Modeling humanoid swarm robots with petri nets
Abstract
Robots have become a hot topic in today‟s electronic world. There are many definitions for it. One of the definition in Oxford dictionary states “a robot is a machine capable for carrying out a complex series of action automatically especially one programmable by a computer”. This paper deals with a special kind of robot, which is also known as humanoid robot. These robots are replication of human beings with head, torso, arms and legs.
A model of human is presented in this paper as discrete event system adapted from “Modeling and simulating motions of human bodies…”[1]. This model consists of sixteen interrelated limbs defined in 3D space, so most limbs/joints are able to make movement in three different angles (α, β and γ). Full details regarding Range of Motion (ROM) of rigid body in forward kinematic is illustrated. Human motions are categorized into two types: stochastic and deterministic motions. Deterministic motions are demonstrated using gait cycle of walking and running of normal adult person.
The main focus of this paper is to model and simulate humanoid robot represented as Discrete Event Systems (DES); in Petri Net using GPenSIM and later expand those group of robots to swarm setting. GPenSIM is General Purpose Petri Net simulator [2] developed as toolbox for MATLAB to model and simulated discrete events using Petri net tools. Each joint‟s angle is treated as a separate Petri Net model which is independent from each other and their movement‟s limits are defined by ROM of normal human body. The instructions relating to the motion of joints for simulation are fed through a file to the instructor. These movements of joints are represented by variation of tokens displayed at the end of simulation in a graphical figure. Further, same structure of model is used in swarm of robots. Instead of feeding instructions to individual robots, a central instructor is created. This instructor acts as a master to robots acting as slaves where slaves include some predetermined commands embedded inside them. With central command system, a proper synchronization is achieved among group of robots working as swarm. A normal routine of group dance or simple group sport can be accomplished with calculated instructions on this swarm of robot.
Description
Master's thesis in Computer science