The neural based sliding-mode control method is developed by using advantageous specifications of sliding-mode control and artificial neural network techniques and applied to an eight storey sample building with active tendon. Success of the designed controller is examined by applying acceleration data belonging to six different earthquakes as the external driving source. A MATLAB package program is used for numerical solutions, and the obtained results are compared in graphical form, presented in tables. For the optimization process, genetic algorithm is used. In this study parametric uncertainties and time delay effects are also considered. It is seen that the success of the controller is quite high and the control force can be used for real applications. Furthermore, the obtained results indicate that the controller provides quite successful control under earthquake having different characteristics.

Purpose: The purpose of the study is to design a three-dimensional (3D) triglide parallel robot with a different approach and to control the manufactured robot via sliding mode control method that has not been applied to the robot before. Design/methodology/approach: The x, y and z coordinates of the end effector of the robot have been given as a reference. The x, y and z reference values are transformed as new reference values of the vertical movement of the robot on the endless screw by using the inverse kinematic equations of the robot. The control of the robot over these reference values is provided by a sliding mode control. The MATLAB/real-time toolbox has been used for creating the interface. The real-time control of the triglide robot has been carried out with a sliding mode controller in the Simulink environment. Findings: When the results of the sliding mode control are examined, it is seen that the desired reference values are provided in about 0.6 s. The velocity of the sliding limbs in each arm of the robot is approximately 50 mm/s. The reference values have been reached using the sliding mode control method, with an average error of 0.01 mm. In addition, the problem of chattering in the system caused by using the sign function has been relatively eliminated by using the saturation function instead of the sign function. Thus, the sliding mode control method with saturation function is more feasible. Originality/value: In this study, the triglide parallel robot was manufactured using a 3D model after taking into consideration the dimensions of the 3D model. After production, the necessary hardware connections were provided, and a real-time sliding mode control method was implemented to the robot by using the interface program in MATLAB/Simulink environment. The literature contribution of the paper is the real-time control of the triglide robot with the sliding mode control method.

In this study, a deep learning-based multimodal framework is presented for forest fire detection using RGB images, which synthetically generates night-vision-like, white-hot, and green-hot pseudo-thermal representations. The synthetic modalities are derived directly from RGB data and integrated into a hardware-independent multimodal learning pipeline to increase visual diversity without relying on additional sensing hardware. Each modality is processed using an ImageNet-pretrained convolutional backbone, and modality-specific feature vectors are combined through feature-level concatenation before classification. The proposed framework was evaluated using multiple backbone architectures, including ResNet18, EfficientNet-B0, and DenseNet121, which were assessed independently under a unified experimental protocol. Experiments were conducted on two datasets with substantially different scales and characteristics: the FLAME dataset (39,375 images, binary classification) and the FireStage dataset (791 images, three-class classification). For both datasets, stratified 80–20% training–validation splits were employed, and online stochastic data augmentation was applied exclusively to the training sets. On the FLAME dataset, the proposed framework achieved consistently high performance across different backbone and modality configurations. The best-performing models reached an accuracy of 99.66%, precision of 99.80%, recall of 99.66%, F1-score of 99.73%, and ROC AUC value of 0.9998. On the more challenging FireStage dataset, the framework demonstrated stable performance despite limited data availability, achieving an accuracy of 93.71% for RGB-only configurations and up to 93.08% for selected multimodal combinations, while macro-averaged F1-scores exceeded 0.92, and ROC AUC values reached up to 0.9919. Per-class analysis further indicates that early-stage fire (Start Fire) patterns can be discriminated, achieving ROC AUC values above 0.96, depending on the backbone and modality combination. Overall, the results suggest that synthetic-modality-based multimodal learning can provide competitive performance for both large-scale and data-limited fire detection scenarios, offering a flexible and hardware-independent alternative for forest fire monitoring applications.