An improved neurogenetic model for recognition of 3D kinetic data of human extracted from the Vicon Robot system

Abstract

These days, it is crucial to discern between different types of human behavior, and artificial intelligence techniques play a big part in that. The characteristics of the feedforward artificial neural network (FANN) algorithm and the genetic algorithm have been combined to create an important working mechanism that aids in this field. The proposed system can be used for essential tasks in life, such as analysis, automation, control, recognition, and other tasks. Crossover and mutation are the two primary mechanisms used by the genetic algorithm in the proposed system to replace the back propagation process in ANN. While the feedforward artificial neural network technique is focused on input processing, this should be based on the process of breaking the feedforward artificial neural network algorithm. Additionally, the result is computed from each ANN during the breaking up process, which is based on the breaking up of the artificial neural network algorithm into multiple ANNs based on the number of ANN layers, and therefore, each layer in the original artificial neural network algorithm is assessed. The best layers are chosen for the crossover phase after the breakage process, while the other layers go through the mutation process. The output of this generation is then determined by combining the artificial neural networks into a single ANN; the outcome is then checked to see if the process needs to create a new generation. The system performed well and produced accurate findings when it was used with data taken from the Vicon Robot system, which was primarily designed to record human behaviors based on three coordinates and classify them as either normal or aggressive.