A biologically inspired spiking neural network for sound source lateralization.

In this paper, a binaural sound source lateralization spiking neural network (NN) will be presented which is inspired by most recent neurophysiological studies on the role of certain nuclei in the superior olivary complex (SOC) and the inferior colliculus (IC). The binaural sound source lateralization neural network (BiSoLaNN) is a spiking NN based on neural mechanisms, utilizing complex neural models, and attempting to simulate certain parts of nuclei of the auditory system in detail. The BiSoLaNN utilizes both excitatory and inhibitory ipsilateral and contralateral influences arrayed in only one delay line originating in the contralateral side to achieve a sharp azimuthal localization. It will be shown that the proposed model can be used both for purposes of understanding the mechanisms of an NN of the auditory system and for sound source lateralization tasks in technical applications, e.g., its use with the Darmstadt robotic head (DRH).     

A low cost, non-individualized surround sound system based upon head related transfer functions: an ergonomics study and prototype development

This paper reports on the types and magnitudes of localization errors of simulated binaural direction cues generated using non-individualized, head-related transfer functions (HRTFs) with different levels of complexity. Four levels of complexity, as represented by the number of non-zero coefficients of the associated HRTF filters (128, 64, 32, 18 non-zero coefficients), were studied. Experiment 1 collected 1728 data runs that were exhaustive combinations of the four levels of complexity, nine simulated directions of sound (no direction (i.e., diotical-mono), 0 degrees , 45 degrees , 90 degrees , 135 degrees , 180 degrees , 225 degrees , 270 degrees , and 315 degrees azimuth angles at 0 degrees elevation), two repetitions, and 24 participants). Binaural cues generated from HRTFs of reduced complexity (from 128 to 18 non-zero coefficients) produced significantly higher localization errors for the directions of 45 degrees , 135 degrees , 225 degrees , and 315 degrees azimuth angles (p<0.01). From the directions of 0 degrees , 90 degrees , and 270 degrees azimuth angles, the cues produced by HRTFs with reduced complexity did not affect the localization error (p>0.2). Surprisingly, cues produced by HRTFs of 128 non-zero coefficients did not have the lowest number of errors. From 45 degrees , 135 degrees , 225 degrees , and 315 degrees , the lowest numbers of errors were obtained from cues produced by HRTFs of 64, 32, 32, and 64 non-zero coefficients, respectively. Based on these findings, a prototype virtual headphone-based surround-sound (VHSS) system was developed. A double-blind usability experiment with 32 participants indicated that the prototype VHSS system received significantly better surround-sound ratings than did a Dolby stereo system (p<0.02). This paper reports results from an original ergonomics study and the application of these results to the design of a consumer product.     

A neural network compensator for uncertainties in robotic assembly

It is difficult to represent the nonlinear characteristics in the dynamics of robot manipulators by means of a mathematical model. An alternative approach of using a neural network to learn the parametric and unstructured uncertainties in robot manipulators is proposed. It is then embedded in the structure of a joint torque perturbation observer to compensate for uncertainties in the robot dynamic model. As the result, an accurate estimate of the joint reaction torque against the environment can be deduced. The approach is applied to monitor the insertion force during electronic components assembly using a SCARA robot. A true teaching signal of neural network for learning the model uncertainties is obtained. Furthermore, a special motion test is conducted to generate the required training data set. After learning, the neural network is capable of reproducing the training data. The generalizing ability of the network enables it to output the correct compensation signal for a trajectory which it has not been trained. With the proposed technique, it is possible to verify the success of component insertion in real time and avoid causing damages to the electronic components.     

Neural network approach to trajectory synthesis for robotic manipulators

The paper deals with the collision free trajectory synthesis for industrial robotic manipulators. A new efficient method is proposed that is based on a neural network collision model. The developed iterative transformation procedure provides small computing times for the C-space synthesis and yields sufficiently precise configuration space map for the manipulators with many degrees of freedom. A topologically ordered neural network model is proposed to find the path in the configuration space. The stability of this model is proved using the Lyapunov function technique. To generate the collision model, a modification of the Radial Basis Function Network (RBFN) is used. The developed technique is illustrated by an application example of designing a robotic manufacturing cell for the automotive industry.     

Numerical solution of a system of fuzzy polynomials by fuzzy neural network

In this paper, a new approach for solving systems of fuzzy polynomials based on fuzzy neural network (FNN) is presented. This method can also lead to improve numerical methods. In this work, an architecture of fuzzy neural networks is also proposed to find a real root of a system of fuzzy polynomials (if exists) by introducing a learning algorithm. Finally, we illustrate our approach by numerical examples.     

基于神经网络的四元球面阵空间声源定向系统

针对移动机器人的目标声源定向问题,应用四元球面麦克风阵列建立了一套空间声源定向系统。为了解决通过阵列的时延值预测出目标声源方向的问题,提出了神经网络算法,并设计了双隐层BP神经网络。通过Matlab仿真验证了该网络可以实现近场和远场的声源定向,并在机器人本体上进行了实验测试,证明了该系统的实用性。     

海流环境下多AUV多目标生物启发任务分配与路径规划算法

针对多障碍物海流环境下多自治水下机器人(AUV)目标任务分配与路径规划问题, 本文在栅格地图构建的 基础上给出了一种基于生物启发神经网络(BINN)模型的新型自主任务分配与路径规划算法, 并考虑海流对路径规 划的影响. 首先建立BINN模型, 利用此模型表示AUV的工作环境, 神经网络中的每一个神经元与栅格地图中的位 置单元一一对应; 接着, 比较每个目标物在BINN地图中所有AUV的活性值, 并选取活性值最大的AUV作为它的获 胜AUV, 实现多AUV任务分配; 最后, 考虑常值海流影响, 根据矢量合成算法确定AUV实际的航行方向, 实现AUV路 径规划与安全避障. 海流环境下仿真实验结果表明了生物启发模型在多AUV水下任务分配与路径规划中的有效性.     

Artificial neural network approach for solving fuzzy differential equations

The current research attempts to offer a novel method for solving fuzzy differential equations with initial conditions based on the use of feed-forward neural networks. First, the fuzzy differential equation is replaced by a system of ordinary differential equations. A trial solution of this system is written as a sum of two parts. The first part satisfies the initial condition and contains no adjustable parameters. The second part involves a feed-forward neural network containing adjustable parameters (the weights). Hence by construction, the initial condition is satisfied and the network is trained to satisfy the differential equations. This method, in comparison with existing numerical methods, shows that the use of neural networks provides solutions with good generalization and high accuracy. The proposed method is illustrated by several examples.     

机器人听觉声源定位研究综述

声源定位技术定位出外界声源相对于机器人的方向和位置,机器人听觉声源定位系统可以极大地提高机器人与外界交互的能力.总结和分析面向机器人听觉的声源定位技术对智能机器人技术的发展有着重要的意义.首先总结了面向机器人听觉的声源定位系统的特点,综述了机器人听觉声源定位的关键技术,包括到达时间差、可控波束形成、高分辨率谱估计、双耳听觉、主动听觉和视听融合技术.其次对麦克风阵列模型进行了分类,比较了基于三维麦克风阵列、二维麦克风阵列和双耳的7个典型系统的性能.最后总结了机器人听觉声源定位系统的应用,并分析了存在的问题和未来的发展趋势.     

基于BP神经网络的正四面体阵列声源定向研究

针对利用正四面体麦克风阵列获取的时延值实现目标声源跟踪这个问题,提出了一种基于BP神经网络的声源定向方法。设计了一个含有双隐层的BP神经网络,使用Matlab神经网络工具箱进行仿真实验,证明可以实现远场和近场的声源定向,进而进行声源跟踪,有较高的实用性。     

A novel solution for maze traversal problems using artificial neural networks

In this paper we have addressed the problem of finding a path through a maze of a given size. The traditional ways of finding a path through a maze employ recursive algorithms in which unwanted or non-paths are eliminated in a recursive manner. Neural networks with their parallel and distributed nature of processing seem to provide a natural solution to this problem. We present a biologically inspired solution using a two level hierarchical neural network for the mapping of the maze as also the generation of the path if it exists. For a maze of size S the amount of time it takes would be a function of S (O(S)) and a shortest path (if more than one path exists) could be found in around S cycles where each cycle involves all the neurons doing their processing in a parallel manner. The solution presented in this paper finds all valid paths and a simple technique for finding the shortest path amongst them is also given. The results are very encouraging and more applications of the network setup used in this report are currently being investigated. These include synthetic modeling of biological neural mechanisms, traversal of decision trees, modeling of associative neural networks (as in relating visual and auditory stimuli of a given phenomenon) and surgical micro-robot trajectory planning and execution.     

A study of neural network based inverse kinematics solution for a three-joint robot

A neural network based inverse kinematics solution of a robotic manipulator is presented in this paper. Inverse kinematics problem is generally more complex for robotic manipulators. Many traditional solutions such as geometric, iterative and algebraic are inadequate if the joint structure of the manipulator is more complex. In this study, a three-joint robotic manipulator simulation software, developed in our previous studies, is used. Firstly, we have generated many initial and final points in the work volume of the robotic manipulator by using cubic trajectory planning. Then, all of the angles according to the real-world coordinates (x, y, z) are recorded in a file named as training set of neural network. Lastly, we have used a designed neural network to solve the inverse kinematics problem. The designed neural network has given the correct angles according to the given (x, y, z) cartesian coordinates. The online working feature of neural network makes it very successful and popular in this solution.     

Adaptive neural network output feedback control for flexible multi-link robotic manipulators

In this paper, a novel approach for adaptive control of flexible multi-link robots in the joint space is presented. The approach is valid for a class of highly uncertain systems with arbitrary but bounded dimension. The problem of trajectory tracking is solved through developing a stable inversion for robot dynamics using only joint angles measurement; then a linear dynamic compensator is utilised to stabilise the tracking error for the nominal system. Furthermore, a high gain observer is designed to provide an estimate for error dynamics. A linear in parameter neural network based adaptive signal is used to approximate and eliminate the effect of uncertainties due to link flexibilities and vibration modes on tracking performance, where the adaptation rule for the neural network weights is derived based on Lyapunov function. The stability and the ultimate boundedness of the error signals and closed-loop system is demonstrated through the Lyapunov stability theory. Computer simulations of the proposed robust controller are carried to validate on a two-link flexible planar manipulator.     

A variational approach to intensity approximation for remote sensing images using dynamic neural networks

Abstract: In remote sensing image processing, image approximation, or obtaining a high‐resolution image from a corresponding low‐resolution image, is an ill‐posed inverse problem. In this paper, the regularization method is used to convert the image approximation problem into a solvable variational problem. In regularization, the constraints on smoothness and discontinuity are considered, and the original ill‐posed problem is thereby converted to a well‐posed optimization problem. In order to solve the variational problem, a Hopfield‐type dynamic neural network is developed. This neural network possesses two states that describe the discrepancy between a pixel and adjacent pixels, the intensity evolution of a pixel and two kinds of corresponding weights. Based on the experiment in this study with a Landsat TM image free of added noise and a noisy image, the proposed approach provides better results than other methods. The comparison shows the feasibility of the proposed approach.     

A neural network controller for flexible-link robots

The object in this paper is to achieve tracking control of a partially unknown flexible-link robot arm. It is shown how to stabilize the internal dynamics by selecting a physically meaningful modified performance output for tracking; this output is the slow portion of the link-tip motions. That is, the tracking requirement is relaxed so that the internal dynamics are controllable through a boundary layer correction. The controller is composed of singular-perturbation based fast control and an outer-loop slow control. The slow subsystem is controlled by a neural network (NN) for feedback linearization, plus a PD outer-loop for tracking, and a robustifying term to assure the closed-loop stability. No off-line learning or training is needed for the NN. Tracking and stability are proven using Lyapunov techniques that yield a novel modified NN weight tuning algorithm.The research is supported by NSF grant IRI-9216545 and EPRI Grant RP8030-09.     

MRNN:一种新的基于改进型递归神经网络的WSN动态建模方法:应用于故障检测

提出了一种适用于无线传感器网络WSN的故障检测方法,该方法运用改进的递归神经网络MRNN为WSN的节点、节点的动态特性以及节点间的关系建立相关模型,对WSN节点进行识别和故障检测。MRNN的输入选择建模节点的先前输出值及其邻居节点的当前及先前输出值,模型基于一种新的改进的反向传播型神经网络,该神经网络的输入以及传感器网络的拓扑结构基于通用的非线性传感器模型。仿真实验将MRNN方法与卡尔曼滤波法进行了全面的比较。实验表明,MRNN在置信因子较小的情况下与卡尔曼滤波方法相比有较高的故障检测精度。     

An adaptive neural network switching control approach of robotic manipulators for trajectory tracking

In this paper, an adaptive neural network (NN) switching control strategy is proposed for the trajectory tracking problem of robotic manipulators. The proposed system comprises an adaptive switching neural controller and the associated robust compensation control law. Based on the Lyapunov stability theorem and average dwell-time approach, it is shown that the proposed control scheme can guarantee tracking performance of the robotic manipulators system, in the sense that all variables of the closed-loop system are bounded and the effect due to the external disturbance and approximate error of radical basis function (RBF) NNs on the tracking error can be converged to zero in an infinite time. Finally, simulation results on a two-link robotic manipulator show the feasibility and validity of the proposed control scheme.     

A novel Parzen probabilistic neural network based noncoherent detection algorithm for distributed ultra-wideband sensors

Ultra-wideband (UWB) has been widely recommended for significant commercial and military applications. However, the well-derived coherent structures for UWB signal detection are either computationally complex or hardware impractical in the presence of the intensive multipath propagations. In this article, based on the nonparametric Parzen window estimator and the probabilistic neural networks, we suggest a low-complexity and noncoherent UWB detector in the context of distributed wireless sensor networks (WSNs). A novel characteristic spectrum is firstly developed through a sequence of blind signal transforms. Then, from a pattern recognition perspective, four features are extracted from it to fully exploit the inherent property of UWB multipath signals. The established feature space is further mapped into a two-dimensional plane by feature combination in order to simplify algorithm complexity. Consequently, UWB signal detection is formulated to recognize the received patterns in this formed 2-D feature plane. With the excellent capability of fast convergence and parallel implementation, the Parzen Probabilistic Neural Network (PPNN) is introduced to estimate a posteriori probability of the developed patterns. Based on the underlying Bayesian rule of PPNN, the asymptotical optimal decision bound is finally determined in the feature plane. Numerical simulations also validate the advantages of our proposed algorithm.     

Semisupervised self-learning granular neural networks for remote sensing image classification

Accuracy of a pattern classification model mostly depends on ample number of training samples, which is the major bottleneck for classifying land cover of remote sensing images. Further, the unbalance scenario typically encountered in hyperspectral remote sensing images, i.e., limited number of training samples with more dimensions, makes the decision-making process cumbersome. Under such inevitable constraints, the article aims to develop an improved classification model using semisupervised self-learning granular neural networks (GNNs) for remote sensing images. The proposed semisupervised method has adopted a new strategy for selecting the potential candidate samples from the unlabeled dataset and used GNN as the base classifier. We have considered GNN because of its transparent architecture that leads to improved performance with less computational complexity compared to the conventional neural networks. Performance of the model is further enhanced with fuzzy granulation of features using class belonging information and selection of granulated features using neighborhood rough sets (NRS). The proposed model thus takes the mutual advantages of GNN architecture, fuzzy granulation with class belonging information, NRS-based feature selection and the most important, improved semisupervised self-learning approach. Performance of the model is compared with other similar methods and verified in terms of different performance measurement indexes, using two multispectral and two hyperspectral remote sensing images.     

A Cross-layer based mapping for spiking neural network onto network on chip

Abstract

Network-on-Chip provides a packet-based and scalable inter-connected structure for spiking neural networks. However, existing neural mapping methods just distribute all neurons of a population into an on-chip network core or nearby cores sequentially. As there is no connection among population, the population based mapping degrades inter-neuron communicating performance between different cores. This paper presents a Cross-LAyer based neural MaPping method that maps synaptic connected neurons belonging to adjacent layers into the same on-chip network node. In order to adapt to various input patterns, the strategy also takes input spike rate into consideration and remap neurons for improving mapping efficiency. The method helps to reduce inter-core communication cost. The experimental results demonstrate the efficient results of the proposed mapping strategy in the aspect of spike transfer latency as well as dynamic energy cost improvement. In the applications of handwritten digits and edge extraction, in which the type of interconnection among neurons is different, the neural mapping algorithm reduces spike average transfer latency by maximum 42.83%, and reduces dynamic energy by maximum 36.29%.