Development of a prototype flexible skin-temperature sensor with printed electronics, an ESP32 MCU and a user-friendly wireless interface. The device was implemented using screen-printing technology with thermochromic inks, carbon, and silver materials, designed as a preventive tool for heatstroke and skin cancer.

This robotic vehicle uses sensors for autonomous navigation, flowerpot detection, and soil moisture monitoring, activating the plant watering system when needed. An LCD displays real-time moisture levels, and a Bluetooth mobile app allows manual control for the robot's movement and plant watering. Combining autonomous and manual modes, the system offers a flexible and efficient solution for automated plant care.

This project involves developing an object detection system using Arduino and Simulink (MATLAB). It integrates hardware and software to detect objects within 10 cm, controlling a servo motor and a LED to execute specific actions. This system offers hands-on experience in hardware programming and serves as a foundation for security applications, showcasing its adaptability for real-world use.

This project explores the use of Multilayer Perceptrons (MLPs) to predict CPU performance based on system characteristics. Using MATLAB programming and the MATLAB's Neural Network Training tool, the study highlights the advantages of MLPs in capturing non-linear relationships compared to linear models. Key findings reveal that increased RAM consistently improves performance, while cache memory effects are more complex. This study talks about backpropagation through training, validation, and testing of the model and the practical applications of MLPs in performance prediction.

This project involves the design and implementation of a spam email detection system using a Multilayer Perceptron (MLP) neural network. By analyzing email characteristics, the system classifies emails as spam or not spam, achieving validation and testing accuracies of 92.4% and 91.8%, respectively. The study demonstrates the effectiveness of MLPs in Pattern Recognition and the importance of robust data preprocessing and evaluation techniques.

This project analyses the effects of arginine and urea on bacterial growth using MATLAB-generated data and statistical tests, including t-tests, a Rank-Sum test, and regression analysis. The study found significant differences in growth, with antibiotics inhibiting bacterial proliferation. Power analysis emphasized the importance of larger sample sizes for sensitivity. The regression model offered predictive insights into bacterial growth trends, showcasing the importance of statistical methods in optimizing experimental outcomes.

This project investigates blood viscosity characterisation using non-Newtonian models and capillary rheometry data. It demonstrates how haematocrit levels affect viscosity and yield stress, with the Casson model and the Weissenberg-Rabinowitsch (WR) computational approach demonstrating the highest accuracy at higher concentrations, due to the model’s ability to capture shear-thinning behaviour. This study highlights the practical application of advanced rheological models in understanding blood flow dynamics, with implications for medical research and clinical diagnostics.

This project analyses human gait through kinematic and kinetic methods using MATLAB and 3D motion capture data. Ground reaction forces indicated asymmetric loading phases during foot contact. Transformation matrices and Euler rotations calculated intersegmental hip angles, showing cyclic patterns typical of gait mechanics. Muscle force contributions were also estimated by solving moment balance equations, identifying the gluteus maximus as the least involved during the task. This study highlights the importance of accurate modelling and data interpretation for biomechanical applications in predicting movement patterns and muscle contributions for gait assessment and rehabilitation.