Posts by Collection

portfolio

Clardia - Prediction of Diabetes Using Non Invasive Photoplethysmography (PPG) Measurements & Physiological Characteristics.

Photoplethysmography (PPG) can be identified as a non invasive, inexpensive, optic technique which measures the blood volume changes in blood vessels through which oxygen saturation, blood pressure, cardiac output could be measured. In recent researches it has been identified that PPG is a promising technique towards early screening of diseases as the PPG waveform possess significant information embedded within.

Flisma - A Wearable Device to Predict Injuries Due to Throwing

The modern sports arena in comparison to early days have become intensely competitive in nature, which has enhanced the focused on integrating technology towards sports enhancement. The amalgamation of technology with sports could be mainly identified in the areas of sports performance enhancement & sports injury prevention. This project focuses on using existing technologies and concepts to develop a wearable system which could be used for early injury prediction of athletes.

Clardia - A Comprehensive Family Health Assistant

Clardia is a Comprehensive Family Health Assistant which can extract health parameters such as the Weight, Body Fat, Heart Rate, Blood Oxygen Levels, Temperature and the PPG Signal of the user. Each family member is just required to stand on top of the Clardia device for 20 seconds. The obtained parameters are thus analyzed to identify diseases symptoms in advance and identify the possibility of potential diseases.

Analyzing Skewed Distributions

We use different data analyzing techniques to extract meaningful information. These techniques could vary from statistical methods, machine learning methods etc. In some cases we assume the normal distribution of data but in reality the data we obtain may not possess these ideal characteristics that we assume. In such cases some sought of a transformation is required to convert the real data to a normal distribution so that we could carry out the analysis. The practical distributions could be in the forms such as Right skewed, Left skewed , or with a couple of peaks.This article is mainly focussing on skewed distributions, and methods to convert them to a normal distribution.

Gym Tracker – An IoT weight scale to record the weight of Gym User’s

Gym Tracker, as the name suggests is a device capable of recording the weight of the users in order to identify the weight changes based on the requirements of the user. The system is able to identify the user via the RFID in the Gym Membership card and thus record the value of the particular user and communicate through wifi to record the data in the cloud server. The user is able to track the movement of his weight through the iWeight application developed.

Rampage – Not another Conventional Line Following Robot

The task was set to design an Analog Line Following Robot without the use of any Micro controllers. This was a quite straight forward task which requires the use of a closed loop feedback from line sensors (usually a combination of a LED/LDR), which is used by the Controller which includes a comparator. But me and my team thought of a different method to achieve the required task by thinking of a new design.

publications

Identifying the Optimum Region of the Human Sole to Extract the PPG Signal for Pulse Rate Estimation

Published in Proceedings of the 9th International Conference on Signal Processing Systems, 2017

Photoplethysmography (PPG) can be identified as a non invasive technique which measures the blood volume changes in the blood vessels, through which key health indicators could be identified and used for abnormality detection. PPG has been integrated in numerous modern applications. This research is conducted in order to evaluate the possibility of applying PPG to the human sole and identifying the optimum regions of measurement. The research calculates the normalized cross correlation between the PPG extracted from the finger and different regions of the sole, and develops an algorithm for Pulse Rate estimation.

Recommended citation: Hettiarachchi, Chirath, Buddhishan Manamperi, and Dilshan Uthpala. "Identifying the Optimum Region of the Human Sole to Extract the PPG Signal for Pulse Rate Estimation." Proceedings of the 9th International Conference on Signal Processing Systems. ACM, 2017. http://chirathyh.github.io/files/paper1.pdf

A Machine Learning Approach to Predict Diabetes Using Short Recorded Photoplethysmography & Physiological Characteristics.

Published in 17th Conference on Artificial Intelligence in Medicine (AIME) - Poland 2019, 2019

Diabetes is a global epidemic, which leads to severe complications such as heart disease, limb amputations and blindness, mainly occurring due to the inability of early detection. Photoplethysmography (PPG) signals have been used as a non-invasive approach to predict dia- betes. However, current methods use long, continuous signals collected in a clinical setting. This study focuses on predicting Type 2 Diabetes from short (~2.1s) PPG signals extracted from smart devices, and readily available physiological data such as age, gender, weight and height. As this kind of PPG signals can be easily extracted using mobile phones or smart wearable technology, the user can get an initial prediction without entering a medical facility. Through the analysis of morphological features related to the PPG waveform and its derivatives, we identify features related to Type 2 Diabetes and establish the feasibility of predicting Type 2 Diabetes from short PPG signals. We cross validated several classification models based on the selected set of features to predict Type 2 Diabetes, where Linear Discriminant Analysis (LDA) achieved the highest area under the ROC curve of 79%. The successful practical implementation of the proposed system would enable people to screen themselves conveniently using their smart devices to identify the potential risk of Type 2 Diabetes and thus avoid austere complications of late detection.

Recommended citation: Hettiarachchi, Chirath, and Charith Chitraranjan. "A Machine Learning Approach to Predict Diabetes Using Short Recorded Photoplethysmography and Physiological Characteristics." Conference on Artificial Intelligence in Medicine in Europe. Springer, Cham, 2019. http://chirathyh.github.io/files/AIME_2019_paper_89.pdf

A Wearable System to Analyze the Human Arm for Predicting Injuries Due to Throwing

Published in 41st Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) - Berlin 2019, 2019

There is limited understanding on factors that contribute to throwing related injuries that frequently occur in sports such as Baseball, Cricket and Javelin throwing. This preliminary study focuses on the development of a real time wearable system focusing on extracting key parameters related to potential upper arm injuries associated with the throwing action in the game of Cricket. A wearable system is developed to analyze Electromyography (EMG) signals for detecting muscle activity and Inertial Measurement Unit (IMU) data for monitoring the arm motion. The extracted parameters are then used for analysis, focusing on detecting established indicators of potential injuries. Additionally, an unsupervised learning algorithm is developed towards identifying novel relationships indicating potential injuries.

Recommended citation: Hettiarachchi, Chirath, et al. "A Wearable System to Analyze the Human Arm for Predicting Injuries Due to Throwing." 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). IEEE, 2019. http://chirathyh.github.io/files/EMBC19_0212_FI.pdf