基于ANFIS的非线性系统辨识研究

系统辨识是控制系统设计的基础，对非线性系统进行辨识是当前的难点；文献[1]提出了用模糊建模方法，文献[2]提出了用神经网络方法，在总结上述方法不足的基础上，该文提出了用自适应神经模糊推理系统(ANFIS)对非线性系统进行辨识的方法，仿真结果表明，ANFIS进行非线性系统辨识是可行的，其辨识精度很高。     

基于改进遗传算法的神经网络模型辨识

利用前向神经网络对非线性动态系统建模时存在着很大的缺陷，因此提出采用递归网络(RNN)对非线性动态系统建模。并在权值的修正上，摆脱常用BP算法的束缚，采用改进的遗传算法搜索最优权值。最后对一高阶非线性系统进行建模仿真实验，结果表明该方法是有效的。     

模糊判决支持向量机在自动语种辨识中的研究

支持向量机(SVM)是在统计学习理论的基础上发展起来的一种新的通用学习方法。自动语种辨识是语音信号处理中新出现的分支,也是一项较难的课题。该文提出的模糊判决支持向量机(FDSVM)是对支持向量机的判决结果的合理化改进,并应用于自动语种辨识系统。利用OGI-TS电话语音库对新算法的性能进行测试,然后给出实验结果。结果表明,该算法相对于传统算法是一种更有效的方法。     

Experimental validation of a resilient monitoring and control system

Complex, high performance, engineering systems have to be closely monitored and controlled to ensure safe operation and protect public from potential hazards. One of the main challenges in designing monitoring and control algorithms for these systems is that sensors and actuators may be malfunctioning due to malicious or natural causes. To address this challenge, this paper addresses a resilient monitoring and control (ReMAC) system by expanding previously developed resilient condition assessment monitoring systems and Kalman filter-based diagnostic methods and integrating them with a supervisory controller developed here. While the monitoring and diagnostic algorithms assess plant cyber and physical health conditions, the supervisory controller selects, from a set of candidates, the best controller based on the current plant health assessments. To experimentally demonstrate its enhanced performance, the developed ReMAC system is then used for monitoring and control of a chemical reactor with a water cooling system in a hardware-in-the-loop setting, where the reactor is computer simulated and the water cooling system is implemented by a machine condition monitoring testbed at Idaho National Laboratory. Results show that the ReMAC system is able to make correct plant health assessments despite sensor malfunctioning due to cyber attacks and make decisions that achieve best control actions despite possible actuator malfunctioning. Monitoring challenges caused by mismatches between assumed system component models and actual measurements are also identified for future work.     

Hierarchical backstepping-based control of a Gun Launched MAV in crosswinds: Theory and experiment

This paper considers the problem of controlling the position and the orientation of a Gun Launched Micro Aerial Vehicle – GLMAV – despite unknown aerodynamic efforts. The proposed approach overcomes the problem of gyroscopic coupling by taking advantage from the structure of the thrust mechanism, which is made up of two counter rotating propellers. An adaptive hierarchical controller is designed, allowing the trajectory tracking and the stabilization of the vehicle's position and orientation while the unknown aerodynamic efforts are estimated by means of an identifier. The overall process is shown to be stable for constant, or slowly time varying, aerodynamic efforts. However numerical simulations demonstrate the satisfying controller's performance even with nonconstant aerodynamic efforts. Experimental results are also provided.     

Near-optimal state feedback design for singularly perturbed systems by unified approach

A near optimal state feedback design for singularly perturbed systems by a unified approach using the delta operator is presented with an example of the aircraft longitudinal motion. The main contribution of this paper is to explore the use of the i -operator that has attracted a new attention in systems science. The i -operator system unifies the continuous system and the discrete system together without loosing any characteristics of both systems. The paper offers the following; Finite-word-length-characteristics are improved using the i -operator. Floating-point-operations are reduced by block diagonalization and by time-invariant optimal feedback gain from the algebraic Riccati equation. The results of adopting those approaches are illustrated in the simulation figures and compared with the earlier one.     

A deep learning-based method for detecting non-certified work on construction sites

The construction industry is a high hazard industry. Accidents frequently occur, and part of them are closely relate to workers who are not certified to carry out specific work. Although workers without a trade certificate are restricted entry to construction sites, few ad-hoc approaches have been commonly employed to check if a worker is carrying out the work for which they are certificated. This paper proposes a novel framework to check whether a site worker is working within the constraints of their certification. Our framework comprises key video clips extraction, trade recognition and worker competency evaluation. Trade recognition is a new proposed method through analyzing the dynamic spatiotemporal relevance between workers and non-worker objects. We also improved the identification results by analyzing, comparing, and matching multiple face images of each worker obtained from videos. The experimental results demonstrate the reliability and accuracy of our deep learning-based method to detect workers who are carrying out work for which they are not certified to facilitate safety inspection and supervision.     

The Identification Problem for a Mixture of Observations from Two Normal Populations

A comparative study is made of the performances of various procedures for identifying the population of origin of each observation in a sample made up of observations from two normal populations, the parameters of which may or may not be known.     

工业机器人运动学逆向建模

针对工业机器人在缺乏运动学参数的情况下,进行机器人运动学逆向建模的研究,给出了机器人机座坐标系位置及杆件参数的识别方法,建立了机器人的运动学方程。并分析了计算的末端位姿与示教器显示末端位姿存在的偏差,指出从机器人示教器上读出的关节转角数据存在舍入误差,采用遗传算法分别以位置误差和位姿误差为目标函数进行了辨识。辨识结果表明,以位姿误差为目标函数来辨识关节转角误差的辨识结果是准确的。试验结果表明,提出的机器人运动学逆向建模方法是有效的。     

FUZZIFYING THE “PRECISE”

Based on the fuzzy-integral model, methods and algorithms are developed for identifying the “input–output” operator of continuous and stationary discrete extremal fuzzy dynamic systems (EFDS). The EFDS “input–output” operator is restored by means of experimental data with possibilistic uncertainty, the source of which is extremal fuzzy time intervals. The regularization conditions for obtaining quasi-optimal estimates are substantiated by the proved theorems. The corresponding algorithms are provided. The results obtained are illustrated by examples in the case of a finite set of EFDS states.