about approaches that rely
on custom sensing platforms.
Transportation mode detection can be subdivided into two
main subtasks: (i) determining whether the user is moving;
and, in case movement is detected, (ii) what kind of means
the person is using for moving around. The former task,
stationarity detection, has been widely explored in different
domains. For example, the LOCADIO positioning system
classifies the user as mobile or stationary based on changes in
the WiFi signal environment [16]. Muthukrishnan et al. [24]
detect mobility by analyzing spectral characteristics of the
WiFi signal environment. Kim et al. [14] use variance of ac-
celerometer values to detect when a person is staying within
one place. Kjaergaard et al. [15] combine accelerometer vari-
ance with a threshold on GPS speed to separate motorised
transportation modalities from stationary behavior.
The latter task, locomotion detection, has been increas-
ingly explored on smartphones. Typical locomotion types
include different pedestrian modalities (e.g., walking, run-
ning or moving in stairs), non-motorised transportation (e.g.,
bicycling, roller skating) and motorised transportation (e.g.,
bus, train or car). In terms of sensors, the accelerometer
is the most widely used sensor for detecting locomotion. A
number of early systems used the embedded accelerometer
for detecting different pedestrian and non-motorised modal-
ities, such as walking and running [20, 12], ascending or
descending stairs [4] or cycling [2]. Wang et al. [31] compare
features extracted from the L
2
norm with features extracted
from horizontal and vertical representations. From each rep-
resentation, an extensive set of features is extracted over 8
second non-overlapping windows, and classification is per-
formed using a decision tree. In contrast to our work, the
authors obtain the best results using only the L
2
norm repre-
sentation, which is due to inaccurate estimation of the grav-
ity component and insufficiently expressive feature space.
While previous accelerometer-based systems have been effec-
tive at detecting pedestrian and non-motorised transporta-
tion modalities, achieving typically over 90% accuracies, their
performance has been significantly lower for stationary and
motorised transportation modalities [26, 31]. Our work im-
proves on these approaches by demonstrating that capturing
features from the acceleration/deceleration periods from ve-
hicular motion can be used for accurate and fine-grained
detection of motorised transportation modalities.
Instead of relying on the accelerometer, Zheng et al. [36,
37] detect transportation modalities using features extracted
from GPS measurements. In addition to speed and location
information, the authors consider features that characterize
changes in movement direction, velocity and acceleration.
Together with information about street segments, the au-
thors reach an average accuracy of 76% in classifying be-
tween stationarity, walking, biking, driving and traveling by
bus. Recent work has focused on decreasing energy con-
sumption by requiring only sparse GPS data [3], introducing
more effective graph-based postprocessing techniques [36],
and improving the detection accuracy by fusing in external
information on the real-time location of the transportation
vehicles [30]. Reddy et al. [26] combine GPS and accelerom-
eter to recognize between stationary, walking, running, bik-
ing and motorised transportation, achieving high, over 90%
accuracies. Classification is performed with a hybrid clas-
sifier consisting of a decision tree and a first order discrete
HMM classifier. However, in contrast to our work, Reddy et
al. make no distinction between different motorised modal-
ities and mainly rely on GPS speed for detecting motorised
transportation. The main drawbacks of all these approaches
are that GPS receiver has high power consumption, requires
inconsistent time for obtaining satellite lock, and is unavail-
able or unreliable when view to satellites is obstructed, e.g.,
when the user is underground, inside a station, moving in
urban canyons or is insufficiently close to a window in a
transportation vehicle.
An alternative to GPS is to estimate movement by mon-
itoring changes in the user’s signal environment. Sohn et
al. [27] use changes in the GSM signal environment for coarse-
grained detection of transportation modalities. Mun et al. [23]
combine GSM and WiFi for detecting between dwelling,
walking and driving, reaching accuracies in the range of
80 − 90%. While energy-efficient compared to GPS, these
techniques are susceptible to varying WiFi access point den-
sity and GSM cell sizes between different locations. Conse-
quently, these techniques are unreliable outside urban areas
and require careful calibration, and thus, struggle to gener-
alize to new environments.
3. TRANSPORTATION MODE DETECTION
We have developed a novel solution for transportation
mode detection that can provide robust, accurate and fine-
grained detection of transportation modalities despite rely-
ing solely on the embedded accelerometer of the smartphone.
The key technical contributions of our work are (i) an im-
proved algorithm for estimating the gravity component of
the accelerometer measurements, (ii) a novel class of fea-
tures, extracted from the horizontal accelerometer represen-
tation, that are capable of capturing characteristics of accel-
eration and breaking patterns for different motorised trans-
portation modalities; and (iii) a hierarchical decomposition
of the overall detection task. As our experiments demon-
strate, the combination of these contributions provides sig-
nificant improvements in the accuracy of transportation mode
detection, in particular for motorised transportation modali-
ties. We have implemented our approach on Android smart-
phones and integrated it as part of a mobile application that
aims at motivating people to reduce their CO2 consump-
tion [13]. In the remainder of this section we describe the
different components of our approach in detail.
3.1 Overview
Our approach decomposes transportation mode detection
hierarchically into subtasks, proceeding from a coarse-grained
classification towards a fine-grained distinction of transporta-
tion modality. At the core of our system are three classifiers,
which are organized into a hierarchy; see Fig. 1. At the root
of the hierarchy is a kinematic motion classifier which per-
forms a coarse-grained distinction between pedestrian and
other modalities. When the kinematic motion classifier fails
to detect substantial physical movement, the process pro-
gresses to a stationary classifier, which determines whether
the user is stationary or in a motorised transport. When
motorised transportation is detected, the classification pro-
ceeds to a motorised classifier which is responsible for clas-
sifying the current transportation activity into one of five
modalities: bus, train, metro, tram or car.
Changes in transportation behavior typically occur in-
frequently and each activity has duration of several min-
utes. Furthermore, changes from one motorised transporta-
tion modality to another are typically separated by eas-
ily detectable walking segments [36]. Our approach treats
the different transportation activities as segments instead of
Figure 1: Overview of the classifiers used in our
system and their dependencies.
performing solely frame-by-frame classification. Segments
contain more information than the frames can express in-
dividually, resulting in improved classification performance.
Specifically, as more evidence to support one of the modal-
ities accumulates during a segment, the prediction becomes
increasingly accurate. The segment-wise classification is con-
tinued until change in transportation mode is detected. Sta-
tionary periods within motorised modality are interpreted as
being in a stopped vehicle, e.g., due to traffic lights or stop-
ping at a station. A detailed description of the frame and
segment-based classifiers are is given in Sec. 3.4.
3.2 Preprocessing and Gravity Estimation
We consider three dimensional acceleration measurements
obtained from contemporary smartphones. We preprocess
the raw measurements by applying a low-pass filter that re-
tains 90% of energy. This is performed to remove jitter from
the measurements and is in line with current best practices.
Next, we aggregate the measurements using a sliding win-
dow with 50% overlap and a duration of 1.2 seconds. The
length of the window was selected to ensure the monitoring
can rapidly react to changes in the transportation behav-
ior of the user. Once the measurements have been filtered,
we project the sensor measurements to a global reference
frame by estimating the gravity component along each axis
and calculating gravity eliminated projections of vertical and
horizontal acceleration. We consider a novel method for esti-
mating the gravity component from accelerometer measure-
ments that improves the robustness of gravity estimation,
particularly in the presence of sustained acceleration.
Currently the dominant approach for estimating the grav-
ity component from accelerometer measurements is to use
the mean over a window of fixed duration [19, 22]. While ele-
gant and simple, this approach, first proposed by Mizell [22],
suffers from two fundamental limitations. First, this ap-
proach is inherently based on the assumption that, given
a sufficiently long window of measurements, noise and ob-
served accelerometer patterns are uncorrelated over time.
This assumption does not hold during sustained accelera-
Algorithm 1 Gravity (Accelerometer
window
, TH
var
)
1: W
mean
= mean(Accelerometer
window
)
2: W
var
= var(Accelerometer
window
)
3: if ||W
mean
− G
est
|| ≥ 2m/s
2
then
4: T H
var
= Reset variance threshold
5: end if
6: if W
var
< 1.5 then
7: if W
var
< T H
var
then
8: G
est
= W
mean
9: T H
var
= (W
var
+ T H
var
)/2
10: V arIncrease = T H
var
∗
inc
11: else
12: T H
var
= T H
var
+ V arIncrease
13: end if
14: else
15: G
est
= MizellEstimate(5s)
16: end if
tion, e.g., during any motorised transportation. Second,
when the orientation of the sensor suddenly changes, e.g.,
when the user sits down or stands up, there is a consider-
able lag before the gravity estimates are accurate again. The
lag of the approach can be reduced by shortening the time
window over which gravity is estimated, as has been used in
several sensing systems [19, 35]. However, this improvement
comes with a decrease in the accuracy of the gravity esti-
mates, making it difficult to detect sustained acceleration.
To illustrate these limitations, Figure 2(a) shows the grav-
ity estimation produced during a tram ride by the approach
used in the Jigsaw system [19], i.e., by using the mean over
a four second window. From the figure we can observe that
this approach tracks the raw acceleration measurements too
closely, removing all information that is relevant for distin-
guishing between the different transportation modalities.
To overcome the above mentioned limitations, we have de-
veloped a novel algorithm for estimating the gravity compo-
nent of accelerometer measurements. Our approach, sum-
marized in Alg. 1, considers short data windows and esti-
mates the gravity by opportunistically identifying periods
where the variation in sensor measurements is sufficiently
small, i.e., below a suitable threshold. During these peri-
ods, the sensor is approximately stationary, which means
that the main force exerting the sensor values is gravity.
In many situations, such as walking, bicycling, or traveling
with motorised transportation along an uneven road, the
measurements contain large variation for a sustained period
of time and no opportunities for gravity estimation occur.
To estimate gravity during these situations, we dynamically
adjust the variance threshold according to the current move-
ment patterns. We allow the variance threshold to increase
until a hard upper threshold is reached (currently, we use
variance of 1.5), after which the gravity estimates would
become overly inaccurate, and utilizing Mizell’s technique
becomes more suitable.
To reduce the influence of orientation changes on the grav-
ity estimate, we reset the estimate of the gravity component
when a large shift in orientation is observed. These shifts
are typically caused by extraneous activities, such as user
interaction or shifts in orientation due to, e.g., standing up
or sitting down. We detect shifts in orientation by compar-
ing the current gravity estimate against the mean of the last
measurement window. Whenever these differ by more than
(a) Mizell with a four second window.
(b) Our approach.
Figure 2: Comparison of the gravity estimation be-
tween the algorithm of Mizell and our approach.
The estimated gravity corresponds to the solid red
line.
Our Miz-1 Miz-10 Miz-30
Bus 0.32 0.09 0.24 0.27
Train 0.54 0.16 0.25 0.42
Metro 0.51 0.18 0.32 0.41
Tram 0.35 0.09 0.31 0.32
Table 1: Correlation coefficient between integral of
the horizontal gravity eliminated acceleration and
GPS speed using our method versus using Mizell’s
method.
a specific threshold (currently, we use 2m/s) along any of
the axes, we re-initialize the gravity estimate for each axis
to the mean of current accelerometer window. As illustrated
in Fig. 2(b), our gravity estimation is particularly effective
while traveling within motorised transportation where pe-
riods of low variance typically are interleaved within accel-
eration and deceleration patterns. In the scenario depicted
in the figure, the phone’s orientation was relatively stable
throughout the tram ride, implying that the estimated grav-
ity component should be approximately constant.
To further demonstrate the benefits of our gravity estima-
tion algorithm, we have conducted a small-scale experiment
using a dataset consisting of slightly over 7 hours or data
from different motorised transportation modalities. In this
experiment, we have compared speed information obtained
from GPS with the numeric integration of the gravity elimi-
nated horizontal projection of the acceleration. The numeric
integration calculates the area under the gravity eliminated
horizontal acceleration, which can be used to estimate the
speed of the user [6]. The results of this experiment, pre-
sented in Table 1, demonstrate a strong correlation between
the gravity eliminated horizontal projection and the speed
information obtained from the GPS. Compared to the ap-
proach of Mizell, our algorithm provides better correlation
with the speed obtained from GPS for all the evaluated
cases, despite using a short time window.
3.3 Feature Extraction
Once the sensor values have been preprocessed and trans-
formed, we construct gravity eliminated horizontal and ver-
tical representations of the accelerometer measurements. We
extract features on three different levels of granularity. The
three sets of features are referred to as frame-based, peak-
based and segment-based features with respect to the feature
source. Below we detail each feature set and describe their
function in the detection task. For a full list of features; see
Table 2.
Frame-based features
The frame-based features considered in our study were cho-
sen based on an analysis of accelerometer features conducted
by Figo et al. [6]. From each frame, we extract 27 fea-
tures from both vertical and horizontal representations, i.e.,
the total number of features we consider from each frame
is 54. The features we extract include statistical features
(e.g., mean, variance and kurtosis), time-domain metrics
(e.g., double integral, auto-correlation and zero crossings)
and frequency-domain metrics (e.g., energy, six first FFT
components, entropy and the sum of FFT coefficients). The
frame-based features are able to effectively capture char-
acteristics of high-frequency motion caused by, e.g., user’s
physical movement during pedestrian activity, or during mo-
torised periods, from vehicle’s engine and contact between
its wheels and surface.
Peak-based features
While the frame-based features can effectively capture in-
formation from high-frequency motion, they are unable to
capture movement with lower frequencies, such as accelera-
tion and breaking periods of motorised vehicles, which are
essential for distinguishing between the different motorised
transportation modalities. To capture features from these
key periods of vehicular movement, we use the horizontal
acceleration projection to extract a set of novel peak-based
features that characterize acceleration and deceleration peri-
ods. As the kinematic activities are largely characterized by
high-frequency motion, we extract the peak-based features
only during stationary and motorised periods, i.e., when the
kinematic classifier fails to detect substantial, cyclic kine-
matic movement.
To extract these features, we identify so-called peak ar-
eas that correspond to acceleration or breaking events; see
Figure 3. We identify peak areas by first applying a stream-
based event detection algorithm to identify significant changes
in the gravity eliminated horizontal acceleration. Once a
significant change has been observed, we mark the corre-
sponding time instant as the starting boundary of the peak
area. We buffer subsequent measurements until a significant
decrease in the magnitude of gravity eliminated horizontal
acceleration is observed, i.e., until the horizontal accelera-
tion levels out. Currently we use a pre-defined threshold of
Domain Features
Statistical Mean, STD, Variance, Median, Min,
Max, Range, Interquartile range
Kurtosis, Skewness, RMS
Time Integral, Double integral, Auto-Correlation,
Mean-Crossing Rate,
Frequency FFT DC,1,2,3,4,5,6 Hz, Spectral Energy,
Spectral Entropy, Spectrum peak position,
Wavelet Entropy, Wavelet Magnitude
Peak Volume (AuC), Intensity, Length,
Kurtosis, Skewness
Segment Variance of peak features (10 features),
Peak frequency (2 features),
Stationary duration, Stationary frequency
Table 2: Full list of the features considered for our
classifiers.
Figure 3: Peak areas detected from gravity elimi-
nated horizontal acceleration during a metro ride.
0.2m/s
2
as the threshold for identifying the end boundary
of the peak area. Once the starting and ending boundaries
have been identified, we extract a set of statistical features
that characterize the peak area; see Table 2 for the features
that are considered. We calculate these features separately
for peaks corresponding to acceleration and to breaking pe-
riods, resulting in 10 peak features.
Segment-based features
In addition to the frame and peak-based features, we extract
segment-based features that characterize patterns of accel-
eration and deceleration periods over the observed segment,
i.e., during a period of stationary or motorised movement.
The segment-based features we consider are the frequency of
acceleration and breaking periods, the frequency and dura-
tion of the intermittent stationary periods, and the variance
of individual peak-based features. The former two of these
are analogous to the velocity change rate and stopping rate
features that Zheng et al. [36, 37] use as part of their GPS-
based transportation mode detection approach. In total we
consider 14 segment-based features.
To illustrate the potential of the peak and segment-based
features to provide fine-grained detection of motorised trans-
