The field of robotics has grown a lot over the years due to the increasing necessity of industrial production and the search for quality of industrialized products. The identification of a system requires that the model output be as close as possible to the real one, in order to improve the control system. Some hybrid identification methods can improve model estimation through computational intelligence techniques, mainly improving the limitations of a given linear technique. This paper presents as a main contribution a hybrid algorithm for the identification of industrial robotic manipulators based on the recursive least square (RLS) method, which has its matrix of regressors and vector of parameters optimized via the Kalman filter (KF) method (RLS-KF). It is also possible to highlight other contributions, which are the identification of a robotic joint driven by a three-phase induction motor, the comparison of the RLS-KF algorithm with RLS and extended recursive least square (ERLS) and the generation of the transfer function by each method. The results are compared with the well-known recursive least squares and extended recursive least squares considering the criteria of adjustable coefficient of determination ( R _a ² ) and computational cost. The RLS-KF showed better results compared to the other two algorithms (RLS and ERLS). All methods have generated their respective transfer functions.

The use of industrial robots has grown over the years, making production systems increasingly efficient. Within this context, some limitations appear that can delay the productive process causing damages to the production. These limitations are robot stops, for example. Stops can be caused by various factors, such as accidents, collisions of manipulator robots with operators or other equipment. The main contribution of this research is to improve the Artificial Potential Field (APF) using Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Differential Evolution (DE) by optimizing the APF parameters in collision avoidance. We present as results: the trajectories generated by a planar manipulator robot; the position errors between the final position and the last position of the generated trajectories; and the computational cost of the PSO, GA and DE algorithms to find the parameters of the APF algorithm.

Solar energy is one of the main renewable energy sources capable of contributing to global energy demand. However, the solar resource is intermittent, making its integration into the electrical system a difficult task. Here, we present and compare two machine learning techniques, deep learning (DL) and support vector regression (SVR), to verify their behavior for solar forecasting. Our testing from Spain showed that the mean absolute percentage error for predictions using DL and SVR is 7.9% and 8.52%, respectively. The DL achieved the best results for solar energy forecast, but it is worth mentioning that the SVR also obtained satisfactory results.

This paper presents the identification of the inverse kinematics of a cylindrical manipulator using identification techniques of Least Squares (LS), Recursive Least Square (RLS), and a dynamic parameter identification algorithm based on Particle Swarm Optimization (PSO) with search space defined by RLS (RLSPSO). A helical trajectory in the cartesian space is used as input. The dynamic model is found through the Lagrange equation and the motion equations, which are used to calculate the torque values of each joint. The torques are calculated from the values of the inverse kinematics, identified by each algorithm and from the manipulator joint speeds and accelerations. The results obtained for the trajectories, speeds, accelerations, and torques of each joint are compared for each algorithm. The computational costs as well as the Multi-Correlation Coefficient ( $R^2$ ) are computed. The results demonstrated that the identification accuracy of RLSPSO is better than that of LS and PSO. This paper brings an improvement in RLS because it is a method with high complexity, so the proposed method (hybrid) aims to improve the computational cost and the results of the classic RLS.