Learning fuzzy controllers in mobile robotics with embedded preprocessing

The automatic design of controllers for mobile robots usually requires two stages. In the first stage, sensorial data are preprocessed or transformed into high level and meaningful values of variables which are usually defined from expert knowledge. In the second stage, a machine learning technique is applied to obtain a controller that maps these high level variables to the control commands that are actually sent to the robot. This paper describes an algorithm that is able to embed the preprocessing stage into the learning stage in order to get controllers directly starting from sensorial raw data with no expert knowledge involved. Due to the high dimensionality of the sensorial data, this approach uses Quantified Fuzzy Rules (QFRs), that are able to transform low-level input variables into high-level input variables, reducing the dimensionality through summarization. The proposed learning algorithm, called Iterative Quantified Fuzzy Rule Learning (IQFRL), is based on genetic programming. IQFRL is able to learn rules with different structures, and can manage linguistic variables with multiple granularities. The algorithm has been tested with the implementation of the wall-following behavior both in several realistic simulated environments with different complexity and on a Pioneer 3-AT robot in two real environments. Results have been compared with several well-known learning algorithms combined with different data preprocessing techniques, showing that IQFRL exhibits a better and statistically significant performance. Moreover, three real world applications for which IQFRL plays a central role are also presented: path and object tracking with static and moving obstacles avoidance.     

新一代多模型迭代算法在70 kVp成人副鼻窦CT扫描中的应用

探讨成人在70 kVp副鼻窦CT检查中降低有效辐射剂量的同时,获取最佳图像质量的新一代多模型迭代算法(Adaptive Statistical Iterative Reconstruction-V,Asir-V)的权重。本研究对20例成人患者进行副鼻窦Revolution CT检查,采用70 kVp、230 mA条件下进行螺旋CT扫描,扫描完成后对每例患者的原始数据采用不同权重(0,20%,40%,60%,80%,100%)的后置新一代多模型迭代算法(Adaptive Statistical Iterative Reconstruction-V,Asir-V)重建的方法,共获得120幅图像,并对每例6组图像进行客观评价及主观评分。客观评价包括噪声(standard deviations, SD)和对比噪声比(contrast noise ratio,CNR),主观评价由2名评价医生对图像质量评分并对两者的主观一致性进行评价。每例患者扫描完成后,记录其剂量长度乘积(dose-length product,DLP)和CT容积剂量指数(CT dose index volume,CTDI_vol),并计算其平均有效辐射剂量(effective dose,ED)。客观评价结果表明不同权重的Asir-V图像噪声随着Asir-V权重的提升而降低,CNR值随着Asir-V权重的提升而升高,且有效辐射剂量ED较常规头颈部CT检查的平均辐射剂量明显减少;主观评价结果表明不同权重的Asir-V图像评分不同,在Asir-V 60% 时评分最高,且2名医生主观评价一致性好。结合主客观评价结果,认为在70 kVp、230 mA成人副鼻窦Revolution CT检查中,Asir-V 60％的图像既可满足临床诊断的需要,也可明显降低有效辐射剂量。     

Precise contour following for biaxial systems via an A-type iterative learning cross-coupled control algorithm

This paper proposes an A-type iterative learning cross-coupled control (CCC) algorithm for biaxial systems. An algebraic equation based contour error model is used as the CCC input. This model has the advantage that it is zero if and only if the real value vanishes. The iterative learning CCC is designed to make its input converge to zero. Hence, it is expected to that the contour error will converge to zero as well. After analyzing the control algorithm convergence condition in the frequency domain, the proposed method is implemented on a motion stage. Experimental results show that the algorithm perfectly follows contours as the cycles approach infinity regardless of whether tracking errors are small or large.     

基于线特征及迭代最近点算法的地基建筑物点云自动配准方法

为克服迭代最近点(ICP)算法易陷入局部最优的缺陷，提出一种基于线特征及ICP算法的地基建筑物点云自动配准方法。首先，基于法向一致性进行建筑物点云平面分割；接着，采用alpha-shape算法进行点簇轮廓线提取，并拆分和拟合处理得到特征线段；然后，以线对作为配准基元，以线对夹角和距离作为相似性测度进行同名特征匹配，实现建筑物点云的粗配准；最后，以粗配准结果为初值，进一步采用ICP算法完成点云精确配准。利用两组部分重叠的建筑物点云进行配准实验，实验结果表明，采用由粗到精的配准方法能有效改善ICP算法对初值依赖的问题，实现具有部分重叠的建筑物点云的有效配准。     

Oscillation mode analysis for power grids using adaptive local iterative filter decomposition

This study proposes an algorithm based on adaptive local iterative filtering decomposition (ALIFD), which is applicable for the feature extraction of a power oscillating signal in a power system. The ALIFD algorithm uses the Fokker–Planck equation to construct the filter function as well as filter sifting to obtain the intrinsic mode function (IMF) with stable features. This algorithm has a solid mathematical foundation and can effectively avoid the mode-mixing problems in the empirical mode decomposition (EMD) algorithm. In this study, the ALIFD algorithm is initially used to obtain the oscillation component. Subsequently, Hilbert Transformation (HT) of each component is performed, and oscillation characteristic parameters are extracted. Analysis results of the test signal, the simulation signal, and the measured data verify the effectiveness of the proposed algorithm. Meanwhile, the comparative results of the EMD algorithm prove that the proposed method is highly adaptive to extracting the characteristics of power oscillation in a power system.     

Shunt capacitor placement in radial distribution networks considering switching transients decision making approach

This paper provides a new approach in decision making process for shunt capacitor placement in distribution networks. The main core of the evaluation process is a multi-objective framework to allocate the capacitor banks. The power loss and the total harmonic distortion (THD) are the objective functions of the system under study in a long-term planning horizon. In order to select the executive plan introduced by using a multi-objective model, transient switching overvoltages have been considered. As the size and location of shunt capacitors may result in unacceptable overvoltages, the proposed technical decision making framework can be applied to avoid corresponding damages. In this paper, an iterative conventional power flow technique is introduced. This technique can be applied to evaluate THD for distribution networks as well as other power flow based objectives, such as power losses calculation and voltage stability assessment. The presented framework is a two stage one where at the first stage, a non-dominated sorting genetic algorithm (NSGA-II) augmented with a local search technique is used in order to solve the addressed multi-objective optimization problem. Then, at the second stage, a decision making support technique is applied to determine the best solution from the obtained Pareto front. In order to evaluate the effectiveness of the proposed method, two benchmarks are addressed in this paper. The first test system is a 9-bus distribution network and the second one is an 85-bus large scale distribution network. The simulation results show that the presented method is satisfactory and consistent with the expectation.     

Stochastic resonance in parallel concatenated turbo code decoding

In an array of threshold devices, we examine the effect of noise in improving performance of turbo code decoding. Such a phenomenon of noise-enhanced effect is termed stochastic resonance (SR). When signal is subthreshold, SR is observed by using Gaussian noise during iterative decoding. That is, the minimal bit error ratio (BER) is achieved at some non-zero noise intensity level. Besides, the larger the number of threshold devices is, the more remarkable the SR effect becomes. Especially when noise intensity is nearly optimal, BER approximates to zero after a few decoding iterations. Moreover, when Gaussian mixture noise is utilized, suprathreshold stochastic resonance (SSR) occurs in turbo decoding. These results show the beneficial effect of noise in channel coding and decoding.