The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
This paper presents new results on multistability of networks when neurons undergo self-excitation and second-order synaptic connectivity. Due to self-excitation of neurons, we split state space into invariant regions and establish new criteria of coexistence of equilibria (periodic orbits) which are exponentially stable. It is shown that high-order synaptic connectivity and external inputs play an...
The adjusted GPS height is the height above the WGS-84 ellipsoid. It is necessary to convert a GPS height into a normal height in engineering applications. GPS height conversion is usually used the standard BP (back-propagation algorithm) neural network model, but there are some defects in standard BP algorithm: low efficiency and easy to fall into local minimum. Aiming at overcoming the slow convergence...
Innovation ability of Industrial clusters is an important measure of regional innovation. Industrial clusters with strong innovation ability can promote the development of innovative enterprises, which are the key element of regional innovation. Therefore, it is important to establish models to evaluate innovation ability for industry clusters. In this paper, an improved BP neural network model was...
An improved functional link neural network was proposed for the identification of the dynamic system. In the improved method, the partial derivatives of the network outputs w.r.t its weights were re-deduced, and the more accurate evaluations of the derivatives were obtained. As a result, a novel recursive algorithm was developed to update the weights of the FLNN and a faster learning could be expected...
Original particle swarm optimization (OPSO) algorithm was modified in the paper, and a self-adaptive PSO (SPSO) was proposed. In this algorithm, SPSO combines Elman neural network (ENN) and forms SPSO-ENN hybrid algorithm. Compared with ENN algorithm, the experiment results show that SPSO-ENN has less adjustable parameters, faster convergence speed and higher precision in the nonlinear function identification.
Artificial Neural Networks (ANN) is gaining significant importance for pattern recognition applications particularly in the medical field. A hybrid neural network such as Counter Propagation Neural Network (CPN) is highly desirable since it comprises the advantages of supervised and unsupervised training methodologies. Even though it guarantees high accuracy, the network is computationally non-feasible...
Artificial neural networks (ANN) and fuzzy systems are the widely preferred artificial intelligence techniques for biological computational applications. While ANN is less accurate than fuzzy logic systems, fuzzy theory needs expertise knowledge to guarantee high accuracy. Since both the methodologies possess certain advantages and disadvantages, it is primarily important to compare and contrast these...
Evaluation of certain properties of calcined alumina or special grade alumina is necessary and important to its manufactures. Generally it is determined in the laboratories using different instrumental and manual methods, which is cost and time intensive. In the present work, evolving neural network has been used for the estimation of a property given few others. To evolve the neural network model...
First, the forecasting principle and improved algorithms about BP ANN are briefly introduced. Then, an improved algorithm about BP ANN is put forward which based on subordinating degree function, and conduct simulating tests. The result indicates that convergence is rapid without changing the forecasting precision. Based on this and combined with the characteristic of power load forecasting, a model...
This paper presents a neural network framework for implementing unknown time-varying mappings. A unified architecture of time-varying neural networks is proposed, and the methodology of iterative learning is used for the network training. Convergence results of the iterative learning least squares algorithm are derived under assumption of bounded input signals. Periodic neural networks are explored...
Shortwave power amplifiers (PAs) are usually considered as memoryless devices in most existing predistortion techniques. Nevertheless, in shortwave communication systems, PA memory effects can no longer be ignored and memoryless predistortion cannot linearize PAs effectively. By analyzing the characteristics of the power amplifier, an improved predistortion method for memory power amplifier is presented...
The principal of terrain contour matching (TERCOM) technique is to measure the present altitude and compare it with the geographical reference altitude of the vehicle. Accordingly, with maximum correlation computation, it can estimate the present position and correct the guidance error of flight vehicle. TERCOM is an important and a crucial technique using in today's cruise missile or some special...
In this paper, we construct a neural network iteration method for simultaneous extraction of all roots of algebraic polynomial with variable learning rate. Its convergence was researched. The specific examples showed that the proposed method can simultaneously find all roots of algebraic polynomial at a very rapid convergence and very high accuracy with less computation.
In this paper, the theories of genetic algorithm (GA) and back propagation (BP) algorithm are introduced. For the purpose of overcoming the disadvantages of standard BP algorithm, such as local optimum and low convergence speed, the paper adopts genetic algorithm optimizing BP neural network for training. By analyzing computer stimulation results and comparing with traditional blind equalization algorithm,...
In this paper, a pose-variant face recognition system is presented for study of human-robot interaction design. An iterative fitting algorithm is proposed to extract feature-point positions based on active appearance model (AAM). Comparing with the traditional Lucas-Kanade algorithm, the proposed iterative algorithm improves the capability of correct convergence as a larger variation of head posture...
A fast and exact neural-network algorithm is proposed to find zeros of polynomials which were not solved by the most other methods. Its convergence rule was presented and proved. The computation is carried out by simple steepest descent rule with variable step-size. The specific examples illustrated that the proposed method can find the roots of polynomials at a very rapid convergence and very high...
In order to eliminate the shortcomings of traditional neural networks in handwritten Chinese characters recognition, such as the premature convergence, a novel intelligent method is presented, which uses the particle swarm optimization (PSO) algorithm with adaptive inertia weight to train the neural networks. The main idea is that the optimum weights and thresholds of the neural networks is acquired...
Fast convergence-rate, low computation complexity and good stability are important goals in the researching area of neural network learning algorithm. A kind of parallel computing lagged-start hybrid optimization algorithm is studied, it not only integrates the basic gradient method and the unconstrained optimization algorithm to realize the supplement of their advantages, but also makes full use...
In order to decrease negative effects brought by the particularity and complexity of imaging environment, and satisfy the real-time need of the underwater task, combined invariant moments are extracted as recognition features. Furthermore, an underwater target recognition system based on neural network which improved by Artificial Fish Swarm Algorithm (AFSA) is proposed. AFSA is of capable of attaining...
The study includes the correlation dimension and the largest Lyapunov exponent of sea clutter based on real radar data obtained with IPIX X-band polarimetric coherent radar, which proved that sea clutter has chaotic characteristics. A method of prediction about sea clutter signal based on chaotic neural network and theory of phase-space reconstruction is established, which has a in-depth analysis...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.