China
Africa
Explore chapters and articles related to this topic
Robotic Skill Acquisition Based on Biological Principles
Published in Abraham Kandel, Gideon Langholz, Lotfi A. Zadeh, Hybrid Architectures for Intelligent Systems, 2020
David A. Handelman, Stephen H. Lane, Jack J. Gelfand
An RSA2net (Fig. 3b) contains fields characterizing its associated CMAC neural network [17–19]. Originally introduced by Albus as a model of the cerebellum, a CMAC module is a perceptron-like associative memory that is capable of learning multidimensional nonlinear functions over particular regions of the function space. CMACs learn by example, and map similar inputs to similar outputs, thereby providing automatic generalization (interpolation) among input/output pairs. Although CMACs can be implemented in highly parallel computer hardware for optimum execution speed, the CMAC algorithm also can be made to operate efficiently on traditional computing architectures.
Learning Based Classifiers
Published in Anil Kumar, Priyadarshi Upadhyay, A. Senthil Kumar, Fuzzy Machine Learning Algorithms for Remote Sensing Image Classification, 2020
Anil Kumar, Priyadarshi Upadhyay, A. Senthil Kumar
With more straightforward and convergent training algorithms, type of neural networks can be explored. One such kind of neural network is called CMAC – cerebellar model articulation controller. In CMAC, learning rates or randomly generated initial weights are not required. The training activity while applied in a batch can be guaranteed to converge, while arithmetical computation in training algorithm behaves linearly, dependent on number of neurons considered.
Intelligent Soft-Computing Techniques in Robotics
Published in Osita D. I. Nwokah, Yildirim Hurmuzlu, The Mechanical Systems Design Handbook, 2017
CMAC is an associative neural network using the feature that only a small part of the network influences any instantaneous output. The associative property built into CMAC enables local generalization; similar inputs produce similar outputs while distant inputs produce nearly independent outputs. As a result, we have fast convergence properties. It is very important that practical hardware realization using logical cell arrays exists today.
Adaptive output feedback control with cerebellar model articulation controller-based adaptive PFC and feedforward input
Published in SICE Journal of Control, Measurement, and System Integration, 2022
Nozomu Otakara, Kota Akaike, Sadaaki Kunimatsu, Ikuro Mizumoto
CMAC is well known as a type of neural network based on a mathematical model of the mammalian cerebellum. In the CMAC, inputs to the input space are transformed into the label set, and it outputs the average value of the weights by referring the weights in the activated cells with a distributed shared memory structure based on the input label. Since the CMAC is trained based on variables in the specified area, it is simple and might achieve quick learning and adjusting the weights in the activated cells through the error observed at the output. We utilize this CMAC strategy for adjusting parameters for the linearly approximated system in a specific domain. Thus, for a system satisfying Assumptions 3.1 and 3.2, in each subsets and , we adjust parameters of PFC and of FF control input via CMAC strategy.
Research on parallel control of CMAC and PD based on U model
Published in Automatika, 2021
Fengxia Xu, Junhua Xu, Jiaqi Zhang, Chunda Zhang, Zifei Wang
Based on the biological finding that the cerebellum makes reflexive responses without thinking when controlling limb movements, scholars have proposed a new type of neural network, named the cerebellar model neural network, or cerebellar model articulation controller (CMAC) network for short. CMAC networks have strong associative capabilities, and the output of the network is determined by only a small number of neurons corresponding to the network weights, the input and output of a CMAC network appears to be a linear relationship. Such characteristics make CMAC have fast learning ability, strong fault tolerance, fast convergence speed, no local minimal problems and other characteristics, so it is widely used in robot arm control, adaptive control, robot control, pattern recognition, signal processing and other fields. Adaptive fuzzy CMAC neural network controller for pneumatic artificial muscle-driven spring mass position control system [8]. In order to solve the control problem of uncertain nonlinear systems and the problem of system mixed interference upper bound that is difficult to measure in practical applications, a recursive CMAC neural network model decomposition control algorithm is proposed [9]. In order to compensate the hysteresis nonlinearity inherent in the telescopic actuator caused by the super magnetism and to improve its accuracy, a real-time hysteresis compensation control strategy is proposed by combining CMAC neural network and proportional integral derivative (PID) control to achieve high-precision tracking control [10]. In view of the nonlinear, large inertia and time-varying characteristics of the temperature control system of central air-conditioning room, the composite control of CMAC neural network and single neuron PID was proposed [11]. A nonlinear quantization-based CMAC neural network algorithm is proposed, which adaptively designs the conceptual mapping of CMAC and improves the computational speed and accuracy of CMAC to meet the needs of nonlinear real-time control in complex dynamic environments [12]. To address the nonlinearity and various uncertainty factors in high-precision servo systems, the fast learning of PD + feed-forward control + CMAC neural network algorithm is proposed, which ensures fast real-time tracking and further improves tracking accuracy [13]. Taking electro-hydraulic servo system as the control object, the control strategy of combining CMAC network and PID controller is discussed. The CMAC neural network is used as a feedforward controller to achieve the inverse dynamic control [14]. To address the problem of poor control effect of temperature control system in metal heat treatment process, a compound control algorithm based on CMAC and PD is proposed to realize the fast tuning and self-learning function of PID parameters [15].