Chih-Li Huo

A novel self-constructing Radial Basis Function Neural-Fuzzy System

This paper proposes a novel self-constructing least-Wilcoxon generalized Radial Basis Function Neural-Fuzzy System (LW-GRBFNFS) and its applications to non-linear function approximation and chaos time sequence prediction. In general, the hidden layer parameters of the antecedent part of most traditional RBFNFS are decided in advance and the output weights of the consequent part are evaluated by least square estimation. The hidden layer structure of the RBFNFS is lack of flexibility because the structure is fixed and cannot be adjusted effectively according to the dynamic behavior of the system. Furthermore, the resultant performance of using least square estimation for output weights is often weakened by the noise and outliers. This paper creates a self-constructing scenario for generating antecedent part of RBFNFS with particle swarm optimizer (PSO). For training the consequent part of RBFNFS, instead of traditional least square (LS) estimation, least-Wilcoxon (LW) norm is employed in the proposed approach to do the estimation. As is well known in statistics, the resulting linear function by using the rank-based LW norm approximation to linear function problems is usually robust against (or insensitive to) noises and outliers and therefore increases the accuracy of the output weights of RBFNFS. Several nonlinear functions approximation and chaotic time series prediction problems are used to verify the efficiency of self-constructing LW-GRBFNIS proposed in this paper. The experimental results show that the proposed method not only creates optimal hidden nodes but also effectively mitigates the noise and outliers problems.

Optimal UAV flight path planning using skeletonization and Particle Swarm Optimizer

The purpose of this paper is to search the best flight route efficiently for unmanned aerial vehicle (UAV) in the 3-dimention complicated topography. The proposed method for the best flight route is mainly utilizing evolutionary algorithm, and give the proper initial population of evolutionary algorithm through skeletonization, efficient pre-processing procedure. In order to provide a smooth flight route for UAV, this paper adopts B-spline Curve method. Several control points of B-spline Curve method must be determined to generate flight route. The best control points can be calculated by Particle Swarm Optimizer (PSO). In this paper, the initial population of PSO is provided by skeletonization. The skeletonization of pre-processing procedure mainly includes two parts: one is Skeletonization and the other is candidate path searching. The purpose of pre-processing procedure is to reduce computation time, to prevent search the best solutions aimless, and execute evolutionary process efficiently. This paper uses Matlab as the experiment environment. The results of the experiments present the proposed method can provide the best flight route for UAV efficiently.

Variable feedback gain control design based on particle swarm optimizer for automatic fighter tracking problems

The main focus of this paper is to develop an optimization method for the automatic fighter tracking (AFT) problem. The AFT problem is similar to a general evader-pursuer maneuvering automation problem between the dynamic systems of two highly interactive objects. This paper proposes a particle swarm optimizer-based variable feedback gain controller (PSO-based VFGC) for dealing with AFT problems. The PSO-based VFGC is designed to obtain the control value of a pursuer through an error-feedback gain controller. Once conditions of system closed-loop stability have been satisfied, the optimal feedback gains can be obtained through PSO, and the actual control values can be derived from the obtained values. Simulation results confirm the capabilities of the proposed method by comparing the results against two other methods in the field: the weight matrix value defined Ricatti equation, and the linear matrix inequality (LMI) based linear quadratic regulator (LQR). The performance of the proposed method is superior to that of its alternatives.

Intelligent flight task algorithm for unmanned aerial vehicle

This paper proposes an intelligent flight task algorithm for unmanned aerial vehicles (UAV) to effectively determine and search the best flight routes in three-dimensional and complicated topographies, where the topography data includes the height of the object on the earth and the aerial imagery map of the earth is obtained by the Satellite. The proposed method is composed mainly of the particle swarm optimizer (PSO), skeletonization, and B-spline curve. Skeletonization is regarded as a pre-processing procedure for topography data to reduce complexity of the searching space and inhibit aimless searches of solutions. B-spline curve method is adopted to provide a smooth flight route for UAV. The best control points of a B-spline curve are determined by PSO. In this paper, the initial population of PSO is acquired through skeletonization to improve the efficiency of the search for the best flight route. The results of the simulation demonstrate the abilities of the proposed method to provide the best flight routes efficiently for UAV.

Out-of-focus blur estimation for blind image deconvolution: Using particle swarm optimization

This study addresses the blind image deconvolution which uses only blurred image and less point spread function (PSF) information to restore the original image. To identify the blind image it is a very important step for restoring the image. Therefore, the first step is to look for PSF model. In this paper, particle swarm optimization (PSO) is utilized to seek the unknown PSF. The objective function is based on the wavelet transform. It can identify the parameters of PSF exactly. Finally, the feasibility and validity of proposed algorithm are demonstrated by several simulations.

Intelligent Maneuvering Decision System for Computer Generated Forces Using Predictive Fuzzy Inference System

The purpose of this paper is to develop an intelligent maneuvering decision system (IMDS) for computer generated forces (CGF). The proposed CGF can take actions similar to a human pilot to gain an advantageous status over the enemy target using the IMDS. The IMDS will produce the best control command from the control alternatives for the CGF in an air combat environment. In this paper, a predictive fuzzy inference system (PFIS) is proposed as the IMDS for CGF, which incorporates and mimics human thinking capability and the maximum capacity of CGF. Before PFIS executes the fuzzy inference system (FIS) process, it will generate the control alternatives from CGF’ s decision space, and allow CGF to predict its future posture. This study assumes that CGF can accurately predict an enemy target’s future position, and then PFIS applies the predicted data to generate the best control command. In this paper, the proposed algorithm is verified with two types of fighter flying data that are used as the enemy target’s flying trajectories. The simulation and discussion of the proposed algorithm shows that PFIS will enable CGF to obtain the best status in an air combat environment and the performance of the proposed algorithm will be affected by the CGF’s prediction ability for enemy target.

A novel calibration method based on heuristic B-spline model for fish-eye lenses

Distortion model vary from lens to lens even are same type camera because some unavoidable imperfections during manufacturing. Therefore we propose a novel calibration method based on heuristic B-spline model to identify the distortion model. Unlike common distortion models, proposed heuristic B-spline model can get better calibrated result. In order to evaluate the result of calibration, a test image is considered to compare proposed method with other distortion models. The experiment is shown that the proposed method has less mean square error (MSE) than others.

Effectively multi-swarm sharing management for differential evolution

This paper presents a novel multi-swarm sharing management for differential evolution (MsSDE) to deal with numerical optimization effectively. Multi-swarm is an effective search concept to keep the original search characteristic or effective balance strategies. However, it still has some defects need to overcome, such as weak search ability for smaller swarm and easy to fall into local optimal position. In order to overcome the problem mention above, the proposed multi-swarm sharing management can adjust each swarm size, share and analyze their information for other swarms to get more effective search ability. Testing and comparing results with original DE and EPUS-PSO by several benchmark functions, it showed that the proposed method has satisfying performance.

Road area detection based on image segmentation and contour feature

This paper developed two procedures to extract road area. It uses initial road detection and continuous image tracking to reduce computation cost. Experiment using three difference environments to verify that this algorithm is can be realized.

Time sequence based lane-marking identification

In this paper, time sequence based lane-marking identification method is proposed to deal with the classification and robust improvement of lane-marking detection mechanism which is proposed in our previous works. The proposed method collects information of several consecutive image frames to perform identification. The experimental results show that the developed system can effectively identify lane-marking in various driving environment.