Department of Electrical Engineering
Permanent URI for this communityhttps://repository.cuilahore.edu.pk/handle/123456789/18
Browse
3 results
Search Results
Item Adaptive Nonlinear Gain Approximation of a Robust Nonlinear Controller to Handle Uncertain Disturbances for UAV Systems Using Neural Networks(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2025) Waqas Ahmad; CIIT/FA23-REE-003/LHR; Dr. Mirza Tariq Hamayun; LHR TP 10045The quadrotor unmanned aerial vehicle (UAV) has been a part of scientific studies owing to its applications in military, agriculture, infrastructure and construction, and environmental monitoring and conservation. However, its control design is challenging due to its nonlinear underactuated dynamics, external disturbances, parametric uncertainties, and the aerodynamic variations encountered during flight operations. Sliding mode control (SMC) is the most widely used control technique for handling nonlinear dynamics in the presence of external disturbances and modeling uncertainties. However, conventional SMC requires a priori knowledge of the upper bound on the overall disturbance at all times. Usually, these upper bounds on disturbances are derived from past data, environmental conditions, and empirical modeling; therefore, a conservative bound is used in SMC design, leading to high control effort and draining more energy from the system. To address this issue, adaptive sliding mode control (ASMC) strategies based on a neural network (NN) are presented in this dissertation. Neural networks are recognized as the leading approach to intelligent computation and are highly effective in handling nonlinearity, fault tolerance, adaptation, and continuous online learning. In this research, NN-based ASMC strategies are presented for a quadrotor UAV to improve flight performance by online estimation of adaptive modulation gain in response to variations in disturbance magnitude. Different neural network schemes are employed, including feed-forward neural networks (FNNs) using backpropagation (BP) and Levenberg-Marquardt (LM) algorithms for optimization, and a radial basis function neural network (RBFNN) with a Gaussian activation function, to adaptively estimate the modulation gains. Furthermore, the stability analysis of the proposed controllers is proven using Lyapunov theory.Item Implementation of Machine Learning Models for Predicting Hydrogen Production from Renewable Energy(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2025) Adnan Ayub; CIIT/FA23-REE-002/LHR; Dr. Muhammad Yaqoob Javed; LHR TP 10044The world continues to be largely reliant on fossil fuels, such as coal, oil and natural gas, which are major contributors of greenhouse gases, air pollution and climate change caused by the release of gases like CO2, NO o and SO2. Even though renewable energy sources such as solar and wind provide a cleaner substitute, their nature as intermittent and weather-dependent sources become a big problem in terms of large scale and long-term energy storage making traditional battery systems economically impractical. Consequently, the utilization of the surplus renewable energy through the process of water electrolysis to create green hydrogen has become a viable and alternative way of storing energy in the long term. This work suggests a two-stage machine learning-based predictive model of hydrogen production when using renewable energy, based on real-world working data of a 40.5 MW grid-connected photovoltaic (PV) power facility. The initial step involves predicting photovoltaic power output from meteorological variables using multiple regression and deep learning models, such as Support Vector Regression (SVR), Random Forest, Decision Tree, and Gated Recurrent Units (GRU). The second stage is to incorporate the predicted solar energy into a hybrid electrochemical model of hydrogen production, and the same machine learning models serve as data-driven correction models to more effectively predict hydrogen yield. It is a two-stage method that integrates both physical modelling and machine learning to model nonlinear system behaviour and real-world losses of operation. The findings show that SVR was always better in both phases than the other models with an R 2 value of 0.965 in photovoltaic power prediction and 0.968 in hydrogen production prediction. The given framework minimized the mistake in the production of hydrogen annually to about 3 percent, which is much better than theoretical models and deep learning alternatives. Moreover, a Sobol based global sensitivity analysis revealed that Global Horizontal Irradiance (GHI) made the greatest contribution to the uncertainty in the hydrogen production process, then AC power output and temperature at the module. The outcomes of these studies support the premise that the two stages proposed framework is a viable, precise, and large-scale solution to real-time forecasting, optimization of the system, and successful grid integration of green hydrogen system.Item Energy Aware Path Planning and Guidance for Non- Holonomic Robots in a Manufacturing Workshop(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2025) Rimsha Rehman; CIIT/FA23-REE-005/LHR; Dr. Mujtaba Jaffery; LHR TP 10046Automated Guided Vehicles (AGVs) are now a core part of modern manufacturing workshops, enabling fast, reliable, and flexible material movement while reducing dependence on manual handling. Despite their importance, most existing path planning approaches still focus mainly on minimizing distance or travel time. These strategies overlook a crucial factor, energy consumption, which directly affects operational cost, battery health, sustainability goals, and overall system efficiency. Addressing this limitation, this thesis presents an Energy Aware Path Planning (EAPP) framework specifically designed for single-load, non-holonomic AGVs used in structured workshop environments. The proposed framework models the workshop layout as an undirected graph, where nodes represent workstation points or intersections and edges represent feasible routes that comply with AGV kinematic and turning constraints. Unlike traditional planners, our approach integrates a physics-based energy model into the A* algorithm, allowing each edge to be evaluated not only by its geometric length but also by its expected energy usage, considering acceleration, deceleration, rolling resistance, turning angles, and standby power. This enables the AGV to prioritize paths with fewer turns, even when they are slightly longer, ultimately reducing total energy consumption. The method is implemented and validated using MATLAB simulations, where both distance-based A* and the proposed energy-aware A* are compared under realistic AGV parameters. The results show that the EAPP framework significantly lowers energy consumption and maintains competitive travel times, while still producing feasible and safe navigation paths within constrained workshop networks. By directly embedding energy considerations into the path planning process, this work contributes to more sustainable intralogistics and provides a practical, scalable solution suitable for real-time AGV navigation and industrial deployment.