Thin-walled parts, especially those with complex surface geometries, are prone to geometric inaccuracies and structural distortions, which can lead to issues during finishing operations like sanding. Using a toolpath based on the reference part to machine deformed workpieces may compromise surface quality or even damage the machined surface. In robotic applications, force control is often used to adapt the toolpath, but its ability to compensate for tool orientation remains limited. Digitizing the part allows for toolpath and tool orientation corrections before machining, preventing surface defects caused by workpiece deviations. This paper proposes an algorithm for toolpath and tool orientation correction, with the developed software solution simulated on a virtual twin of the robot control unit.

The variations in the physical and mechanical properties of different types of stones have a significant impact on the dynamics of the machining process. Hardness stands out as one of the most crucial properties of stone, directly affecting cutting forces and, consequently, the stability of tools and/or workpieces. Therefore, an analysis was conducted on the application of cutting forces and servomotor current signals of the main spindle and feed drives of the drilling machine in the classification of stone hardness. Stone samples were drilled using a newly designed drill bit with internally cooling capabilities. Laboratory testing with this type of drill bit revealed notably lower cutting forces and reduced tool wear intensity. In comparison to experiments conducted with commercially available industrial drill bits, the analyzed signals exhibited significantly improved accuracy in stone hardness classification, surpassing 98%. This high precision is evidently attributed to the absence of adverse effects caused by stone particles in the cutting zone, effectively eliminated by the analyzed internally cooled drill bit.

In this paper, an improved Levenberg-Marquardt-based feedforward neural network, with variable weight decay, is suggested. Furthermore, parallel implementation of the network on graphics processing unit is presented. Parallelization of the network is achieved on two different levels. First level of parallelism is data set level, where parallelization is possible due to inherently parallel structure of the feedforward neural networks. Second level of parallelism is Jacobian computation level. Third level of parallelism, i.e. parallelization of optimization search steps, is not implemented due to the variable weight decay, which makes third level of parallelism redundant. Suggested weight decay variation enables the compromise between higher accuracy with oscillations on one side and stable, but slower convergence on the other. To improve learning speed and efficiency, modification of random weight initialization is included. Testing of proposed algorithm is performed on two real domain benchmark problems. The results obtained and presented in this paper show effectiveness of proposed algorithm implementation.

Forecasting performances of feed-forward and recurrent neural networks (NN) trained with different learning algorithms are analyzed and compared using the Mackey-Glass nonlinear chaotic time series. This system is a known benchmark test whose elements are hard to predict. Multi-layer Perceptron NN was chosen as a feed-forward neural network because it is still the most commonly used network in financial forecasting models. It is compared with the modified version of the so-called Dynamic Multi-layer Perceptron NN characterized with a dynamic neuron model, i.e., Auto Regressive Moving Average filter built into the hidden layer neurons. Thus, every hidden layer neuron has the ability to process previous values of its own activity together with new input signals. The obtained results indicate satisfactory forecasting characteristics of both networks. However, recurrent NN was more accurate in practically all tests using less number of hidden layer neurons than the feed-forward NN. This study once again confirmed a great effectiveness and potential of dynamic neural networks in modeling and predicting highly nonlinear processes. Their application in the design of financial forecasting models is therefore most recommended.

In this paper, reactive vehicle navigation in unknown environment with obstacles, using fuzzy controller, is presented. In order to prevail conventional fuzzy controller drawbacks, analytical fuzzy controller is proposed. Reference trajectory is generated in real-time using potential field method. Vehicle dynamics model is full nonlinear, eleventh order model, which includes three degrees of freedom, wheel dynamics and nonlinear tyre friction model. Controller capabilities are tested depending on different vehicle's initial velocity and different road conditions.

This work considers the application of radial basis function neural network (RBFNN) for tool wear determination in the milling process. Tool wear, i.e., flank wear zone widths, have been estimated in two phases using two types of RBFNN algorithms. In the first phase, RBFNN pattern recognition algorithm is used in order to classify tool wear features in three wear level classes (initial, normal and rapid tool wear). On behalf of these results, in the second phase, RBFNN regression algorithm is utilized to estimate the average amount of flank wear zone widths. Tool wear features were extracted in time and frequency domain from three different types of signals: force, acoustic emission and nominal currents of feed drives.

This paper presents the stability analysis of an analytic fuzzy PID controller for robot manipulators. The analytic fuzzy control is a nonconventional approach that uses an analytic function for output determination, instead of a fuzzy rule base. The stability analysis is based on Lyapunov's direct method and does not require representation of the plant dynamics in the form of Takagi-Sugenos's fuzzy model. The stability criterion, which ensures local asymptotic stability, is obtained. Finally, an example is given to demonstrate the obtained results.