Francesco Cupertino

Compact Differential Evolution

This paper proposes the compact differential evolution (cDE) algorithm. cDE, like other compact evolutionary algorithms, does not process a population of solutions but its statistic description which evolves similarly to all the evolutionary algorithms. In addition, cDE employs the mutation and crossover typical of differential evolution (DE) thus reproducing its search logic. Unlike other compact evolutionary algorithms, in cDE, the survivor selection scheme of DE can be straightforwardly encoded. One important feature of the proposed cDE algorithm is the capability of efficiently performing an optimization process despite a limited memory requirement. This fact makes the cDE algorithm suitable for hardware contexts characterized by small computational power such as micro-controllers and commercial robots. In addition, due to its nature cDE uses an implicit randomization of the offspring generation which corrects and improves the DE search logic. An extensive numerical setup has been implemented in order to prove the viability of cDE and test its performance with respect to other modern compact evolutionary algorithms and state-of-the-art population-based DE algorithms. Test results show that cDE outperforms on a regular basis its corresponding population-based DE variant. Experiments have been repeated for four different mutation schemes. In addition cDE outperforms other modern compact algorithms and displays a competitive performance with respect to state-of-the-art population-based algorithms employing a DE logic. Finally, the cDE is applied to a challenging experimental case study regarding the on-line training of a nonlinear neural-network-based controller for a precise positioning system subject to changes of payload. The main peculiarity of this control application is that the control software is not implemented into a computer connected to the control system but directly on the micro-controller. Both numerical results on the test functions and experimental results on the real-world problem are very promising and allow us to think that cDE and future developments can be an efficient option for optimization in hardware environments characterized by limited memory.

Real-Time Estimation of Fundamental Frequency and Harmonics for Active Shunt Power Filters in Aircraft Electrical Systems

A novel algorithm for fundamental frequency and harmonic components detection is presented in this paper. The technique is based on a real-time implementation of discrete Fourier transform, and it allows fast and accurate estimation of fundamental frequency and harmonics of a distorted signal with variable fundamental frequency. It is suitable for active shunt filter applications, when fast and accurate tracking of the reference signal is required to achieve a good control performance. The main application for the algorithm is aircraft ac power systems, where the fundamental frequency can be either fixed on 400 Hz and its actual value fluctuates around the nominal value, or variable in the range 360-900 Hz. Hence, a real-time estimation of fundamental frequency is essential for active filter control. The proposed algorithm has been at first implemented in Matlab/Simulink for computer simulation, and it has been compared with a Phase Locked Loop (PLL) algorithm for frequency detection and the synchronous dq reference method for harmonic detection. Experimental tests have been carried out in order to validate the simulation results. The distorted current absorbed by a nonlinear load is analyzed and processed by means of a digital implementation of the algorithm running on the active shunt power filter control DSP, in order to calculate the active filter compensating current.

Real-Valued Compact Genetic Algorithms for Embedded Microcontroller Optimization

Recent research on compact genetic algorithms (cGAs) has proposed a number of evolutionary search methods with reduced memory requirements. In cGAs, the evolution of populations is emulated by processing a probability vector with specific update rules. This paper considers the implementation of cGAs in microcontroller-based control platforms. In particular, to overcome some problems related to the binary encoding schemes adopted in most cGAs, this paper also proposes a new variant based on a real-valued solution coding. The presented variant achieves final solutions of the same quality as those found by binary cGAs, with a significantly reduced computational cost. The potential of the proposed approach is assessed by means of an extensive comparative study, which includes numerical results on benchmark functions, simulated and experimental microcontroller design problems.

Sliding Mode Neuro-Adaptive Control of Electric Drives

An innovative variable-structure-systems-based approach for online training of neural network (NN) controllers as applied to the speed control of electric drives is presented. The proposed learning algorithm establishes an inner sliding motion in terms of the controller parameters, leading the command error towards zero. The outer sliding motion concerns the controlled electric drive, the state tracking error vector of which is simultaneously forced towards the origin of the phase space. The equivalence between the two sliding motions is demonstrated. In order to evaluate the performance of the proposed control scheme and its practical feasibility in industrial settings, experimental tests have been carried out with electric motor drives. Crucial problems such as adaptability, computational costs, and robustness are discussed. Experimental results illustrate that the proposed NN-based speed controller possesses a remarkable learning capability to control electric drives, virtually without requiring a priori knowledge of the plant dynamics and laborious startup procedures

Fuzzy control of a mobile robot

This paper describes the design of a new fuzzy logic-based navigation algorithm for autonomous robots. This design effectively achieves correct environment modeling and noisy and uncertain sensory data processing on low-cost hardware equipment. A hierarchical control strategy is presented in which three different reactive behaviors are fused in a single control law by means of a fuzzy supervisor guaranteeing robot safety and task accomplishment. Due to the inherent transparency of fuzzy logic, the proposed algorithm is computationally light, easily reconfigurable, and well-performing in a wide range of differing operating conditions and environments

End Effects in Linear Tubular Motors and Compensated Position Sensorless Control Based on Pulsating Voltage Injection

The sensorless position control of permanent-magnet (PM) synchronous motors can be successfully implemented by superimposing a high-frequency voltage signal on the control voltage. In this paper, the position estimation is obtained by means of a high-frequency sinusoidal voltage signal injected along the estimated -axis. Several methods proposed in the literature obtain the position estimation by tracking the zero condition of the high-frequency current component. We propose a new approach that also exploits the -axis high-frequency current component and allows working with injected voltage signal of reduced amplitude, thus reducing noise and additional losses. The main contribution of this paper relies in the compensation of the motor end effects due to the finite length of the tubular motor armature. These effects must be taken into account in the motor modeling because they cause an error in the position estimation that varies with the motor position. The modeling of the phenomenon and a proper compensation technique are proposed in this paper. Last, a simplified integral-type controller is used to estimate motor position instead of the commonly adopted proportional-integral controller plus integrator, and this requires a low-effort design. Experiments on a linear tubular PM synchronous-motor prototype are presented to validate the theoretical analysis and evidence the feasibility of the proposed sensorless technique.

On-line genetic design of anti-windup unstructured controllers for electric drives with variable load

In this paper, we describe an evolutionary design procedure for discrete-time anti-windup controller for electrical drives. Using a genetic algorithm devised to test and compare controllers of different orders, we search for the discrete anti-windup controller achieving the optimal compromise of weighted cost and performance indices. The search is performed on-line, on the physical hardware, by continuously downloading and testing new solutions on a microprocessor running the control algorithms in real time. The controller obtained through genetic search significantly outperforms alternative schemes obtained with conventional design techniques.

Accurate Inverter Error Compensation and Related Self-Commissioning Scheme in Sensorless Induction Motor Drives

The inverter non-linear effects limit the performance of Sensorless Field Oriented Control (SFOC) of Induction Motor (IM) drives. When no direct voltage measurement is available and the reference voltages are used instead, the problem relies in the non-linear error that affects the motor back-emf estimation. The back-emf signals are fundamental for estimating the motor flux in most of SFOC schemes, either oriented on the stator flux or on the rotor flux. Many methods have been proposed for compensating such non-linearities (IGBT dead-time and on-state voltage drops) by means of the digital control algorithm. In many references the compensation relies on an inverter model based on the signum function of the motor phase currents. In the considered cases, the identification of the compensation parameters is performed off-line, often with additional processing tasks. Very few schemes are integrated into the control code of the IM drive. The goal of the paper is to present a model of the inverter non-linear effects that is more accurate than the signum-based one, and the related self-identification method. The proposed inverter model can be identified directly by the digital controller at the drive start-up with no extra measures other than the motor phase currents and dc-link voltage. After the commissioning session, the compensation does not require to be tuned furthermore and is robust against temperature detuning. The identification is based on the feedback signal of the closed-loop flux observer in dc current conditions. The experimental results, presented here for a rotor flux oriented SFOC IM drive for home appliances, demonstrate the feasibility of the proposed solution.

Running DFT-Based PLL Algorithm for Frequency, Phase, and Amplitude Tracking in Aircraft Electrical Systems

Phase-locked loop (PLL) algorithms are commonly used to track sinusoidal components in currents and voltage signals in three-phase power systems. Despite the simplicity of those algorithms, problems arise when signals have variable frequency or amplitude, or are polluted with harmonic content and measurement noise, as can be found in aircraft ac power systems where the fundamental frequency can vary in the range 360-900 Hz. To improve the quality of phase and frequency estimates in such power systems, a novel PLL scheme based on a real-time implementation of the discrete Fourier transform (DFT) is presented in this paper. The DFT algorithm calculates the amplitudes of three consecutive components in the frequency domain. These components are used to determine an error signal which is minimized by a proportional-integral loop filter in order to estimate the fundamental frequency. The integral of the estimated frequency is the estimated phase of the fundamental component, and this is fed back to the DFT algorithm. The proposed algorithm can therefore be considered to be a PLL in which phase detection is performed via a DFT-based algorithm. A comparison has been made of the performances of a standard PLL and the proposed DFT-PLL using computer simulations and through experiments.

Sensorless Position Control of Permanent-Magnet Motors With Pulsating Current Injection and Compensation of Motor End Effects

The sensorless position control of permanent-magnet motors is successfully implemented by superimposing a high-frequency voltage signal on the voltage reference or adding a high-frequency current signal to the current reference. The former approach is usually preferred because of its simplicity, although the latter one may allow better performance. This paper presents a new algorithm for the sensorless control of low-saliency permanent-magnet synchronous motors based on high-frequency sinusoidal current signal injection into the d-axis. Different from the related literature, the position information is derived by analyzing the measured high-frequency currents. The amplitude of the d-axis voltage reference is also exploited to improve performance. A proportional-integral (PI) controller plus a resonant term (PI-RES) is adopted to ensure the accurate tracking of both the dc and high-frequency components of the d -axis current reference. The main advantages of the proposed approach are the increased accuracy and sensitivity with respect to the approach based on voltage injection, the insensitiveness to inverter nonlinearities that are compensated by the current regulation loop, the actual control on the injected current value, and the practical absence of acoustic noise. Experiments on a linear tubular permanent-magnet synchronous motor prototype have been carried out to verify the aforementioned advantages. This paper also presents a discussion of the parameters of the PI-RES.