IMPLEMENTASI JARINGAN SARAF TIRUAN BERBASIS RAM PADA ROBOT WALL FOLLOWER DENGAN ALGORITMA MAZE MAPPING UNTUK MENENTUKAN RUTE TERCEPAT
Daftar Isi:
- Robot wall followers who navigate through the wall must study the position of the surrounding wall so that they can navigate safely without damaging the walls. Position distance wall learning will use RAM-based artificial neural network methods. There are 3 RAM nodes to store the pattern received from the left sensor, right sensor, and front sensor. 8 bits of input pattern received will be optimized so that only 4 bits are stored in the RAM node. So the computing process is simpler. A discriminator RAM has 3 RAM nodes each RAM node has 4 bit word (x = 4) with total 12 bit input vector (n = 12), so each discriminator RAM can receive 48 binary input patterns. Robot wall follower can come out of the maze using the left hand rule algorithm and robot can come out faster than the maze by converting from 8 intersections or dead ends to 4 intersections.