This project explores the development of an Energy Management System (EMS) designed to optimize power generation from wind turbines, battery storage, and load management using Reinforcement Learning (RL). By leveraging load profile and wind speed data, we create an RL agent that makes optimal decisions in a fluctuating energy landscape. The EMS incorporates key cost parameters, including USD 0.20 per kWh for imported energy, USD 0.05 per kWh for exported energy, and USD 0.10 per kWh for battery usage. The main objective of the agent is to maximize rewards by minimizing costs associated with energy consumption while improving the efficiency of both energy generation and storage.

Through extensive training and simulation, the RL agent adapts to varying conditions, effectively balancing energy supply and demand in response to changes in wind energy generation. Preliminary results indicate that the EMS not only enhances cost efficiency but also improves overall energy utilization. This demonstrates the viability of applying RL techniques in the management of renewable energy resources.

The findings of this research significantly contribute to the advancement of smart energy systems and the integration of sustainable energy sources, providing a framework for developing more efficient and resilient energy networks. By showcasing the potential of RL in optimizing energy management, this project paves the way for future innovations in renewable energy applications.

The robust control of quadcopters is crucial for maintaining stability and performance in dynamic and unpredictable environments. This paper investigates the effectiveness of two optimization techniques, Particle Swarm Optimization (PSO) and Grey Wolf Optimization (GWO), for tuning LQR-PID controllers specifically designed for a constrained quadcopter limited to rotational degrees of freedom. The objective is to enhance attitude stabilization and perform a comparative robustness analysis of these optimized controllers under various disturbance conditions.

LQR-PID controllers are designed for the quadcopter model using PSO and GWO to optimize the Q and R matrices of the LQR controller. Both algorithms aim to minimize a cost function based on the quadcopter’s attitude error and control effort. The optimized controllers are tested in a Simulink environment where disturbances such as wind, initial condition perturbations, and sudden impulse disturbances are introduced. Wind disturbances represent varying external forces, initial condition perturbations simulate small deviations from the expected starting state, and sudden impulse disturbances model unexpected sharp forces. These disturbance types were selected to reflect real-world operational challenges faced by quadcopters and are introduced by perturbing the feedback vector of the quadcopter’s control system.

The comparative analysis shows that while both PSO- and GWO-optimized controllers achieve effective attitude stabilization, they display different robustness characteristics. The PSO-optimized LQR-PID controller demonstrates better performance in terms of faster convergence and higher sensitivity to disturbances, whereas the GWO-optimized controller excels under extreme parameter variations.

This study contributes to the current state of the art by providing a detailed comparison of PSO and GWO for LQR-PID tuning in quadcopter attitude control. The results offer valuable insights for selecting the most suitable optimization method based on specific performance and robustness criteria, ultimately aiding in the development of more resilient and reliable quadcopter control systems.

This study introduces a sophisticated control framework for managing a pair of quadrotor drones in a leader-follower configuration, leveraging a combination of backstepping control optimized through Particle Swarm Optimization (PSO) and Proportional-Integral-Derivative (PID) control for maintaining inter-drone distance. The mathematical model of quadrotor is presented as a first step then the leader drone’s trajectory is governed by a backstepping approach, which is fine-tuned using PSO to achieve optimal performance. This optimization enhances the backstepping controller’s ability to manage complex trajectory tracking tasks, ensuring the leader drone follows its designated path with precision, robustness and minimal deviation. Simultaneously, the follower drone’s position relative to the leader is managed by a PID control system, specifically designed to maintain a constant distance between the two drones. The PID controller adjusts the follower drone’s position dynamically, responding in real-time to any variations in the leader’s trajectory and ensuring that the desired separation distance is consistently maintained. Simulation experiments validate the effectiveness of the proposed control strategy. Results demonstrate that the leader drone adheres to its trajectory with exceptional accuracy, facilitated by the PSO-optimized backstepping control. The follower drone, in turn, effectively maintains the desired distance from the leader, showcasing the robustness and adaptability of the PID control system. The proposed approach significantly improves both trajectory tracking and distance maintenance, providing a reliable solution for coordinated multi-drone operations. This strategy not only enhances precision in trajectory adherence but also ensures stable and robust distance control, making it highly effective for complex and dynamic flying scenarios.

The inverted pendulum is a classical benchmark in control theory, known for its inherent instability and nonlinearity, making it a challenging problem for control engineers. In this paper, we propose a novel optimized backstepping control approach to address the challenges of stabilizing the inverted pendulum while ensuring robust performance, fast response, and high precision. Backstepping, a recursive design methodology for nonlinear systems, is employed due to its ability to systematically handle the pendulum’s nonlinear dynamics. To further enhance the performance of the controller, an optimization technique is applied to fine-tune the backstepping parameters, focusing on achieving a balance between robustness, speed, and precision in the pendulum’s stabilization. The optimization process is designed to minimize the control error while maintaining stability under various disturbances and model uncertainties. Simulation results validate the effectiveness of the proposed approach. The optimized backstepping controller demonstrates superior performance compared to traditional control methods, particularly in terms of robustness to external disturbances and parameter variations, fast convergence to the desired state, and precise tracking of the pendulum’s upright position. Additionally, the system exhibits low overshoot and minimal steady-state error, making it well-suited for applications requiring high control accuracy. The results highlight the potential of this optimized backstepping methodology for controlling complex nonlinear systems, providing a robust, fast, and precise solution for stabilizing the inverted pendulum, even in the presence of disturbances.

Quantum-Dot Cellular Automata (QCA) emerges as a promising nanotechnology, offering a viable alternative to traditional technologies for high-performance, low-power computing at high operating speeds in a compact area. It is well known that parity generators and checkers are crucial components in processors and communication circuits. In line with recent trends in nanocircuits in electronics, this work designs a parity generator and checker using QCA, an emerging nanotechnology that is gaining popularity for nanocomputing tasks. This paper presents a dual-tasking QCA-based 3-bit parity generator and checker, implemented using an optimized modified majority voter (MMV)-based XOR gate in QCA circuitry. It is a dual-tasking circuit, as the single circuit can both generate and check parity. The efficient XOR gate design significantly enhances the reliability and importance of parity checker circuits in QCA, which is vital for error detection in communication systems. A full input–output-accessible,single-layer and scalable layout was designed and simulated using QCADesigner 2.0.3 without incorporating crossovers to minimize fabrication complexity. Detailed power dissipation analysis, conducted using two tools, QCADesigner-E and QCAPro, shows that the circuit consumes 23.86 meV, demonstrating 86% energy efficiency as well as 59% area efficiency compared to the latest reported QCA-based parity checkers. The proposed circuit exemplifies a significant advancement in nanocomputing, providing a scalable, energy-efficient solution for next-generation computing systems, further establishing QCA as a key technology for future low-power, high-density applications.

This paper deals with the development of a new 6-degree-of-freedom (DOF) visual servoing control law. The traditional visual servoing (VS) methods depend on geometric features to design the control law, which limits their versatility due to reliance on visual tracking algorithms. To address these limitations, direct visual servoing (DVS) approaches have been introduced, showing that the design of photometric visual servoing (PVS) can bypass geometric feature extraction by directly considering the luminance of all image pixels to control the robot. However, these methods are sensitive to illumination changes and partial occlusions, which make the control task non-robust. To overcome this, we develop a new direct visual servoing control approach based on a Sobel filter to enhance the precision of image information under changing lighting conditions by extracting image gradients. These gradients are then used to design the interaction matrix that allows real-time updates to the motion of the robot for adaptability and precision. Also, the proposed control scheme has been tested on the VISP platform and was compared to the classical photometric visual servoing in order to evaluate its efficiency in nominal and unfavorable conditions. Experimental results validate that the approach provides more performance and reliability under variable illumination conditions.

This study focuses on improving geolocation accuracy in satellite systems using passive sensors through relativistic error estimation. Geolocation, the determination of a target's position using signals from electromagnetic emitters, becomes increasingly challenging when relativistic effects are considered, especially in space-based systems. The system under study involves a passive receiver mounted on a satellite and an emitter located on Earth's surface. The primary goal is to integrate relativistic corrections—such as time dilation, changes in potential energy, and satellite orbit eccentricity—into traditional geolocation algorithms, which primarily rely on signal time delay.

To achieve this, a theoretical model is developed, which examines the relativistic contributions to position errors arising from the finite speed of light and the effects of the satellite relative motion. Using an analytical approach, the study evaluates how these relativistic factors influence geolocation accuracy. A comparison between a Newtonian model for time delay and a modified algorithm incorporating relativistic corrections is presented. The relativistic correction algorithm, implemented through adjustments in the software layer, does not require system-level changes, making it feasible for current satellite operations.

Simulation in worst case scenarios results demonstrate that while relativistic effects are often considered negligible in many applications, they can introduce significant errors, particularly in high-precision tracking scenarios or in systems with highly eccentric orbits. In conclusion, incorporating relativistic corrections in satellite-based geolocation algorithms can enhance the precision of passive sensors, offering valuable improvements in both Earth and deep-space missions.

Autonomous navigation in partially known or unknown environments, such as agricultural fields, poses significant challenges for mobile robots. The effective guidance of these robots is crucial for their successful operation in dynamic settings. Artificial Potential Fields (APFs) are widely employed for this purpose; however, they often lead to issues such as oscillations and local minima, which can hinder the performance. This study proposes an innovative optimization of the parameters of Artificial Potential Fields using a genetic algorithm (GA) to address these limitations. The GA fine-tunes the attractive and repulsive constants of the potential fields, significantly enhancing the navigation performance. Comprehensive simulations were conducted in a dynamic environment, incorporating various static and mobile obstacles to rigorously test the proposed method. The results demonstrate a significant improvement in the robot performance, highlighted by smoother trajectories, reduced collisions, and improved handling of dynamic obstacles. Specifically, the APF-GA method decreased the time to reach the goal from 18.8 to 16.1 seconds and the distance traveled from 7.61 to 6.43 meters. This integration of the genetic algorithm into the APF method not only enhances the smoothness of the trajectory but also increases the navigation safety in complex environments. These promising results have important implications for real-world applications, particularly in agriculture and logistics, paving the way for more efficient robotic systems.

Estimation technique is a very important tool in the control of electrical machines, especially in the control of induction motors. The technique is based on the concept of "observing" unobservable states of the system by a mathematical model. There are two kinds of observer techniques available in the literature: the open loop type and the close loop type (observer).

In our study we use Direct Torque Control (DTC)which is a control strategy that allows direct control of the motor’s torque and flux. This technique uses estimators or observers to calculate motor parameters, including rotor speed, torque, and stator flux.

A comparative study between the experimental implementation of an open loop observer and an Extended Kalman Filter (EKF) observer with Direct Torque Control (DTC) of an induction motor can be framed around several performance criteria. These criteria include accuracy, dynamic response, robustness, computational cost, noise sensitivity, and ease of implementation. The Open Loop Observer is a simpler observer, based on the mathematical model of the induction motor. It assumes perfect knowledge of the system and does not correct for measurement or model inaccuracies. Where Extended Kalman Filter (EKF) Observer is the more advanced observer that uses a stochastic approach, incorporating statistical models of noise. It estimates motor parameters by recursively updating based on measurement and system dynamics, correcting for errors and noise.

DC machine usage in Electric Vehicles (EVs) has been gaining a notable amount of focus. The speed of DC motor-driven wheels in an EV changes as it encounters disturbances like a reduction in tire air volume, or a corrugated and rugged surface on which it is driven. This continuous disturbance and variation in speed could result in the exertion of EV circuits, which can be fatal for passengers. Hence, a control method that could respond to the disturbance and give a signal to the motors of the wheels for automatic speed control is required. Thus, this paper proposes artificial neural network (ANN)-based control strategies for enhanced speed regulation in DC motor-driven electric vehicles. This paper highlights the focus on ANN control strategies in the context of DC motors of EVs. Different ANN architectures such as radial bias network (RBNN), probabilistic neural network (PNN), feed-forward network (FFNN), Elman network, NARX network, NAR network, and recurrent neural network (RNN) are implemented to design the gains of the PID (Proportional–Integral–Derivative) control loop of the DC motor. A thorough analysis concerning different activation functions, mean squared error, mean absolute error, and weight-bias functions is provided. The efficacy of all these methods is tested when the EV system is subjected to key disturbances, namely, step, ramp, sinusoidal, and chirp. System responses under all these test conditions for all the ANN architectures are drawn. A better ANN architecture to tune the PID controller is recommended based on these transient characteristics and disturbance rejection ability. From the results, it is observed that the performance of FFNN is superior to that of other ANNs due to its shorter rise time, less peak overshoot, lower delay time, and lower steady-state error. Thus the proposed work leverages the usefulness of ANNs to achieve more precise speed control, enhancing the overall performance of EVs.