Processing of Accelerometer Signals to Provide Weightlifting Insights

This report delves into the processing of accelerometer signals to gain insights into efficient weightlifting techniques, focusing on the Australian context. The study begins with an introduction to accelerometers and their application in measuring weightlifting performance, followed by a discussion of research aims, objectives, and questions. A comprehensive literature review explores accelerometer sensors, efficiency in weightlifting, and signal analysis methods. The research methodology outlines data collection techniques, including the use of a 3-axis accelerometer and a 3D camera, along with signal processing methods like Naïve Bayes Classifier and Hidden Markov Model, implemented using MATLAB. The report details the formulas used for acceleration responses and discusses data collection techniques, including primary and secondary sources. The report also includes a Gantt chart outlining the project timeline and concludes with a list of references, providing a structured approach to understanding and analyzing accelerometer data for improved weightlifting performance.

Running head: PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS
ON EFFICIENT WEIGHTLIFTING
Processing of accelerometer signals to provide insights on efficient weightlifting
Name of the student
Name of the University
Author note

1
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
Abstract
This paper discusses about the analysis of accelerometer measurements on weightlifting. The
chosen country is Australia. This paper includes a discussion of the basic requirements. This
paper has also discussed about the methods that will be utilized in making a successful research.
Moreover, the timeline to be followed for this paper is also listed and depicted in the form of a
Gantt chart. This paper has also utilized and discussed about the associated literature related to
the topic. These will be used in considering the research of the topic.

2
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
Table of Contents
1.0 Introduction................................................................................................................................4
1.1 Background of the study........................................................................................................4
1.2 Research aims and objectives................................................................................................4
1.2.1 Aims....................................................................................................................................4
1.2.2 Objectives...........................................................................................................................4
1.3 Research questions.................................................................................................................5
2.0 Literature review........................................................................................................................5
2.1 Accelerometer sensors...........................................................................................................5
2.2 Efficiency in weightlifting.....................................................................................................7
2.3 Analysis of signals in weightlifting.......................................................................................9
3.0 Research methodology...............................................................................................................9
3.1 Introduction............................................................................................................................9
3.2 Acceleration responses for weightlifting.............................................................................10
3.3 Algorithm of recognition.....................................................................................................10
3.4 Data collection techniques...................................................................................................11
3.5 Sampling technique.............................................................................................................11
3.6 Data analysis technique.......................................................................................................11
3.7 Gantt chart...........................................................................................................................12
4.0 References................................................................................................................................14

3
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING

4
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
1.0 Introduction
In this section of the research paper, the main focus will be on the background of the
research, aims and objectives of the paper and the research questions.
1.1 Background of the study
An accelerometer is considered to be a device that can measure the acceleration of a body
according to its own state of reference. The use of accelerometer is widely being adopted in
various industrial based adoptions. In addition, the use of accelerometer in the measurement of
weightlifting based performances is also a trend (Kalgan, Bahl & Kumar, 2015). This helps in
aiding as a tool for measuring the metrics of weightlifting. Moreover, this also gives a general
idea to the coach and the weight lifter about the necessary works that should be done for the
process.
1.2 Research aims and objectives
1.2.1 Aims
The aim of this paper is to evaluate the processing of an accelerometer signal which will
be used to get insights on weightlifting.
1.2.2 Objectives
The objectives of this paper are:
 To obtain insights from accelerometer signals in weightlifting
 To estimate velocity and displacement during the weight lifting cycle
 To successfully evaluate the insights gained from the analysis and get measured from
these estimates

5
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
1.3 Research questions
The research questions of this paper are:
1. Can actionable insights be consistently obtained from accelerometer measurements
during weight lifting?
2. What type of data processing will be needed to extract robust estimates of velocity and
displacement during the weight lifting cycle?
3. What insights can be gained from these estimates and are there specific measures derived
from these estimates that could provide actionable information?
2.0 Literature review
In this section, various articles and scholarly papers will be evaluated for making the
review. This section will also include theoretical findings of the associated literature such that
the evaluation can be done effectively.
2.1 Accelerometer sensors
The uses of sensors are mainly adopted for detecting, analyzing and recording the
physical parameters that are difficult to analyze using conventional instruments. These are
mainly done by converting analog to digital signals which are then sampled and analyzed. The
main involvement of a sensor is the conversion of physical parameters to electronic signals
(Struber et al., 2015). The accelerometer sensors are used for measuring the acceleration of any
body by keeping its state of rest as reference. The most common method of utilization of an
accelerometer is the airbag deployment in case of automobiles. The criteria set for its use is that
these are deployed when the acceleration of the vehicle exceeds from 30 to 50 dyne. In case of
this, an accident is usually termed which in turn leads to the deployment of airbags. The use of

6
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
an accelerometer is also involved with gyroscopes for guidance based mechanisms. These are
mainly used in small guidance units in rockets and aircrafts.
For the purpose of weightlifting, the accelerometer involved is a 3 axis based analog
device with a low profile capacitive action (Ong et al., 2018). The structure of the device
involves using a semi-conductor material which is modeled as a set of beams which are attached
to a movable central mass between fixed beams. When acceleration is sensed, the movable
beams moves from their axis. In this diagram, there are two variable amount of capacitance on
both of the moving side. The changes due to this on the beams are responsible for incurring
changes in the output voltage which in turn is responsible for the calculation of acceleration. The
selective sensitivity range of the accelerometer device is considered to range from (+-) 1.5g to
(+-) 6g. The reference node of the sensor is termed to be 0g which will give a logic high output
when the accelerometer is not experiencing any kind of acceleration. For this reason, the sleep
pin is pulled to a low power mode. The input and output node of the device will be used in
association to a microcontroller for getting accurate amounts of data (Krüger et al., 2017). Using
the tilt angle, the accelerometer will get data. The main reason this device was chosen is that the
simplicity and ease of use of this device helps in better handling of the equipments needed.
Figure 1: 3- axis accelerometer model
(Source: Krüger et al., 2017, pp 400)

7
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
2.2 Efficiency in weightlifting
Acceleration of a body always involves the influence of the Earth’s gravitational field.
The body’s own acceleration and the gravitational field of Earth are considered by an
accelerometer. The use of gravity is helpful in measuring the tilt sensor of the vehicle. This is
usually done by negating the effects of gravity which then forms a three axis based sensors. This
in turn is utilized for navigation purposes (Attal et al., 2015). The accelerometer based sensors
can also be used to change the status of any device. For example, the use of an accelerometer can
be adopted for a washing machine where changes in the acceleration of the water can be made to
change the functioning of it. By utilizing a two axis based accelerometer, vibrations help in
detecting any out-of-balance load.
For reaching the goals of this paper, various weightlifting exercises will be considered for
utilizing accelerometer signal analysis. These exercises are most commonly used in case of
weightlifting requirements and the need to track the signals will be presented here. Based on the
target muscle impact, these exercises will be analyzed for this paper (Rawat, 2016). For example,
for arm based weightlifting exercises, people often use biceps curl or triceps curl. The table
below shows the various exercises that can be utilized in adopting weightlifting exercises.
S.no Exercise Muscle groups Body part Posture
1 Biceps curl Biceps Arms Standing/Sitting
2 Triceps curl Triceps Standing/Sitting
3 Bench press Chest Upper body Lying
4 Flye Lying
5 Bent-over row Upper back Standing
6 Lateral raise Shoulder Standing

8
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
7 Overhead dumbbell press Standing/Sitting
8 Dead-lift Quadriceps Lower body Standing
Table 1: Representative and commonly performed exercises for each muscle group
( Source: Shoaib et al., 2016)
For tracking the exercises from these weightlifting movements, the use of an
accelerometer will be emphasized. This in turn will help in incorporating acceleration data with
machine learning techniques. As the weights are involved in movement and rotation, a three axis
accelerometer is thus utilized. Additionally, it can also be used to differentiate the posture of the
people with respect to standing or sitting. Due to the trade-off among cost, nature of
weightlifting and cost, the use of a three axis accelerometer is adopted in this case. In addition, as
human motion is restricted to a few motions in case of weightlifting exercises, the use of this
three axis based accelerometer is utilized which will in turn help in monitoring of the parameters
in an effective way.
Figure 2: Illustrations of (a) overhead dumbbell press, (b) bench press, (c) bent-over row,
(d) lateral raise, (e) bicep curls, and (f) dead-lift
(Source: Chang, Chen & Canny, 2014, pp 21)

9
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
2.3 Analysis of signals in weightlifting
For this research, the device must be able to analyze the effective weight that is being
used in exercising. For this reasons, the free weights can be enabled by attaching RFIDs to the
weights and attaching RFID readers in the gloves of a weightlifter. In addition, the utilization of
RFID mapping can also help in tracking the signals from the devices (Chang, Chen & Canny,
2014). In this paper, there are various parameters which will be considered for denoting the
acceleration of the body. For the three axis accelerometer, the value of the three axes is to be
analyzed by considering all the parameters associated.
Figure 3: The direction of three axes on the accelerometer and the posture clip
(Source: Chen, Jafari & Kehtarnavaz, 2015, pp 55)
2.4 3D camera
For the use of the 3-axis accelerometer, the adoption of the 3-dimensional camera will be
utilized. This will help in capturing the analog signals from the sources which will then be
converted into its digital aspects. This will not only help in efficient conversion but will also help
in making a successful analysis of this paper concerned. The associated co-ordinates of this type
of camera is the X, Y and Z axis. This will be utilized in making the analysis by using the
equations listed later in this paper.

10
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
3.0 Research methodology
3.1 Introduction
The collection of information and analyzing them is included in this section of the paper.
The information and data from sources are arranged in a logical and sequential manner. The
utilization of hypothetical techniques and numerical values will be the sole intent of this section
in the paper. This paper has utilized the positivism philosophy due to the fact that the utilization
of scientific approaches for collecting and analyzing data will be adopted in the analysis used for
this research. In this paper, the deductive approach is utilized for analyzing the relevant sources.
3.2 Acceleration responses for weightlifting
AccelerationX =(((accelX∗Vref )/1023)−accelXzeovolatge)/ accelSe nsitivity Equation 1
AccelerationY =(((accelY∗Vref )/1023)−accelYzeovolatge)/accelSensitivity Equation 2
AccelerationZ=(((accelZ∗Vref )/1023)−accelZzeovolatge)/accelSensitivity Equation 3
With reference to equation 1, accelX is considered to be the value from the ADC
converter. Here, Vref is considered to be the reference voltage which will be used to scale the
readings. Similarly, accelXzerovoltage is termed as the nominal voltage when stationary position
is implied. Lastly, accelSensitivity is termed as the voltage response for every unit per g-force. In
a similar manner, the other axes parameters are evaluated using these formulas.
3.3 Signal processing
For the use of this device, the adoption of Naïve Bayes Classifier and Hidden Markov
Model has been emphasized. The first requirement for any real time analysis involves filtering
out the noise from the sensor data. For this reason, this paper will utilize the adoption of

11
PROCESSING OF ACCELEROMETER SIGNALS TO PROVIDE INSIGHTS ON
EFFICIENT WEIGHTLIFTING
MATLAB which will help in filtering out the noise from a 3-axis accelerometer. The next step
will involve the detection of the sensor signal. This must be done for checking the threshold level
of the device so that an effective data collection is implied. This will also be done by utilizing the
MATLAB software. Any kind of raw or filtered signal from the sensor can be used. Lastly, the
utilization of a 3D camera will be adopted which will help in collection of the signals in the 3-
axes accelerometer.
3.4 Data collection techniques
Data collection is the process which involves the collection of data from various sources.
The data and information is then collected from various sources. There are two different types of
data collection techniques, the primary and secondary data collection. In case of the primary
methods, the information is collected from surveys and interviews while in case of secondary
data collection, the information is collected from online sources like journals, books and
websites. For this research paper, the researcher has utilized the primary data collection method.
3.5 Sampling technique
The number of devices used and the instrument involved in collecting the data is
discussed in this section of the paper. For this paper, the researcher has utilized the primary data
collection methods. This paper will analyze the signals from the device which will be considered
as the main instrument for this paper. For this paper, the three axis accelerometer will be used for
analyzing the different types of data.
3.6 Data analysis technique
The analysis of the collected data is involved in this section. The data that is collected is
then analyzed and presented in the research paper. The utilization of data analysis methods is