Paraconsistent Artificial Neural Cell in Application of Learning Rodrigo Silvério da Silveira, João Inácio da Silva Filho, Dorotéa Vilanova Garcia, Maurício Conceição Mário, Cláudio Luis Magalhães Fernandes

Abstract

In Artificial Intelligence "Learning from Demonstration - Apd (LdD) is the process in which it relies on teaching new tasks to machines by means of repetitive manipulation, without the need for programming. In order to facilitate the assignment of movements in industrial machines, a Paraconsistent Artificial Neural Cell (PAN_Cell) was implemented in a Programmable Controller (CP), constituting a Paraconsistent Demonstration Learning System (SPApD-PAL2v). For validation tests the SPApD-PAL2v was applied in a machine with electropneumatic parts movement process, composed of linear actuators, suction cup and linear positioning analog sensors. The results obtained in this work demonstrate that the SPApD-PAL2v presents fundamentally as a potential alternative for new implementations of controls in industrial processes.