Artificial Neural Networks Applied in the Solution of the Inverse Kinematics Problem of a 3D Manipulator Arm

A neural autonomous robotic manipulator with three degrees of freedom

“…3 On the other hand, trajectory planning determines kinematic characteristics, such as acceleration and speed of the manipulator. 23 Thus, the trajectory planning problem concatenates different computational techniques. In this context, a possible solution involves applying computational algorithms and data collected from sensors to obtain the angle of a robotic arm or the position of an object in the robot's operating area.…”

“…The trajectory planning can be performed in the joint space through polynomials defined from the initial boundary conditions such as the initial and final angles, already obtained through inverse kinematics. 23 Thus, one way to carry out trajectory planning is through a fifthorder polynomial Equation (1), where it is possible to obtain the relationship between position, velocity, and acceleration for the desired movement. It is important to highlight that there are other trajectory planning algorithms in the literature, such as the Linear Segments with Parabolic Blends, which can be observed in [20].…”

“…This study starts with the camera calibration procedure in two and three dimensions [39] and then [23], in which an ANN type perception multilayer (MLP) was employed in the 3D inverse kinematics solution. As for the development of this prototype, more techniques and components were added to advance and link these investigations carried out.…”

“…The trajectory is a path with time limitations 3 . On the other hand, trajectory planning determines kinematic characteristics, such as acceleration and speed of the manipulator 23 . Thus, the trajectory planning problem concatenates different computational techniques.…”

“…In general, an ANN is a parallel distributed processor consisting of tiny structures that, combined, can store knowledge from experience. ANNs have two main characteristics 23 : (i) Knowledge is acquired by the network from its environment through a learning process; (ii) Interneuron connecting forces, known as synaptic weights, are used to store the acquired knowledge.…”