Development of an adaptive fuzzy logic-based inverse dynamic model for laser cladding process

Learning Controller Design using Fuzzy Modeling, Sliding Mode Control, and PID Controller for Robots

“…whereˆ( , , , ) F q q q t is the known part (the approximated part) of the manipulator inverse dynamic model and can be approximated using fuzzy modeling method [3,4]. Fuzzy logic is capable of modeling vagueness, which cannot be described by precise mathematical models; of handling uncertainty; and of supporting human-type reasoning.…”

“…Therefore, for an n-DOF manipulator, a MISO fuzzy model for joint k, (k = 1, …, n) expresses the variation of that joint's torque/force, as a result of the motion of all joints, in the following form of the rules Parameter identification which, in this paper, consists of identification of the optimum values of the parameters of the FCM clustering algorithm and the parameters of the reasoning mechanism, for more information please refer to Zeinal and Notash [3]. In this paper the methodology of fuzzy model construction from available input-output data is based on the improved systematic fuzzy modelling method reported in [3,4], through the following three general steps:…”

“…The controller proposed in this paper aims at building such a control system using a novel approach to combine adaptive fuzzy modeling algorithm, fuzzy c-means clustering algorithm [3,4], adaptive sliding mode control [5] and PID controller. The controller combines all the salient features of fuzzy modeling (FM), neural network modeling, sliding mode control (SMC), proportional integral derivative (PID), and repetitive control (the adaptive fuzzy model makes the controller a model-based learning or internal model learning control).…”

An experimentally derived hybrid intelligent tool for analysing and optimising the clad height and melt-pool depth in laser solid freeform fabrication process

“…Alimardani and Toyserkani (2008) used adaptive neuro-fuzzy inference system to dynamically predict the clad height and coating thickness as a function of laser pulse energy, laser pulse frequency and traverse speed. Zeinali and Khajepour (2010) developed an adaptive fuzzy-logic-based inverse dynamic model to predict the scanning speed as a function of the cladding parameters in particular the clad height. Mozaffari, Fathi, Khajepour, and Toyserkani (2013) proposed a hybrid self-learning evolutionary technique for predicting the melt pool geometry and optimising the LSFF process.…”

Deposição de materiais metálicos em pó a laser baseado em controle por visão computacional e aprendizagem de máquina

“…2,[46][47] Os sistemas de deposição mais comuns, geralmente, contemplam controle de parâmetros pré-definidos, não realizando ajustes automáticos no decorrer do processo (malha aberta Outras abordagens mais modernas incluem, por exemplo, utilizar a análise de espectro do plasma emitido pelo laser durante sua interação com a superfície do material para monitorar o foco, obtendo, por exemplo, informações como o seu posicionamento e intensidade. 51,52 Há, ainda, técnicas explorando modelos computacionais, 30,40 escaneamento a laser, 43 visão computacional, 28 inteligência artificial [61][62][63][64] e até mesmo sistemas híbridos que fazem uso de vários desses métodos elencados anteriormente. [25][26]…”