Human Activity Recognition

Using Sensor Data of Smartphones and Smartwatches

Bishoy Sefen

, Sebastian Baumbach

2,3

, Andreas Dengel

2,3

and Slim Abdennadher

German University in Cairo, Cairo, Egypt

University of Kaiserslautern, Kaiserslautern, Germany

German Research Center for Artiﬁcial Intelligence (DFKI), Kaiserslautern, Germany

Keywords:

Activity Recognition, Fitness Games, Sensors, Smartphones, Smartwatches, Machine Learning, Data Mining.

Abstract:

Unobtrusive and mobile activity monitoring using ubiquitous, cheap and widely available technology is the

key requirement for human activity recognition supporting novel applications, such as health monitoring. With

the recent progress in wearable technology, pervasive sensing and computing has become feasible. However,

recognizing complex activities on light-weight devices is a challenging task. In this work, a platform to com-

bine off-the-shelf sensors of smartphones and smartwatches for recognizing human activities in real-time is

proposed. In order to achieve the best tradeoff between the system’s computational complexity and recogni-

tion accuracy, several evaluations were carried out to determine which classiﬁcation algorithm and features

to be used. Therefore, a data set from 16 participants was collected that includes normal daily activities and

several ﬁtness exercises. The analysis results showed that naive Bayes performs best in our experiment in both

the accuracy and efﬁciency of classiﬁcation, while the overall classiﬁcation accuracy is 87% ± 2.4.

1 INTRODUCTION

Physical activity is well-known by the general pub-

lic to be crucial for leading a healthy life. Thus, re-

searchers are seeking a better understanding of the re-

lationship between physical activity and health. Pre-

cise recording of the conducted activities is an essen-

tial requirement of their research. (Bauman et al.,

2006)

This data can be used to design and con-

struct activity recognition systems. These systems

allow physicians to check the recovery develop-

ment of their patients automatically and constantly

(da Costa Cachucho et al., 2011). Another rising con-

cern is the sedentary life many people live, due to the

shift in lifestyle occurring in the modern world, where

work and leisure tend to be less physically demand-

ing (Gyllensten, 2010). Several reports have already

found links between common diseases and physical

inactivity (Preece et al., 2009). Thus, activity recog-

nition can be used by recommender systems to help

the users track their daily physical activity and pro-

mote them to increase their activity level.

With the recent progress in wearable technology,

unobtrusive and mobile activity recognition has be-

come reasonable. With this technology, devices like

smartphones and smartwatches are widely available,

hosting a wide range of built-in sensors, at the same

time, providing a large amount of computation power.

Overall, the technological tools exist to develop a mo-

bile, unobtrusive and accurate physical activity recog-

nition system. Therefore, the realization of recogniz-

ing the individuals’ physical activities while perform-

ing their daily routine has become feasible. So far,

no-one has investigated the usage of light-weight de-

vices for recognizing human activities.

An activity recognition system poses several main

requirements. First, it should recognize activities in

real-time. This demands that the features used for

classiﬁcation are computable in real-time. Moreover,

short window durations must be employed to avoid

delayed response. Finally, the classiﬁcation schemes

should be simple, light-weight and computationally

inexpensive to be able to run on hand-held devices.

In this work, a method for recognizing human

activities using the acceleration sensors incorporated

in smartphones and smartwatches is proposed, while

fulﬁlling the requirements stated above (Section 3).

Therefore, an experiment is done to collect data from

16 participants (Section 4). Different classiﬁcation al-

gorithms are evaluated in order to ﬁnd the best trade-

off between computational complexity and recogni-

488

Sefen, B., Baumbach, S., Dengel, A. and Abdennadher, S.

Human Activity Recognition - Using Sensor Data of Smartphones and Smartwatches.

DOI: 10.5220/0005816004880493

In Proceedings of the 8th International Conference on Agents and Artiﬁcial Intelligence (ICAART 2016) - Volume 2, pages 488-493

ISBN: 978-989-758-172-4

tion accuracy, in addition to, evaluating the best fea-

tures to be used with them (Section 5). Finally, the

results show that this platform is able to recognize

various human activies and using a smartwatch com-

bined with a smartphone improves the accuracy of the

recognition process (Section 6).

2 RELATED WORK

Past work focused on the use of multiple accelerom-

eters placed on several parts of the user’s body.

Bao and Intille used ﬁve bi-axial accelerometers dis-

tributed across the user’s body. They tested their ap-

proach with data of twenty users. (Bao and Intille,

2004). Krishnan et al. used two accelerometers to

recognize ﬁve activities (Krishnan et al., 2008). They

collected data from only three users. Parkka et. al.

created a system using twenty different types of sen-

sors in order to recognize activities such as football,

croquet, and using the toilet (Parkka et al., 2006).

Subramayana et. al. addressed normal daily ac-

tivities by using data not only from a tri-axial ac-

celerometer, but from micro-phones, temperature sen-

sors and barometric pressure sensors as well (Subra-

manya et al., 2012). These systems using multiple ac-

celerometers and other sensors were capable of iden-

tifying a wide range of activities. However, they are

not practical as they involve the user wearing multiple

sensors distributed across his body.

Other studies focused on the use of a single ac-

celerometer for activity recognition. Long et al.

placed a tri-axial accelerometer worn at the user’s

waist, recognizing walking, jogging, running, cy-

cling, and sports of twenty four users (Long et al.,

2009). Lee et. al. used a single accelerometer at-

tached to the left waists of only ﬁve users (Lee, 2009).

However, all of these studies used devices speciﬁcally

made for research purposes.

Several investigations have considered the use of

widely available mobile devices. Ravi et. al. col-

lected data from only two users wearing a single

accelerometer-based device and then transmitted this

data to the phone carried by the user (Ravi et al.,

2005). Lester et. al. used accelerometer data from

a small set of users along with audio and barometric

sensor data to recognize eight daily activities (Lester

et al., 2006). However, the data was generated using

distinct accelerometer-based devices worn by the user

and then sent to the phone for storage.

Some studies took advantage of the sensors incor-

porated into the phones themselves. Yang developed

an activity recognition system using a smartphone to

distinguish between various activities (Yang, 2009).

However, stair climbing was not considered and their

system was trained and tested using data from only

four users. Brezmes et. al. developed a real-time

system for recognizing six user activities (Brezmes

et al., 2009). In their system, an activity recognition

model is trained for each user, i.e., there is no univer-

sal model that can be applied to new users for whom

no training data exists. Bayat et al. gathered acceler-

ation data from only four participants, performing six

activities. (Bayat et al., 2014) Shoaib et al. evaluated

different classiﬁers by collecting data of smartphone

accelerometer, gyroscope, and magnetometer for four

subjects, perfoming six actvities. (Shoaib et al., 2013)

3 METHODOLOGY

In this section, the activity recognition process is de-

scribed, containing four main stages.

3.1 Data Collection

The ﬁrst step is to collect multivariate time series data

from the phone’s and the watch’s sensors. The sensors

are sampled with a constant frequency of 30 Hz. Af-

ter that, the sliding window approach is utilized for

segmentation, where the time series is divided into

subsequent windows of ﬁxed duration without inter-

window gaps (Banos et al., 2014). The sliding win-

dow approach does not require preprocessing of the

time series, and is therefore ideally suited to real-time

applications.

3.2 Preprocessing

Filtering is performed afterwards to remove noisy val-

ues and outliers from the accelerometer time series

data, so that it will be appropriate for the feature

extraction stage. There are two basic types of ﬁl-

ters that are usually used in this step: average ﬁlter

(Sharma et al., 2008) or median ﬁlter (Thiemjarus,

2010). Since the type of noise dealt with here is sim-

ilar to the salt and pepper noise found in images, that

is, extreme acceleration values that occur in single

snapshots scattered throughout the time series. There-

fore, a median ﬁlter of order 3 (window size) is ap-

plied to remove this kind of noise.

3.3 Feature Extraction

Here, each resulting segment will be summarized by

a ﬁxed number of features, i.e., one feature vector

per segment. The used features are extracted from

Human Activity Recognition - Using Sensor Data of Smartphones and Smartwatches

489

both time and frequency domains. Moreover, all fea-

tures are going to be computed from the 3 acceleration

components A

, A

, in addition to, a 4

compo-

nent derived as

+ A

, which is known as

the magnitude component (see Figure 1). Therefore

for each device (phone and watch), there are going to

be 4 values per feature type, i.e., 8 values combined

per feature type.

Time (s)

0 1 2 3 4 5 6 7 8 9 10

Acceleration (m/s

)

-10

-5

mag

Figure 1: The accelerometer time series data of ascending

the stairs, consisting of the three axial components plus the

derived magnitude component (mag).

In the time domain, the following statistical fea-

tures are computed: Mean, Minimum, Maximum,

Range, Standard Deviation, and Root-Mean-Square.

Since, many activities have a repetitive nature, i.e.,

they consist of a set of movements that are done pe-

riodically like walking and running. This frequency

of repetition, also known as dominant frequency, is

a descriptive feature and thus, it has been taken into

consideration. (Telgarsky, 2013). Consequently, fast

Fourier transform (FFT) is performed on the time se-

ries outputting a list of frequencies along with their re-

spective magnitudes (Sharma et al., 2008). Then, the

frequency with the highest magnitude will be selected

as the dominant frequency. Moreover, to increase the

descriptiveness of the features, the 2

dominant fre-

quency will also be selected.

To sum up, 8 different types of features will be

computed, 6 from the time domain, while 2 from the

frequency domain. Since each feature type is ex-

tracted from 4 components, therefore, 32 features will

be used to summarize the accelerometer time series.

Finally, the features computed from both the phone’s

and the watch’s sensors will be combined, producing

a 64 value feature vector.

3.4 Standardization

Since, the time domain features are measured in

(

), while the frequency ones in (Hz), therefore, all

features should have the same scale for a fair compar-

ison between them, as some classiﬁcation algorithms

use distance metrics. In this step, Z-Score standard-

ization is used, which will transform the attributes to

have zero mean and unit variance, and is deﬁned as

new

x−µ

, where µ and σ are the attribute’s mean and

standard deviation respectively (Gyllensten, 2010).

4 EXPERIMENT

An experiment was conducted, where the participants

wore the watch on the left hand and placed the phone

in the front right pocket of the pants while performing

diverse activities. A total of 16 participants (8 males

and 8 females) performed the complete set of the se-

lected activities (see table 1). This activities can be

grouped under normal everyday activities, or ﬁtness

exercises. The four everyday activities are walking,

jogging, idle (both standing and sitting), and using

the stairs (both ascending and descending). While the

selected ﬁtness exercises are rope jumping, pushups,

crunches, and squats.

Table 1: The details of the collected recordings for each

activity type.

Activity Recording 1 Recording 2 Recording 3 Recording 4

Walking 100 meters 100 meters 100 meters ˜

Jogging 100 meters 100 meters 100 meters ˜

Idle

Standing

20 seconds

Sitting

20 seconds

Standing

20 seconds

Sitting

20 seconds

Stair Climbing

Ascending

16 stair steps

Descending

16 stair steps

Ascending

16 stair steps

Descending

16 stair steps

Rope jumping 20 seconds 20 seconds 20 seconds ˜

Pushups 5 repetitions 5 repetitions 5 repetitions ˜

Crunches 5 repetitions 5 repetitions 5 repetitions ˜

Squats 5 repetitions 5 repetitions 5 repetitions ˜

The datasets used later in the evaluations were de-

rived afterwards. The window durations used for seg-

menting the recordings are 1, 2, 3, 4, and 5 seconds.

Therefore, 5 different datasets resulted. Information

about the composition of these datasets is shown in

Table 2, where the samples number for each activity,

and the total number of samples are shown for each

dataset. Finally, dataset x is the one resulting from

segmenting the recordings using a x seconds window.

Table 2: The composition of the used datasets.

Dataset

Idle

Walking

Jogging

Stairs

Rope jumping

Pushups

Crunches

Squats

Total

Dataset 1 1228 3190 1334 531 817 430 449 448 8427

Dataset 2 605 1581 655 248 402 204 211 210 4116

Dataset 3 397 1048 429 158 262 126 134 137 2691

Dataset 4 277 778 317 107 184 91 89 95 1938

Dataset 5 213 620 245 82 140 64 74 67 1505

ICAART 2016 - 8th International Conference on Agents and Artiﬁcial Intelligence

490

5 EVALUATION

Several evaluations were conducted in order to ﬁnd

the optimal parameters and conﬁgurations to be used,

taking into consideration mainly the recognition ac-

curacy, speed of execution, and their applicability to

real-time recognition.

To obtain reliable user-independent results, leave-

one-participant-out cross validation (LOPOCV) is

used, where the left out part in each iteration is the

entire collection of samples from a single participant.

Since the collected data consists of 16 participants,

therefore, 16 iterations of train-test will be performed.

A common greedy feature selection algorithm

known as forward selection is used to ﬁnd the best

scoring feature subset for each classiﬁcation algo-

rithm candidate. However, the search of possible fea-

ture subsets is likely to ﬁnd a misleadingly well scor-

ing one by chance. To prevent this, a complex scor-

ing function that evaluates feature subsets against sev-

eral datasets is used. At ﬁrst, it uses the leave-one-

participant-out cross validation (LOPOCV) to com-

pute the accuracies for the ﬁve datasets, then, it com-

putes the ﬁnal score as the average of the resulting

accuracies.

In this work, the following classiﬁcation algo-

rithms are evaluated: Support Vector Machine: linear

and polynomial kernels, One-Vs-All method, Deci-

sion Tree: Gini index as a split criteria, Naive Bayes:

Gaussian distribution, Discriminant Analysis: linear

and quadratic, K-Nearest Neighbors: Euclidean dis-

tance metric, K from 1 till 10.

5.1 Classiﬁcation Accuracy

The ﬁrst evaluation aims at comparing the perfor-

mance of different classiﬁcation algorithms, in addi-

tion to, determining the best feature subset for each

one. However, one of the variables that can affect the

results is the window duration. At the same time, per-

forming this evaluation while ﬁxing the window dura-

tion, would lead to the argument that the ﬁndings are

dependent on the used window duration and dataset,

i.e., they can not be generalized to other window du-

rations.

The results of the forward feature selection shows

that naive Bayes reached the highest average accu-

racy, which is 89.4% for the subset consisting of 18

features. Moreover, this 89.4% accuracy is derived

from averaging the individual LOPOCV accuracies

of each dataset, which are 85.5%, 88.6%, 90.5%,

90.9%, and 91.5% respectively. Finally, the best

performing feature subset for naive Bayes consists

of 18 features, which are: PMinZ, WStdM, PStdM,

WMinM, WFq1M, PFq1Z, PStdY, WAvgZ, WStdY,

PFq2Z, PMinM, PFq2M, WFq2M, WFq1Y, WAvgX,

PFq1X, WFq2Z, and PFq2Y.

P: phone, W: watch; Avg: average, Min: minimum,

Std: standard Deviation, Fq1: most dominant fre-

quency, Fq2: second most dominant frequency; X:

x-component, Y: y-component, Z: z-component, M:

magnitude component.

The confusion matrix for the highest scoring fea-

ture subset is presented in Table 3, where the rows

correspond to the actual performed activities, while

the columns correspond to the predicated activity la-

bels. This matrix is derived from the classiﬁcation re-

sults of the 5 datasets, which means it is the accumu-

lation of 5 confusion matrices (one for each dataset).

Table 3: The confusion matrix for the naive Bayes classiﬁer.

Activity

Idle

Walking

Jogging

Stair Climbing

Rope jumping

Pushups

Crunches

Squats

Recall

Idle 2589 0 0 0 0 94 36 1 95.2

Walking 0 6620 15 424 0 1 0 157 91.7

Jogging 0 2 2769 0 209 0 0 0 92.9

Stair Climbing 6 528 14 522 3 9 3 41 46.4

Rope jumping 0 0 220 0 1585 0 0 0 87.8

Pushups 8 6 2 4 0 826 38 31 90.3

Crunches 34 7 23 5 0 71 735 82 76.8

Squats 0 28 16 44 0 14 75 780 81.5

Precision 98.2 92.1 90.5 52.3 88.2 81.4 82.9 71.4

The results in Table 4 shows that naive Bayes is

the most accurate classiﬁer, scoring an average accu-

racy of 89.4% using 18 features subset. However,

the classiﬁcation accuracy lies between 84.6% and

89.4%, at which the differences in performance are

negligible.

Table 4: The highest accuracies obtained for each classiﬁ-

cation algorithm along with the size of their feature subsets.

Algorithm Number of Features Average Accuracy (%)

SVM (Linear) 40 84.6

SVM (Polynomial) 20 87.9

Decision Tree 17 86.3

Naive Bayes 18 89.4

LDA 45 87.9

QDA 15 89.1

K-NN (K = 6) 13 89.3

Weighted K-NN (K = 5) 14 89.3

5.2 Classiﬁcation Speed

The aim for the computational efﬁciency is to com-

pare the candidate classiﬁcation algorithms from the

previous evaluation based on their efﬁciencies. More-

over, each algorithm will be tested with it’s best per-

forming feature subset.

The algorithms’ efﬁciencies was compared by

Human Activity Recognition - Using Sensor Data of Smartphones and Smartwatches

491

measuring the time they take to complete on a spe-

ciﬁc machine. This run-time evaluation is performed

while computing the LOPOCV using Dataset 5, i.e., it

measures the amount of time spent in the training and

the classiﬁcation phases of this validation algorithm.

As described earlier, the LOPOCV splits the dataset

16 times, and in each time, it trains using the data of

15 participants and tests the classiﬁer using the data

of the left participant.

Table 5: Training and testing times for the candidate classi-

ﬁcation algorithms.

Algorithm Training Time (sec) Classiﬁcation Time (sec)

SVM (Linear) 15.8 13.3

SVM (Polynomial) 15.3 12.7

Decision Tree 0.9 2.9

Naive Bayes 0.1 3.0

LDA 2.0 5.8

QDA 0.4 6.2

KNN (K = 6) 0.4 6.1

WKNN (K = 5) 0.4 6.0

Table 5 shows the accumulative time spent in

both the training and classiﬁcation phases. The re-

sults show that naive Bayes has the best training time

among all of them. While for classiﬁcation, it is the

second fastest algorithm after decision trees, however,

the time difference between them is almost negligible.

Given the accuracy results stated in Table 4, naive

Bayes outperforms all other candidates in terms of

both recognition accuracy and efﬁciency.

5.3 Sampling Frequency

The sampling frequency has a direct impact on the

system’s resources. This means lowering the sam-

pling frequency reduces the amount of operations and

computations done in the feature extraction stage, and

decreases the memory usage of the system. The data

collected throughout the experiment was sampled at

30 Hz which is relatively high. Thus, lower frequen-

cies, achieved by downsampling the collected data,

can be evaluated to determine their recognition accu-

racies.

Figure 2 presents the obtained results, showing the

average accuracies resulting from using naive Bayes

on the downsampled datasets (recomputed for each

frequency). Moreover, it shows that the best fre-

quency is in the ﬁeld of 10 Hz, scoring an average

accuracy of 88.8%, i.e., 0.6 percentage point decrease

when compared to the accuracy obtained using 30 Hz.

5.4 Watch Accuracy

In order to determine the improvement the watch

brings to the recognition system, the system’s accu-

Frequency (Hz)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Average Accuracy (%)

Figure 2: The impact of varying the sampling frequency on

the average accuracy (using naive Bayes).

racy is evaluated using the phone only, and compared

to the one obtained with both. Here, the same pro-

cedure used in the ﬁrst evaluation will be reapplied,

where the forward feature selection will be used to

compute the best performing subset using only the

phone’s features. Table 6 compares the resulting av-

erage accuracies to the ones obtained in the ﬁrst eval-

uation, showing the obtained improvements in the av-

erage recognition accuracy.

Table 6: The average accuracies of the classiﬁcation algo-

rithms with and without using the smartwatch.

Algorithm

Average Accuracy (%)

Phone & Watch

Average Accuracy (%)

Phone Only

SVM (Linear) 84.6 74.8

SVM (Polynomial) 87.9 81.1

Decision Tree 86.3 78.4

Naive Bayes 89.4 82.1

LDA 87.9 82.7

QDA 89.1 85.6

KNN (K = 6) 89.3 85.4

WKNN (K = 5) 89.3 85.4

The highest average accuracy (82.1%) is achieved

by using naive Bayes, where the individual LOPOCV

accuracies are 75.7%, 81.3%, 82.8%, 85.1%, and

85.5% for each dataset respectively. Comparing these

dataset accuracies with the ones obtained using naive

Bayes from the ﬁrst evaluation, it is clear that the

adding the watch to the recognition system improves

the accuracy with at least six percentage point.

6 CONCLUSION

In this paper, a platform to combine sensors of smart-

phones and smartwatches to classify various human

activities was proposed. It recognizes activities in

real-time Moreover, this approach is light-weight,

computationally inexpensive, and able to run on hand-

held devices.

The results showed that there is no clear winner,

ICAART 2016 - 8th International Conference on Agents and Artiﬁcial Intelligence

492

but naive Bayes performs best in our experiment in

both the classiﬁcation accuracy and efﬁciency. The

overall accuracy lies between 84.6% and 89.4%, at

which the differences are negligible. Thus, this plat-

form is able to recognize various human activities.

However, all of the tested classiﬁers confused walk-

ing and using the stairs activities.

The second conclusion is that adding the smart-

watch’s sensor data to the recognition system im-

proves it’s accuracy with at least six percentage point.

Finally, it is computations that the best sampling

frequency is in the ﬁeld of 10 Hz.

Some questions still require to be answered. Most

important is the conducting of larger experiments

with more people in order to perform more robust

evaluation to clearify if indeed one method is better

than the other, or whether, any off-the-shelf method

can do well in this classiﬁcation task. This work

could be furhter extended by incorporating more sen-

sors (e.g. heart rate sensor), recognizing high-level

activities (e.g. shopping or eating dinner) or extrapo-

lating these trained classiﬁers to other people.

ACKNOWLEDGEMENTS

This work was sponsored by Federal Ministry of Ed-

ucation and Research (grant 01IS12050).

REFERENCES

Banos, O., Galvez, J.-M., Damas, M., Pomares, H., and Ro-

jas, I. (2014). Window size impact in human activity

recognition. Sensors, 14(4):6474–6499.

Bao, L. and Intille, S. S. (2004). Activity recognition from

user-annotated acceleration data. In Pervasive com-

puting, pages 1–17. Springer.

Bauman, A., Phongsavan, P., Schoeppe, S., and Owen, N.

(2006). Physical activity measurement-a primer for

health promotion. Promotion & Education, 13(2):92–

103.

Bayat, A., Pomplun, M., and Tran, D. A. (2014). A study

on human activity recognition using accelerometer

data from smartphones. Procedia Computer Science,

34:450–457.

Brezmes, T., Gorricho, J.-L., and Cotrina, J. (2009). Ac-

tivity recognition from accelerometer data on a mo-

bile phone. In Distributed computing, artiﬁcial intel-

ligence, bioinformatics, soft computing, and ambient

assisted living, pages 796–799. Springer.

da Costa Cachucho, R. E. G., Moreira, J. M., and Gama, J.

(2011). Activity recognition using, smartphone based,

accelerometer sensors.

Gyllensten, I. C. (2010). Physical activity recognition in

daily life using a triaxial accelerometer. PhD thesis,

KTH Royal Institute of Technology.

Krishnan, N. C., Colbry, D., Juillard, C., and Panchanathan,

S. (2008). Real time human activity recognition us-

ing tri-axial accelerometers. In Sensors, signals and

information processing workshop.

Lee, M. (2009). Physical activity recognition using a single

tri-axis accelerometer. In Proceedings of the world

congress on engineering and computer science, vol-

ume 1.

Lester, J., Choudhury, T., and Borriello, G. (2006). A prac-

tical approach to recognizing physical activities. In

Pervasive Computing, pages 1–16. Springer.

Long, X., Yin, B., and Aarts, R. M. (2009). Single-

accelerometer-based daily physical activity classiﬁca-

tion. In Engineering in Medicine and Biology Society,

2009. EMBC 2009. Annual International Conference

of the IEEE, pages 6107–6110. IEEE.

Parkka, J., Ermes, M., Korpipaa, P., Mantyjarvi, J., Peltola,

J., and Korhonen, I. (2006). Activity classiﬁcation

using realistic data from wearable sensors. Informa-

tion Technology in Biomedicine, IEEE Transactions

on, 10(1):119–128.

Preece, S. J., Goulermas, J. Y., Kenney, L. P., Howard, D.,

Meijer, K., and Crompton, R. (2009). Activity iden-

tiﬁcation using body-mounted sensors - a review of

classiﬁcation techniques. Physiological measurement.

Ravi, N., Dandekar, N., Mysore, P., and Littman, M. L.

(2005). Activity recognition from accelerometer data.

In AAAI, volume 5, pages 1541–1546.

Sharma, A., Purwar, A., Lee, Y.-D., Lee, Y.-S., and Chung,

W.-Y. (2008). Frequency based classiﬁcation of activ-

ities using accelerometer data. In Multisensor Fusion

and Integration for Intelligent Systems, 2008. MFI

2008. IEEE International Conference on, pages 150–

153. IEEE.

Shoaib, M., Scholten, H., and Havinga, P. J. (2013). To-

wards physical activity recognition using smartphone

sensors. In Ubiquitous Intelligence and Computing,

2013 IEEE 10th International Conference on and 10th

International Conference on Autonomic and Trusted

Computing (UIC/ATC), pages 80–87. IEEE.

Subramanya, A., Raj, A., Bilmes, J. A., and Fox, D. (2012).

Recognizing activities and spatial context using wear-

able sensors. arXiv preprint arXiv:1206.6869.

Telgarsky, R. (2013). Dominant frequency extraction. arXiv

preprint arXiv:1306.0103.

Thiemjarus, S. (2010). A device-orientation independent

method for activity recognition. In Body Sensor

Networks (BSN), 2010 International Conference on,

pages 19–23. IEEE.

Yang, J. (2009). Toward physical activity diary: motion

recognition using simple acceleration features with

mobile phones. In Proceedings of the 1st interna-

tional workshop on Interactive multimedia for con-

sumer electronics, pages 1–10. ACM.

Human Activity Recognition - Using Sensor Data of Smartphones and Smartwatches

493