Toggle menu
Toggle personal menu
Not logged in
Your IP address will be publicly visible if you make any edits.

Robot Study/Introduction to Robotics: Difference between revisions

From ZeroWiki
No edit summary
No edit summary
 
Line 1: Line 1:
__TOC__
[[DeleteThisPage]]
 
== Information ==
Purpose: Learning basic knowledge of robotics
Lecture: CS223A, Stanford University
Date: Jan 21, 2019 ~
 
* Prerequite
* Linear Algebra
* Numerical Analysis
(or graphics programming experience)
== Reference ==
Material: Copy from Stanford
Video clips: https://www.youtube.com/watch?v=0yD3uBshJB0&list=PL65CC0384A1798ADF
 
== Study List ==
=== Lecture 1: Spatial Description ===
General Manipulator: Robot Arm, using Revolute joint, Prismatic joint
* Robot Arm: base, link, joint, end-effector
* Revolute joint: Rotation movement, 1 Degree of Fredom(DoF)
* Prismatic joint: Linear movement, 1 DoF
* Denote joint type using ε(0 for revolute, 1for prismatic)
 
Discription of body1 (9 parameters)
* Link location: 3 points (Each point has 3 parameters)
 
Discription of body2 (6 parameters)
* Body orientation: 3 parameter
* Point on the body: 3 parameter
=> Robot arm(n:links, 1: base) has n DoF
 
Transformation
* Pure Rotation
* Pure Translation
* General Tasformation
* Inverse Transformation
 
Configuration Representation
There is no universial agreement in the field of robotics as to what is the best orientation representation.
Because each representation hase advantages and shortcomings
* Direction Cosines:
* Euler angle representation: ZYX, angle(α, β, γ)
* Fixed angle representation: XYZ, angle(γ, β, α)
* Inverse of an orientation representation
=== Lecture 2: Direct Kinematics ===
Previous
* Independent of the structure of the manipulator
 
Introduction
* A set of parameters specific to each manipulator
* ex) rotation, translation, link of manipulator
* Forware Kinematics
* Inverse Kinematics
 
Link Description
* Manipulator: Consist of a chain of links from base
* Consecutive links are connected by joints which exert the degree of freedom.
 
D-H Parameter
* link length(a): length along the common normal from axis (i-1) to axis i
* link twist(α): angle between this parallel line and axis (i-1)
* link offset(θ): distance alont the line on axis i between the common normal for link (i-1) and common normal for link i
* joint angle(d): angle between the two common normal for link (i-1) and common normal for link i
* Revolute joint: joint angle(variable), link offset(constant)
* Prismatic joint: joint angle(constant), link offset(variable)
* a, α: describe link
* d, θ: describe the link's connection
 
Conventions for First and Last Link
* Once robot structure is set link length & link twist is determined.
* a(i) and α(i) depend on joint axes i and i+1
Axes 1 to n: determined => a(1), a(2), ,,,, a(n-1) and α(1), α(2), ,,,,a(n-1)
* d(i) and θ(i) depend on
 
Attaching Frames to links
* ex1) RRR (Revolute-Revolute-Revolute) Manipulator
* ex2) RPRR (Revolute-Prismatic-Revolute-Revolute) Manipulator
 
Propagation of Frames
* Show how to calculate matrix about D-H parameter
* Reference
http://www.adrian.zentner.name/content/projects/xml/x3d/robot/res/Denavit-Hartenberg.gif
 
Kinematics of Manipulators
* Example of robot arm (Stanford Scheinman Arm)
* Reference
http://infolab.stanford.edu/pub/voy/museum/pictures/display/robots/StanfordArm.jpg
 
Direct(forward) Kinematics
* Mapping between the joint space of dimension n and the task space of manipulator of dimension m
* Called the "Geometric Model of the manipulator"
(It is determinded solely by knowing the geometry of manipulator)
* q(i) = ε'(i)θ(i) + ε(i)d(i)
* X = f(q)
=== Lecture 3: Inverse Kinematics ===
Introduction
* Difficult task: Multiplicity or non-existence of potential soultions
* Problem: find q given T(B,W) or x / find q = f^(-1)(x)
 
Closed Form Solutions
Algebraic: solution is found using the fact that θ1+θ2+θ3 = a0
Geometric: there are two possible solutions
 
Piper's Solution
???
 
Existence of Solution
* If these two equations are correct, solution of the inverse kinematics exists
* However, sometimes there is no solution because of limitation of robot model
 
Workplace of the Manipulator
* Workspace: the set of points that can be reached with the mainpulator
* Joint limitation is always defined by the mechanical design of the manipulator
* Related question: # of possible solutions
* Reachable Workspace: the set of points that can be reached in at least one conficuration of the manipulator
* Dextrous workspace: the set of points that can be reached by any possible orientation of the end-effector, important in the motion planning with obstacles (Reachable Workspace > Dextrous workspace)
 
# of Solutions
6R manipulator: 16 solutions
5RP manipulator: 16 solutions
4R2P manipulator: 8 solutions
3R3P manipulator: 2 solutions
in-parallel structures: 40 solutions
 
* Puma Robot
Reference
https://d2t1xqejof9utc.cloudfront.net/screenshots/pics/45f6b6d1d881d687d15e29d47f181a6f/large.PNG
 
* Stanford Scheinman Arm
Reference
http://infolab.stanford.edu/pub/voy/museum/pictures/display/robots/IMG_2404ArmFrontPeekingOut.JPG
 
=== Lecture 4: The Jacobian ===
Previous
* Establish the mathematical models which describe the relationships between the static configurations of a mechanism and its end-effector
 
Introduction
* Establish the relationship between δx and δq
* The relationship between δx and δq is described by the Jacobian matirx
* This matrix is key to the relationship between joint torques and end-effector forces
 
Differential Motion
* X = f(q)
Reference
https://ipfs.io/ipfs/QmXoypizjW3WknFiJnKLwHCnL72vedxjQkDDP1mXWo6uco/I/m/74e93aa903c2695e45770030453eb77224104ee4.svg
 
Example
* RR Manipulator
* Stanford Scheinman Arm
 
 
 
== Comments ==
선배님 너무 멋있어여 - [[조예진]]
* 새싹 준비입니다 - [[우준혁]]
 
== Back page ==
* [[우준혁]]



Latest revision as of 00:05, 30 June 2020