Indonesian Journal of Electrical Engineering and Computer Science Vol. 21, No. 2, February 2021, pp. 1238~1246 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v21.i2. pp1238-1246  1238 Journal homepage: http://ijeecs.iaescore.com Recursive least square and control for PUMA robotics Lafta E. Jumaa Alkurawy Department of Electronics Engineering, Universitas of Diyala, Iraq Article Info ABSTRACT Article history: Received Jul 17, 2020 Revised Sep 19, 2020 Accepted Oct 3, 2020 The solution of inverse kinematics system based on recursive least square (RLS) theorem is improved this paper. The task in joints of robotics is inverse kinematics for PUMA robotics. The design the manipulator of robotics is not simple if due to model of algebraic method. I suggested a method of RLS method to get predicts the positions of robot and it is comfortable the applications in real-time. The RLS is used to find the solution of the inverse kinematics for the joints 6-dof of the robotics. This technique is important to compute the joints of each arm space with Cartesian axes in the end-effector. The identification will be in each joint for PUMA by RLS and applied PI controller on each joint to get the response follows the reference input by tuning the values of coefficients of PI. Keywords: Joints 6-dof Kinematics system PI controller PUMA robotics Recursive least square This is an open access article under the CC BY-SA license. Corresponding Author: Lafta Esmaeel Jumaa Alkurawy Department of Electronics Engineering University of Diyala Quds Square, Baquba, Diyala, Iraq Email: lafta_67@yahoo.com 1. INTRODUCTION The objective of this work suggests an algorithm for predicting the robot position from reading measuringof the sensoes some feature of framework. This algorithm is prepared for implementations when readings the sensor are not cheap or is limited so that just comparatively few can get, the noise and errors of readings that are taken. The algorithm suggested enable of approaching to predict the position with high precision by using alittle measurments than other algorithms for implementations. This algorithm is validated by using a PUMA robotic toget information of position with using a controller to ensuare the stability without problems or mistakes. Zhu and Dai [1] proposed a method to deal with data of missing are in the process of fusion, so its accuracy of fusion is decreased. Two algorithms the recursive least square and datch ae suggested to develop the precision the prediction of fusion. The performance of calculation is bad because the the first suggested theory focused on the performs fitting of least square at the same time. The simulation and analysis display that the algorithms of suggested to deal with missing data case and the prediction of fusion is accuracy. Wen et al. [2] proposed a technique of analytic and method of fuzzy logic are put into dynamic modeling for robotic fish with swimming control. The method of fuzzy control is focused on the behavior of the robotic fish dynamic knowledge. The smaller steady state error and fast acceleration are got from fuzzy controller. The method of conventional control is the trust efficiency through the steady. Cabré et al. [3] described a method of learning of project from control of robotics and vision of computer. The target of learning was to evolve a system of computer vision that must to discover the object that located on surface of target and control arm of computer to move it and turn it to destination. The results that be seen that evolution a case of learning focused on vision of computer and motivation of computer has increased.