In this paper, we examine the problem of containment control for linear heterogeneous multi-agent systems with fuzzy uncertainty. Fuzzy uncertainty is considered as granular fuzzy numbers in the parameters of the agents. The design of the controller for each follower is based on the state feedback control approach and the output regulation framework. We also introduce a compensating dynamic for each follower. The leaders are considered as exosystems with fuzzy uncertainty. We use a new formulation for the containment error that guarantees that the followers converge to the fuzzy convex hull spanned by the leaders. Finally, we give a numerical example to show validity of the obtained theoretical results.
In recent years, several industries have widely applied BLDC servomotors for various purposes that they almost need a control system to be accurate and robust. In this paper, we compare an adaptive fuzzy controller and PI controller for BLDC servomotor position control in variable loads. The PI controller is tuned by both the trial-and-error and, Ziegler-Nichols methods. These controllers are implemented to servo system in simulation and physical system real-time with a variable loaded BLDC Motor. The obtained results in simulation and experiment indicate that adaptive fuzzy controller has much better behaviors than PI controller in terms of accuracy and reliability in presence of uncertainties.
It is always very challenging to obtain a unique solution for the forward kinematics of parallel robots due to nonlinear and coupled equations. Soft computing is one of the most widely used methods in identifying systems. In the present paper, some different supervised clustering methods are applied to obtain a model for the forward kinematics of the 3-DoF delta parallel robot. In the proposed process, a hybrid of fuzzy logic and neural networks is used for modeling. In this paper, the Subtractive Clustering Method (SCM), Fuzzy C-means (FCM), and Grid Partitioning (GP) techniques are described, and the simulation results obtained from the explained approaches are examined in detail. Finally, it concluded that the (GP) method for this application provides better performance than (FCM) and (SCM) techniques.
Soft robots, due to their flexible characteristics, have the required potential to communicate and interact with soft, sensitive and living tissues. Therefore, static modeling and behavior analysis of these actuators in a gripper form are studied in this paper. For this purpose, the large deflection theory of curved beams is utilized and the effect of external forces caused by external objects is taken into account. Finally, the shape of the actuator and grasping results are validated by finite element analysis.
Medical image classification plays a crucial role in monitoring and detecting diseases. This paper presents deep learning methods to distinguish images taken by the Optical Coherence Tomography technique from the normal eye and three eye-related diseases named Diabetic Macular Edema (DME), Choroidal neovascularization (CNV), and DURSEN. To achieve this aim, the images undergo a patch extraction process; then, the extracted patches are treated as sequences, and Recurrent Neural Networks are implemented to classify the images. Four pre-trained models, including VGG16, ResNet152V2, NasnetMobile, and Densnet169, and a vision transformer model are also applied and compared. Based on the results, The proposed model has achieved 99.38% test accuracy, higher than other models.
