Cerebellar Model Articulation Controller Learning Research Articles

Basic Theory and Practice Teaching Method Based on the Cerebellar Model Articulation Controller Learning Algorithm

With the continuous in-depth development of science and technology in my country, the field of artificial intelligence has also made considerable progress. For example, the comprehensiveness and operability of technology have been enhanced. Artificial intelligence, as a comprehensive and intersecting research field in the fields of computer science, information science, and mathematics, has also made a certain contribution to the progress of international students in colleges and universities. This article is aimed at studying the problems based on basic theories and time teaching methods, mainly through a comprehensive data analysis of artificial intelligence technology and fusion differential evolution algorithm, and then improving the learning ability of students and comprehensive quality. This paper proposes a swarm intelligence algorithm, K -means clustering algorithm, and vector machine algorithm and summarizes the process and application fields of the swarm intelligence algorithm. In addition, the CMAC learning algorithm is proposed, which shows that robots under the background of industrial intelligence and machine learning can organize themselves to form groups to perform complex tasks. The experimental results of this paper show that the communication range is between 0 and 1, which can enable more robots to communicate in order to improve the individual’s social learning opportunities, thereby enhancing the individual’s learning ability to improve the performance of learning behavior and improving the efficiency of practical teaching and research, making the efficiency reach 0.9.

A fuzzy CMAC learning approach to image based visual servoing system

This study presents a fuzzy robotic joint controller using a cerebellar model articulation controller (CMAC) integrating a Takagi-Sugeno (T-S) framework with an online compensator for an articulated manipulator. The proposed controller is applied to image-based visual servoing (IBVS), including closed-loop feedback control and the kinematic Jacobian calculation. This approach learns a mapping from image feature errors for each joint’s velocity instead of the classical kinematics, thereby reducing the computational complexity and improving the self-regulation ability of the control system. These connecting weights of the cerebellar model learn offline, and an online compensator that uses reinforcement learning is developed to resolve system noise and uncertainties in an unknown environment. Compared with the classical inverse kinematics model, this approach does not need an excessive computational expense so that this proportional controller can be implemented in general scenarios with an eye-in-hand configuration. Experimental results show the proposed method can outperform the classical IBVS controller.

CMAC-based previous step supervisory control schemes for relaxing bound in adaptive fuzzy control

In this paper, a novel scheme of incorporating a learning mechanism into previous step supervisory controllers for adaptive fuzzy control is proposed to relax bounds required in the control process. In traditional supervisory adaptive fuzzy control approaches, the use of fuzzy estimators for approximating system functions and a robust supervisory control law are necessary to deal with any possible uncertainties caused in the system. This kind of supervisory controller depends on the robust bounds of system functions so that it can ensure the Lyapunov stability of controlled systems. However, in those approaches, the output may not be able to follow the reference trajectory well if the robust bounds are predicted improperly. In our implementation, CMAC (Cerebellar Model Articulation Controllers) is used as the learning mechanism because of its quick learning capability. Under the Lyapunov stable criterion, the proposed CMAC learning mechanism can improve the output performance and can relax the robust bound limitation so that practical systems can easily be realized. In summary, the proposed approach not only can relax bounds for previous step supervisory controllers in adaptive fuzzy control, but also can significantly improve the control performance of the system.

Tikhonov training of the CMAC neural network

The architecture of the cerebellar model articulation controller (CMAC) presents a rigid compromise between learning and generalization. In the presence of a sparse training dataset, this limitation manifestly causes overfitting, a drawback that is not overcome by current training algorithms. This paper proposes a novel training framework founded on the Tikhonov regularization, which relates to the minimization of the power of the sigma-order derivative. This smoothness criterion yields to an internal cell-interaction mechanism that increases the generalization beyond the degree hardcoded in the CMAC architecture while preserving the potential CMAC learning capabilities. The resulting training mechanism, which proves to be simple and computationally efficient, is deduced from a rigorous theoretical study. The performance of the new training framework is validated against comparative benchmarks from the DELVE environment.

Robust and fast learning for fuzzy cerebellar model articulation controllers

In this paper, the online learning capability and the robust property for the learning algorithms of cerebellar model articulation controllers (CMAC) are discussed. Both the traditional CMAC and fuzzy CMAC are considered. In the study, we find a way of embeding the idea of M-estimators into the CMAC learning algorithms to provide the robust property against outliers existing in training data. An annealing schedule is also adopted for the learning constant to fulfill robust learning. In the study, we also extend our previous work of adopting the credit assignment idea into CMAC learning to provide fast learning for fuzzy CMAC. From demonstrated examples, it is clearly evident that the proposed algorithm indeed has faster and more robust learning. In our study, we then employ the proposed CMAC for an online learning control scheme used in the literature. In the implementation, we also propose to use a tuning parameter instead of a fixed constant to achieve both online learning and fine-tuning effects. The simulation results indeed show the effectiveness of the proposed approaches.

Improving Conventional Longitudinal Missile Autopilot Using Cerebellar Model Articulation Controller Neural Networks

A way to integrate the cerebellar model articulation controller (CMAC) neural network with the missile longitudinal conventional feedback controller (CFC) to compensate for nonlinearities, unmodeled dynamics, parameter variations, etc., is proposed. The inner loop of the CFC will essentially be left unchanged to improve the stability of the missile. The outer loop of CFC, in addition to playing its traditional role, would work with the CMAC to learn quickly to approximate the dynamic inversion from angle of attack to control deflection, to achieve better tracking in normal acceleration. In this arrangement, the well-known CFC acts as a safety net, whereas additional performance is brought about through CMAC learning.

Credit assigned CMAC and its application to online learning robust controllers

In this paper, a novel learning scheme is proposed to speed up the learning process in cerebellar model articulation controllers (CMAC). In the conventional CMAC learning scheme, the correct numbers of errors are equally distributed into all addressed hypercubes, regardless of the credibility of the hypercubes. The proposed learning approach uses the inverse of learned times of the addressed hypercubes as the credibility (confidence) of the learned values, resulting in learning speed becoming very fast. To further demonstrate online learning capability of the proposed credit assigned CMAC learning scheme, this paper also presents a learning robust controller that can actually learn online. Based on robust controllers presented in the literature, the proposed online learning robust controller uses previous control input, current output acceleration, and current desired output as the state to define the nominal effective moment of the system from the CMAC table. An initial trial mechanism for the early learning stage is also proposed. With our proposed credit-assigned CMAC, the robust learning controller can accurately trace various trajectories online.

A General Learning Scheme for CMAC-based Controller

Cerebellar model articulation controller (CMAC) is a powerful tool for nonlinear control applications. However, it yet lacks an adequate learning scheme. It is found that, with the existing learning scheme, if a complicated learning algorithm is not used, CMAC can destabilize a system that is otherwise stable. Oscillations resulting from the interaction between CMAC and the classical controller were found to contribute to the instability. This paper presents a new CMAC learning scheme that models plant's characteristics based on closed loop errors instead of the original input-output pairs. In this scheme, memory space of the CMAC is partitioned into two parts. One is for dynamic control, in which dynamic information is stored. Another is for steady state control, in which steady state information is adaptively updated for smooth control. Relationship between the two parts of the space is discussed and specified for a stable control. Simulation results on a typical nonlinear plant model and a real electrohydraulic servo system using the proposed scheme demonstrate that the oscillations are eliminated and stable control is obtained. The new scheme demonstrates superior tracking performance, noise rejection property and good robustness.

Learning Convergence and Stability of CMAC-Based Robot Control

ABSTRACTLearning control algorithms based on the Cerebellar Model Articulation Controller (CMAC) have been successfully applied to control non-linear systems in the past. However, it lacked on rigorous mathematical proofs that the closed-loop control system based on CMAC are stable. This paper presents a stability analysis of a CMAC-based learning control structure. In the feedforward loop a CMAC learns the inverse dynamics of a walking machine leg and a Proportional Derivative type feedback controller reduces the remaining errors. By utilizing the Liapunov stability theorem and La Salle's Extension we have proved analytically that the CMAC learning is convergent for the regulation control and the learning control system is asymptotically stable. By using computer simulation we have shown that the CMAC-based control is stable for the trajectory tracking control.

The learning convergence of CMAC in cyclic learning

In this paper we discuss the learning convergence of the cerebellar model articulation controller (CMAC) in cyclic learning. We prove the following results. First, if the training samples are noiseless, the training algorithm converges if and only if the learning rate is chosen from (0,2). Second, when the training samples have noises, the learning algorithm will converge with a probability of one if the learning rate is dynamically decreased. Third, in the case with noises, with a small but fixed learning rate e the mean square error of the weight sequences generated by the CMAC learning algorithm will be bounded byO(e). Some simulation experiments are carried out to test these results.

Learning convergence in the cerebellar model articulation controller

A new way to look at the learning algorithm in the cerebellar model articulation controller (CMAC) proposed by J.S. Albus (1975) is presented. A proof that the CMAC learning always converges with arbitrary accuracy on any set of training data is obtained. An alternative way to implement CMAC based on the insights obtained in the process is proposed. The scheme is tested with a computer simulation for learning the inverse dynamics of a two-link robot arm.