Genetic design of 3D modular manipulators

Abstract

This paper proposes a method for task based design of modular robotic systems using Genetic Algorithms (GA). We introduce a 30 kinematic description for modular serial manipulators and a two-level GA to optimize their topology from task specifications. Revolute and prismatic joints are considered and the number of DOF is let to the G A to determine (allowing redundant manipulators). The upper-level G A is dedicated to the topology evolution and uses the lower-level G A to search for inverse kinematics problem solutions. The topology is evolved for adaptation to a global task constituted by several required end effector configurations (subtasks). The implemented G A optimizes severalperformance criteria under constraints. To illustrate the method capacities, an example is presented for a redundant manipulator an Q cluttered workspace.

Evolutionary cost-optimal composition synthesis of modular robots considering a given task

“…It is obtained by (6), where J is the Jacobian matrix and w is the Yoshikawa manipulability index calculated by w = p det(JJ T ) [21].…”

“…The criteria R, L, A, D, and I are also considered in [6] and O is modified from [6] to obtain less conservative distances. One additional criterion, V , is added to these evaluation functions to minimize and optimize differences between joint angles.…”

“…The first part of the equation represents the method proposed in [6]. Adding the second part of (10) enables us to select the compositions considering the task-related objectives.…”

“…This method mainly consists of a top-level GA to generate compositions from the given modules and a lower-level GA to optimize joint positions. The same authors present another technique which stores both composition and configuration information in the same chromosome [6]. In [7], a GA-based modular robot composition synthesis method for obtaining optimal kinematic configurations for task-based, fault-tolerant manipulators is proposed.…”

Coevolutionary robotics

“…However, an evolutionary design of the robot's configuration was not attempted. Recent work by Chocron and Bidaud (1997) describes an attempt to evolve both the morphology and the inverse kinematics of a modular manipulator composed of prismatic and revolute joints. A genetic algorithm searches for suitable configuration for a task given as a set of required effector configurations.…”