Back-propagation Neural Network (BPNN) algorithm is one of the most widely used and a popular technique to optimize the feed forward neural network training. Traditional BP algorithm has some drawbacks, such as getting stuck easily in local minima and slow speed of convergence. Nature inspired meta-heuristic algorithms provide derivative-free solution to optimize complex problems. This paper proposed a new meta-heuristic search algorithm, called cuckoo search (CS), based on cuckoo bird's behavior to train BP in achieving fast convergence rate and to avoid local minima problem. The performance of the proposed Cuckoo Search Back-Propagation (CSBP) is compared with artificial bee colony using BP algorithm, and other hybrid variants. Specifically OR and XOR datasets are used. The simulation results show that the computational efficiency of BP training process is highly enhanced when coupled with the proposed hybrid method.

A new back-propagation neural network optimized with cuckoo search algorithm

The efficient management of product development projects is important to reduce the required development time and cost. However, each project is unique in nature and the duration of activities involed in a project often cannot be predicted accurately. The uncertainty of activity duration may lead to incorrect scheduling decisions. The objective of this research is to develop a fuzzy scheduling methodology to deal with this problem. Possibility theory is used to model the uncertain and flexible temporal information. The concept of schedule risk is proposed to evaluate the schedule performance. A fuzzy beam search algorithm is developed to determine a schedule with the minimum schedule risk and the start time of each activity is selected to maximize the minimum satisfaction degrees of all temporal constraints. In addition, the properties of schedule risk are also discussed. We show that the proposed methodology can assist project managers in selecting a schedule with the least possibility of being late in an uncertain scheduling environment. An example with an electronic product development project is used to illustrate the developed approach.

(Fuzzy Sets and Systems, 127:099-116)A Fuzzy Project Scheduling Approach to Minimize Schedule Risk for Product Development

Nowadays, computer scientists have shown the interest in the study of social insect’s behaviour in neural networks area forsolving different combinatorial and statistical problems. Chief among these is the Artificial Bee Colony (ABC) algorithm. This paper investigates the use of ABC algorithm that simulates the intelligent foraging behaviour of a honey bee swarm. Multilayer Perceptron(MLP) trained with the standard back propagation algorithm normally utilises computationally intensive training algorithms. One of thecrucial problems with the backpropagation (BP) algorithm is that it can sometimes yield the networks with suboptimal weights becauseof the presence of many local optima in the solution space. To overcome ABC algorithm used in this work to train MLP learning thecomplex behaviour of earthquake time series data trained by BP, the performance of MLP-ABC is benchmarked against MLP trainingwith the standard BP. The experimental result shows that MLP-ABC performance is better than MLP-BP for time series data.

Using Artificial Bee Colony Algorithm for MLP Training on Earthquake Time Series Data Prediction

Inertial Navigation System (INS) is often combined with Global Navigation Satellite System (GNSS) to increase the positioning accuracy and continuity. In complex urban environments, GNSS/INS integrated systems suffer not only from dynamical model errors but also GNSS observation gross errors. However, it is hard to distinguish dynamical model errors from observation gross errors because the observation residuals are affected by both of them in a loosely-coupled integrated navigation system. In this research, an optimal Radial Basis Function (RBF) neural network-enhanced adaptive robust Kalman filter (KF) method is proposed to isolate and mitigate the influence of the two types of errors. In the proposed method, firstly a test statistic based on Mahalanobis distance is treated as judging index to achieve fault detection. Then, an optimal RBF neural network strategy is trained on-line by the optimality principle. The network's output will bring benefits in recognizing the above two kinds of filtering fault and the system is able to choose a robust or adaptive Kalman filtering method autonomously. A field vehicle test in urban areas with a low-cost GNSS/INS integrated system indicates that two types of errors simulated in complex urban areas have been detected, distinguished and eliminated with the proposed scheme, success rate reached up to 92%. In particular, we also find that the novel neural network strategy can improve the overall position accuracy during GNSS signal short-term outages.

https://www.mdpi.com/1424-8220/18/9/3091/pdf

An Optimal Radial Basis Function Neural Network Enhanced Adaptive Robust Kalman Filter for GNSS/INS Integrated Systems in Complex Urban Areas.

This paper investigates the static deflection as well as the nonlinear resonant behaviour of an electrically actuated MEMS resonator subject to a temperature rise. The deformable electrode is actuated by a harmonic AC load superimposed to a constant DC load. Taking into account thermal effects, the size-dependent equation of motion is derived by means of the extended Hamilton's principle together with the modified couple stress theory. The effect of temperature rise on the mechanical properties of the microresonator is also taken into account. Based on the Galerkin method, a high-dimensional discretised model of the microresonator is obtained which is solved by means of the pseudo-arclength continuation technique. In particular, the non-trivial deflected configuration of the deformable electrode, as well as the static pull-in voltage, is obtained. The nonlinear resonant behaviour of the system is analyzed when the system is actuated by the AC voltage near the primary resonance. The effect of temperature rise on the static and dynamic responses of the system is highlighted. Furthermore, it is shown that when the mechanical properties are considered temperature-dependent, rather than constant, the behaviour of the system changes significantly.

Nonlinear thermo-mechanical behaviour of MEMS resonators

An integrating device of variable frequency and speed-down is an important driving and control unit which combines a variable frequency module, a variable frequency motor and a retarder. Its technical core is the variable frequency control module developed in this paper. The paper realized a variable frequency and adjusting speed system controled by vector based DSP and IPM. The operating priciple, hardware and software of the system are analyzed and studied. The testing results proved that it had good dynamic and static performance.

Realization of AC Motor Vector Control Based on DSP

Aiming at improving the stability and real-time characteristics of welding control system,a hardware control system based on digital signal processor(DSP) was designed.The μC/OS-II embedded real-time kernel was applied in the control system software design,and it was transplanted to TMS320LF2407A.According to the system structure and functional requirements of the control system,welding application tasks were created and intertask communications were designed.Experiment results prove that this control system has high reliability.

Application of μC/OS-II in digital welder control system

The multi-function loader is a complex system in which mechanical, electronic, hydraulic and control systems are assembled. It has a significant multi-disciplinary coupling characteristic. The traditional single-domain modeling simulation tools can not do the simulation and optimization of overall performance for modern complex products. In view of this situation, the paper studied the modeling methods of the multi-function loader executing systems, and established the dynamic, hydraulic and control models. At the same time the paper analyzed the coupling relationship between the hydraulic system model and the mechanical system model to ensure the accurate agreement among the loader hydraulic system simulation and the gestures and movements, proposed the realization method of the static pressure transmission and control with the cascade feedback of the diesel engine's output speed and variable pump's output speed to improve the speed, pressure, power adjustment range of the loader and the ability to adapt working conditions. The PID adjustment method was used to control the shovel turning and maintain horizontal in the moving process. In the iSIGHT environment, the numerical optimization algorithm of sequential quadratic programming was adopted as an optimization program to achieve the multi-disciplinary collaborative optimization of the loader.

Research on the Multi-domain Modeling and Optimizing Method for Loader Executing System

To simulate the auto-focus mechanism of human's eyes by using artificial neural networks is a reasonable and available way to improve the auto-focus effect of digital cameras The evaluation of image definition plays a key role in the auto-focus of digital cameras, so constructing the image definition evaluation is an important step of designing auto-focus systems In this paper, we use the pattern recognition method of the RBF neural networks to recognize the image definition The designed neural networks are trained by the training set that is composed of 75 pictures Then, the testing set that is composed of 102 pictures is verified experimentally The experiment results indicate that this method is quite effective The method gives full play to self-adapting ability of neural networks and gets a higher recognition rate

Application of Neural Networks in Image Definition Recognition

For the engineering institutions and universities, the implementation of the students’ creativity, innovation and entrepreneurial abilities, has become imperative. According to our orientation, training model and development needs, combined with the our condition mainly administered by the students of engineering classes and economy & management classes, the education programs of creativity, innovation and entrepreneurial capacity are put forward. The matched experimental platforms are established. According to the educational reform idea of a phased, modular, integrated and open style, the new practical system of a progressive and constant style for the engineering undergraduates is built with the “big engineering education” thought as a guide. Through the years of the “creativity, innovation and entrepreneurship” teaching practice, the student’s initiative was developed, the innovative ability and overall quality of our students were greatly enhanced, and the outstanding achievements in the innovative awareness and practical ability won the praise of the employing units.

Chen Guojin

Papers

Realization of AC Motor Vector Control Based on DSP

Application of μC/OS-II in digital welder control system

Research on the Multi-domain Modeling and Optimizing Method for Loader Executing System

Application of Neural Networks in Image Definition Recognition

Study on the Education Plan of the Creativity, Innovation and Entrepreneurship Ability in University Students