This study introduces an adaptive Fourier linear combiner (FLC) based on a modified least mean kurtosis (LMK) algorithm in order to effectively process sinusoidal signals, which we call FLC-LMK algorithm. In the design procedure of the proposed FLC-LMK algorithm, the classical kurtosis-based cost function is first modified for only sinusoidal signal distributions instead of Gaussian. Then, the FLC-LMK algorithm is derived from the minimization of this cost function and thus updates the weight coefficients of the FLC structure so as to directly process sinusoidal signals. Moreover, in this study, the convergence in the mean of the proposed FLC-LMK algorithm is analysed in order to determine the lower and upper bounds of its step size parameter. The most important contributions of the use of the proposed algorithm in the FLC structure are that it increases the convergence rate, decreases the steady-state error level and also has a robust behaviour against sinusoidal signal distributions due to its modified cost function. The performance of the proposed FLC-LMK algorithm is evaluated on the synthetic and real-world pathological hand tremor data by comparing with that of the FLC based on the classical least mean square (LMS) (FLC-LMS) algorithm. The simulation results support the mentioned properties of the proposed FLC-LMK algorithm.
A method of designing adaptive controller for high performance active magnetic bearings (AMB) is represented. The approach combines two simple architectures often referred to as a Dynamic Feedback Linearization (DFL) and recent advances in adaptive control design techniques to form a new approach for AMB. The procedure enables the designer to explicitly define the desired closed loop dynamics. The result is a straight forward procedure that enables the design of a robust stabilizing adaptive controller that forces the system dynamics to the specified desired dynamics, despite disturbances, modelling uncertainties, and variations in AMB dynamics. This method also guarantees the robustness of both stability and system performance and enables the design for other unstable, magnetic levitation plants.
Batterychargers are the energy transmission part of electric vehicles between the grid (or electric energy source) and the vehicle. This energy transmission process has a great effect on the technology of electric vehicles. Charging a vehicle as fast as possible is necessary for an electric vehicle to compete with the internal combustion engine vehicles. The technology has been dealing with this process near past, while charging a car was taking a few hours, this time is reducing day after day. In this paper it is aimed to give some information about the current technologies of chargers(only the chargers with cable not wireless chargers) and it is also aimed to design a trial Simulink model for chargers. When the literature is considered, simulation samples of chargers are little so increasing these samples is important to improve the technology offast charging.
Development of intelligent care system for elder people have been investigated in recent years. In this study, to detect emergency situations for elder people, activity classification was aimed using on body sensor data. Multi-layer perceptron, radial basis function networks, k- nearest neighbor and support vector machines were used in classification. In feature selection process principal component analysis and ReliefF were used. Accuracy of classification was above 85% for every classifier and the best performance was acquired with 3-NN with 99.8% accuracy. When feature selection was applied 5- NN was showed the highest performance with 99.4%. This study shows that it is possible to develop remote care system by using sensors and classifiers for a more secure life for elder people.
This paper studies the problem of stabilization for a class of linear singular time-delay systems. First, a delay-dependent stability criterion is developed within the context of Lyapunov stability theory and linear matrix inequalities. The stability result is then extended to obtain a stabilizing state-feedback controller within the bilinear matrix inequalities (BMI) framework. Employing the cone complementary linearization approach, the synthesis problem can be resolved through a set of LMI conditions. Several numerical examples are presented to illustrate the application of the theoretical results.
Audio was mainly used for speech and speaker recognition before. Sound event detection (SED) is another field of audio recognition which is the recognition of sounds other than speech and music. If we recognize environmental sounds coming from hazardous events then we can use this for surveillance for security. Audio surveillance can be integrated into video surveillance systems for public security in cities, for surveillance of elderly people living alone and road surveillance etc. In this paper we developed deep neural network (DNN) models to recognize scream and traffic accident (car crash). Our model tests show that the developed models can be used in real applications.
Permanent Magnet Synchronous Machines (PMSMs) are getting popular in the automotive applications which are necessary for high power density in low volumes. Also they have a huge usage potential in other industrial applications like robotics, naval applications, space applications etc. In this paper, design parameters of a special PMSM are analyzed which has two independent winding set for motor and generator operations in a simple stator instead of two separate electrical machine. Important design parameters like skew, airgap length, slot-pole combination, magnet height-width which affect the output performance are investigated for multi tasked PMSM. Performance output values are analyzed for two operation situation of multitasked PMSM and optimum design values are determined. These values are obtained from Maxwell-Simplorer softwares where co-simulation analysis has been run.
In this study, we present a unified motion planner with low-level controller for continuous control of a differential drive mobile robot. Deep reinforcement agent takes 10 dimensional state vector as input and calculates each wheel's torque value as a 2 dimensional output vector. These torque values are fed into the dynamic model of the robot, and lastly steering commands are gathered. In previous studies, navigation problem solutions that uses deep - RL methods, have not been considered with agent's own dynamic constraints, but it has been done by only considering kinematic models. This is not reliable enough for real-world scenarios. In this paper, deep-RL based motion planning is performed by considering both kinematic and dynamic constraints. According to the simulations in a dynamic environment, the agent succesfully navigates through the intersection with 99.6% success rate.
This paper has for objective, the maximization of a one-body wave energy converter using model predictive control. Where two cases are tested. In the second case, a proposition of using the observer model to extract the control law in order to ameliorate control characteristics and reduce the costs of using sensors. The suggested approach is applied to our system and compared to the first case, which represent the classical approach, through simulation results.
In this study, an online-tuning method for derivative order term of Fractional PD and Filtered Fractional PI controllers is presented. For this purpose, closed-loop step response is divided to certain regions and a different tuning strategy is proposed for each region. These tuning strategies basically depend on the error between the system output and reference input. The strategy formulas are formed as linear equations arranged in terms of system error and system time constant. Simulations are performed to show the effectiveness of the proposed on-line tuning method on various systems.
This paper presents a mathematical modeling and simulation of a nonholonomic mobile robot for the low-level control regarding the parameter change using model-based design. The model-based design methodology has been a promising technique in automotive and home appliance industries to address and validate the algorithm development, code generation and its deployment on platform. To investigate the electrical and mechanical behavior of the mobile robot, nonlinear mathematical model is constructed and is added to the control loop with the estimator to identify the parameter change. The recursive least square algorithm is used to identify the inertia and the payload of the robot to adapt the dynamic behavior before applying control law. The performance of the model-based approach is investigated on the simulation environment considering the different payloads with DC motor's constraints to track the reference trajectory using PI controller.
Disturbance attenuation problem is considered as an important topic in control literature. This paper deals with the design of HOSIDF (Higher Order Sinusoidal Input Describing Functions) based Chebyshev structured compensator in order to increase the disturbance attenuation performance of the system involving actuator saturation. This study consists of the proposed compensator design in addition to ℋ∞ dynamic output feedback controller that already exists in the system. The simulation studies are carried out with an active suspension system which is known as a benchmark problem in control literature. The improvement in disturbance attenuation performance of the closed loop system involving HOSIDF-based compensator is illustrated with time-domain and harmonic plots.
Home automation systems give consumers access to control devices in their homes from a mobile phone or a tablet anywhere in the world and there are many products such as door locks, smart lights, garage doors, doorbell cameras, and smart switches with plugs in the home automation technology market. Smart thermostats are also one of the most important products of home automation system. HVAC systems account for over 60% of the energy consumed by buildings and this rate is expected to increase further in the future. For this reason, it is important to pay attention to the efficient use of HVAC systems. In this study, three different studies were performed on how much energy could be saved by using smart thermostats. When the temperature can be reduced during the night hours while consumers are sleeping, the energy savings for electricity and natural gas consumption were calculated as 20.72% and 17.7%, respectively. When consumers are away from the house at specific hours, they can save average 35.41% and 10.28% percent on electricity and natural gas consumption, respectively by turning down the thermostat or turning it off completely.
The role of a double talk detector (DTD) in acoustic echo cancelation (AEC) system is to detect the presence of near-end speech signal with the microphone signal and freeze the filter adaptation to avoid the adaptive algorithm divergence. This paper presents a DTD based on an enhanced Geigel algorithm where a modified form of decision variable is proposed. The aim is to improve the behavior of Geigel algorithm, evaluate performances of the proposed method, and compare it to conventional Geigel algorithm. Recursive Least Squares (RLS) algorithm is used in this case as an adaptive filter which it requires a good DTD due to its fast convergence and its sensitivity to double talk situations.
To recognize and automatically identify the identities of individuals, there are several biometric identification systems based on physiological and behavioral characteristics, in our work we are interested in face recognition, which is a recent biometric authentication technology. This technology offers a reasonable level of precision. In this paper, we propose a method of biometric recognition of a person by their face using the wavelet transform. For our application, we have opted for different types of wavelets in order to decompose the region of interest. The evaluation and judgment of each type in relation to the other is given by the calculation of the parameters Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR). The experimental results were performed using the FEI database.
The share of wind energy in electric energy generation increase day by day. The penetration of wind power plant in power systems has come to considerable size. On the other hand, power quality concerns are on rise. In this study, the data obtained by measurements of power quality parameters performed on a real power system are studied. The data are evaluated by the available legislations related on power quality
The nonlinear hysteretic behavior exists in many systems and makes their analysis and control an arduous task. Based on adaptive fuzzy synergetic control strategy, this paper deals with the control of a class of nonlinear systems preceded by an unknown backlash nonlinearity. Fuzzy logic systems are used to estimate the unknown nonlinear behaviors of the system, and a novel adaptive fuzzy controller is designed via synergetic control theory. Stability is proven, theoretically, in the sense of bounded closed-loop signals, and a tracking error converges to the origin. Simulation results validate the proposed approach and give an overview on the achieved tracking performances.
Autonomous robots are critical components of factories of futures. In this era, autonomous transfer vehicles are expected to play important role for flexible manufacturing. But the system should detect abnormal events itself. In this study, anomaly detection approach is proposed for autonomous transfer vehicles in the smart factories. Decision trees are used to detect stopping and slow down anomalies in internal transportation of the factories. The proposed approach is tested in simulation environment.
A two-cell power chopper system will be studied in this paper. The topology of this chopper is based on a combination of two cell switching interconnected via a flying capacitor. The system is a particular hybrid dynamical one which induces new and difficult control problems. In this paper, such problem is tackled by a new control concept based on Petri Nets modeling. The main advantage of this control is to use a discrete event algorithm for both current tracking and capacitor voltage balancing with the ability to drive directly the chopper switching components, by respecting the tolerance errors of load current and capacitor voltage. Simulation and experiment tests are carried out to verify the feasibility and effectiveness of the proposed control. The obtained results show that; the proposed controller presents good performances, in terms of both current tracking and voltage balancing compared to conventional existing controls according to the variation of the tolerance errors.
Due to the rapid increase in energy demand, the equipment used in the transmission and distribution network is subjected to thermal overloading. Additionally, electrical stress due to high voltage and contaminations from environmental/structural factors lead to the life of insulating material to decrease rapidly. Among the equipment used in power systems, power transformers are one of the most affected ones from these undesired and distorted conditions. Particularly for oil type transformers, it is quite difficult to provide energy continuity due to failures caused by insulation problems, serious economic losses occur and more importantly, there might be cases where life loss is experienced. In order to avoid from such this cases, the electrical and mechanical properties of the insulation material should be analyzed correctly and the behaviors against different stresses should also be examined. In this context, it is crucial to investigate the problem of insulation between windings, which encountered quite often and has an important place among transformer failure reasons. Along with creeping of spacer part between the windings, the windings get closer to each other and by breaking down of the insulation material around the conductor due to overheating cause failure. In this study, a test system is designed and realized to determine the creep characteristic of spacer component, which is used to provide oil flow between the windings and for easier cooling, under different thermal and mechanical stresses. In this designed test set up, the spacer is placed in transformer oil and exposed to both mechanical and thermal stresses whose amplitudes can be controlled. By this way, it is aimed to investigate the effects of various thermal and mechanical stresses on the creeping of the spacer part. The results of the tests show that the creep curves will help to make the power transformers' aging calculations and failure prediction algorithms more accurate.
