Antlion Algorithm Optimized Fuzzy PID Supervised On-line Recurrent Fuzzy Neural Network Based Controller for Brushless DC Motor

In this paper, Antlion algorithm optimized Fuzzy PID supervised on-line Recurrent Fuzzy Neural Network based controller is proposed for the speed control of Brushless DC motor. Learning parameters of the supervised on-line recurrent fuzzy neural network controller, i.e., learning rate (η), dynamic factor (α), and number nodes (N_i) are optimized using Genetic algorithm, Particle Swarm optimization, Ant colony optimization, Bat algorithm, and Antlion algorithm. The proposed controller is tested with different operating conditions of the Brushless DC motor, such as varying load conditions and varying set speed conditions. The time domain specifications such as rise time, overshoot, undershoot, settling time, recovery time, and steady state error and also integral performance indices such as root mean square error, integral of absolute error, integral of squared error, and integral of time multiplied absolute error are measured and compared for above optimized controller. Simulation results show Antlion algorithm optimized Fuzzy PID supervised on-line recurrent fuzzy neural network based controller has proved to be superior than other considered controllers in all aspects. In addition, the experimental verification of proposed control system is presented to test the effectiveness of the proposed controller with different operating conditions of the Brushless DC motor.     

基于量子加权多层级GRU神经网络的综合能源 系统多元负荷短期预测

针对综合能源系统多元负荷短期预测问题,提出一种基于量子加权多层级GRU(quantum weighted multi hierarchy gated recurrent unit, QWMHGRU)神经网络的多元负荷短期预测模型。采用最大信息系数对多元负荷间和负荷与天气因素间的相关性进行分析,选取模型输入量。然后改进GRU的门控结构,形成多层级门控循环单元(multi hierarchy gated recurrent unit, MHGRU),并将量子加权神经元引入MHGRU,构成QWMHGRU多变量负荷预测模型。仿真算例结果表明,QWMHGRU多元负荷预测模型在夏季和冬季的权重平均精度均可达97%以上,相比MHGRU、QWGRU和GRU模型具有更高的预测精度。     

一种混沌神经网络的混沌时间序列预测

混沌是一种普遍存在的非线性动力学行为,针对混沌时间序列预测问题,提出了一种新颖的混沌对角递归神经网络模型,然后,给出了易实现的动量梯度学习算法。为了验证网络的预测性能,采用该神经网络预测模型对Mackey—Glass混沌时间序列进行了仿真。结果表明,所提出的混沌神经神经网络模型能快速、精确地预测混沌时间序列,并能在一定精度上满足多步预测需要,是研究复杂非线性动力系统辨识和控制的一种有效方法。     

基于VMD与改进QRGRU的超短期风电功率概率预测

风电功率概率预测是分析未来风电功率不确定性的有效方法之一。为提高风电功率概率预测精度,文中提出基于变分模态分解(VMD)与改进门控循环单元分位数回归(QRGRU)的超短期风电功率概率预测方法。首先,采用VMD将原始风电功率序列分解成不同特征的模态函数;然后,对每个模态函数分别建立基于QRGRU的概率预测模型,并将变量间的网络结构约束作为目标函数的惩罚项,改进QRGRU权重在迭代修正过程中的平稳性;最后,在不同分位数条件下叠加各个模态函数预测值,并采用非参数核密度估计方法得到未来风电功率的概率密度函数。结合某风电场实测数据开展具体算例分析,结果表明所提方法能够兼顾区间覆盖率,减少区间宽度,在不同预测步长中均能表现较好的预测效果。     

基于CAPSO-RNN的光伏系统短期发电量预测

针对光伏系统的发电特性及影响光伏发电的因素,建立基于混沌自适应粒子群优化算法的反馈型神经网络短期发电量预测模型。该预测模型利用混沌自适应粒子群优化算法的全局优化能力初始化反馈性神经网络权值和阈值,可以克服反馈型神经网络收敛速度慢俄且易陷于局部最优等缺点。同时为提高预测精度,采用隶属度函数对温度进行模糊化处理。预测结果表明,建立的预测模型具有较高的精度。     

采用注意力门控卷积网络模型的目标情感分析

针对传统目标情感分析采用循环神经网络模型导致训练时间长且其他替代模型未能使得上下文和目标词之间实现良好交互等问题,提出了一种用于目标情感分析的注意力门控卷积网络模型。该模型首先将上下文和目标词通过多头注意力机制加强上下文和目标词之间的交互;其次采用门控卷积机制进一步提取关于目标词的情感特征;最后通过Softmax分类器将情感特征进行分类,输出情感极性。实验结果显示,与循环神经网络模型中准确率最高的循环注意力网络模型相比,在SemEval 2014任务四的餐厅和笔记本电脑数据集上的准确率分别提高了1.29%和0.12%;与循环神经网络模型中收敛速度较快的基于注意力的长短期记忆网络模型相比,收敛时间下降了约29.17s。     

轮询多址通信系统的门限服务分析方法

针对多个用户共享一个统计复用的系统时,在轮询多址通信系统的门限服务中,必须采用适当的服务规则（如预约或轮询）的问题。采用嵌入马尔可夫链理论和概率母函数的方法,对轮询排队系统的门限服务用两种不同的分析方法进行分析,得到了相同的门限服务排队系统的平均排队队长和平均循环周期的解析结果。计算机模拟结果表明了仿真与理论的一致性。     

Navigation behavior based on self-organized spatial representation in hierarchical recurrent neural network

ABSTRACT

A cognitive map is an internal model of the external world and contains the spatial representation of the surrounding environment. The existence of the cognitive map was first identified in rats; rats can navigate to their desired destination using cognitive maps while dealing with environmental uncertainty. We performed a mobile robot navigation experiment where obstacles were randomly placed using hierarchical recurrent neural network (HRNN) with multiple timescales. The HRNN was trained to navigate the mobile robot to the destination indicated by a snapshot image. After the training, the HRNN was able to successfully avoid the obstacles and navigate to the destination from any location in the environment. Analysis of the internal states of the HRNN showed that the module with fast timescale handles obstacle avoidance and the one with slow timescale has spatial representation corresponding to the spatial position of the destination. Moreover, in the experiment wherein the novel path appeared, the trained HRNN performed shortcut behavior. The shortcut behavior shows that the HRNN performed navigation using the self-organized spatial representation in the slow recurrent neural network. This indicates that training of goal-oriented navigation, i.e. the navigation motivated by a snapshot image of the destination results in the self-organization of cognitive map-like representation.     

Fabrication of a planar water gated organic field effect transistor using a hydrophilic polythiophene for improved digital inverter performance

A planar water gated OFET (WG-OFET) structure is fabricated by patterning gate, source and drain electrodes on the same plane at the same time. Transistor output characteristics of this novel structure employing commercial regioregular poly(3-hexylthiophene) (rr-P3HT) as polymer semiconductor and deionized (DI) water as gate dielectric show successful field effect transistor operation with an on–off current ratio of 43 A/A and transconductance of 2.5 μA/V. These output characteristics are improved using P3HT functionalized with poly(ethylene glycol) (PEG) (P3HT-co-P3PEGT), which is more hydrophilic, leading to on–off ratio of 130 A/A and transconductance of 3.9 μA/V. Utilization of 100 mM NaCl solution instead of DI water significantly increases the on–off ratio to 141 A/A and transconductance to 7.17 μA/V for commercial P3HT and to 217 A/A and to 11.9 μA/V for P3HT-co-P3PEGT. Finally, transistors with improved transconductances are used to build digital inverters with improved characteristics. Gain of the inverters employing P3HT and P3HT-co-P3PEGT are measured as 2.9 V/V and 10.3 V/V, respectively, with 100 mM NaCl solution.