Advanced Materials Research Vol 216 (2011) pp 250-253 Online: 2011-03-15

© (2011) Trans Tech Publications, Switzerland

doi:10.4028/www.scientific.net/AMR.216.250

Inverse Kinematics Solution for a 6R Special Configuration Manipulators

Based on Screw Theory
Tan Yue-shenga Cheng Peng-leb Xiao Ai-ping c
(Technical School, Beijin Forestry University, Beijing 100083)
a b
tanyuesheng@163.com, chengpengle@yahoo.com.cn, c xxaapp@sina.com.cn

Keywords: motion screw, Paden-Kahan sub-problem, inverse kinematics

Abstract. Three basic sub-problems of screw theory are acceptable for some particular
configuration manipulators’ inverse kinematics, which can not solve the inverse kinematics of all
configuration manipulators. This paper introduces two extra extended sub-problems, through which
all inverse kinematic solutions for 6-R manipulators having closed-form inverse kinematics can be
gained. The inverse kinematic solution for a new particular configuration manipulator is presented.

Introduction
In recent years, screw theory is widely used in kinematics, dynamics and control of the robot and gets
a fine effect. For a general configuration manipulator, Denavit-Hartenberg parameter method is
employed to make mathematical model for kinematics and dynamics of the robot, and then a number
of value iterative methods [1,2] or neural network methods [3,4,5] or genetic algorithm [6] are taken
to solve the inverse kinematics or dynamics of the robot, which not only cost much calculating time
for its’ high computational complexity, but also the integrality and convergency of the solution can
not be warranted. On the other hand, the model based on D-H parameter method requires that each
joint and the end-effector has to be connected with a local coordinate in order to represent four
parameters of each joint, and different configurations correspond to different solution methods, thus it
is not a general method [9]. There are two main advantages to using screw theory. The first is that
they allow a global description of rigid body motion which does not suffer from singularities due to
the use of local coordinates. Such singularities are inevitable when one chooses to represent rotation
via Euler angles, for example. The second advantage is that screw theory provides a very geometric
description of rigid motion which greatly simplifies the analysis of mechanisms [7]. The method to
solve the inverse kinematics problem based on screw theory generally reduces the full inverse
kinematics problem into appropriate sub-problems whose solutions are known. For any configuration
manipulator, the inverse kinematics of the manipulator is rather simple if the sub-problems employed
are not out of the existing ones. Some scholars such as Chen and Gao [8], Zhao Jie [9], etc, have
solved the inverse kinematics for some manipulators having a special configuration basing on the
following three sub-problems and exponential equation. For a manipulator only including rotary
joints, the typical sub-problems includes[7]:
ξθ
Sub-problem 1: Rotation about a single axis, find θ such that e p = q
ξθ ξ θ ˆ ˆ
Sub-problem2: Rotation about two subsequent axes, find θ1 and θ 2 such that e e p = q
1 1 2 2

ˆ
q − e ξθ p = δ
Sub-problem 3: Rotation to a given distance, find θ such that
Sub-problem 4: Rotate p about the axis of ξ1 followed by a rotation about the axis of ξ 2 such
ˆ ˆ
p δ q δ q θ θ eξ θ eξ θ p − q1 = δ1 1 1 2 2

that the final location of is 1 from 1 and 2 from 2 , find 1 , 2 such that
ˆ ˆ
eξ1θ1 eξ2θ2 p − q2 = δ 2
and
The above-mentioned four sub-problems are basic ones. Depending on all the said sub-problems,
it is impossible to solve the inverse kinematics of some manipulator in the case all the axes of the
adjacent joints are neither paralleling nor intersecting, even if they have closed- form solutions. For

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans
Tech Publications, www.ttp.net. (ID: 158.42.28.33, Universidad Politecnica de Valencia, Valencia, Spain-20/04/15,09:52:50)
 Advanced Materials Research Vol. 216 251

this reason, extra two extension sub-problem must be introduced to solve this problem, one is
called by Sub-problem 5[11] herein, which is similar with Sub-problem 2 except that the two axes are
disjoint. Another one is called by Sub-problem 6[12] herein, which is similar with Sub-problem 4[10]
except that the two axes are disjoint.

Fig.1 sub-problem 5 Fig.2 sub-problem 6

Solving inverse kinematics for two special configuration manipulators using sub-problems
Before solving for θi (i = 1,2...6) , a conclusion is requisite during solving inverse kinematics:
ˆ −ξˆ jθ j ˆ ˆ ξˆ j θ j ξˆk θ k
e −ξ k θ k e e −ξiθi p w = p w eξ iθ i e e pw = pw
, if
ξ
where, ξi , j , ξ k are three adjacent axes of joint i , j and k respectively, pw is the common point
of the axes
ξi , ξ j , ξ k . Obviously, the conclusion is true if we premultiply e −ξˆ θ , e −ξˆ θ , e −ξˆ θ to the
i i j j k k

ˆ ξˆ j θ j ξˆk θ k
e pw = pw eξ iθ i e
both side of the equation subsequently.
As shown in Figure 4, the manipulator consists of six joints with a spherical wrist, in which the
axes of the first three joints intersect at one point p w , the axes of the rest joints are unparallel and
disjoint. The inverse kinematics for this 6R manipulator based on screw theory has not any reference
yet.

Fig.3 a 6R manipulator with the axes of the joints 4, 5, 6 neither parallel nor joint
The equation we wish to solve is
ˆ ˆ ˆ ˆ ˆ ˆ6
eξ1θ1 eξ2θ2 eξ3θ3 eξ4θ4 eξ5θ5 eξ θ6 g st (0) = g d (1)
In order to solve the inverse kinematics for this manipulator, rearranging this equation
ˆ ˆ ˆ6 ˆ ˆ ˆ
eξ4θ4 eξ5θ5 eξ θ6 g st (0) g d−1 = e−ξ3θ3 e −ξ2θ2 e−ξ1θ1 (2)
Based on the said sub-problems, the requisite joint angles are determined in four steps.
Step 1 (solve for θ5 and θ 6 ).
252 Optical, Electronic Materials and Applications

pw
Apply both sides of equation (2) to point which is the common point of intersection for the wrist
∧
ξ
axes of ξ1 , ξ 2 , 3 , since
exp(− ξ θ ) pw = pw
, yields
ˆ ˆ ˆ6
eξ4θ4 eξ5θ5 eξ θ6 g st (0) g d−1 pw = pw (3)
Subtract from both sides of equation (3) two point p1 , p2 which are two different points on the axis
of ξ 4 :
ˆ ˆ ˆ6
eξ4θ4 eξ5θ5 eξ θ6 g st (0) g d−1 pw − p1 = pw − p1 (4)
ξˆ θ ξˆ θ ξˆ6θ
e 4 4
e 5 5e 6
g st (0) g d−1 pw − p2 = pw − p2 (5)
Using the property that the distance between points is preserved by rigid motions, take the
magnitude of both sides of equation (4) and (5), we can get:
ˆ ˆ6
eξ5θ5 eξ θ6 g st (0) g d−1 pw − p1 = pw − p1 (6)
ˆ ˆ6
eξ5θ5 eξ θ6 g st (0) g d−1 pw − p2 = pw − p2 (7)

Equations (6) and (7) are in the form of Sub-problem 6, with

p = g st (0) g d−1 pw , q1 = p1 and
q2 = p2 . Applying Sub-problem 6, we solve for θ 4 and θ5 .
Step 2 (solve for θ 4 ).
Since
θ5 and θ6 are known, equation (3) becomes
ˆ ˆ ˆ
eξ4θ4 (eξ5θ5 eξ θ6 g st (0) g d−1 pw ) = pw
6
(8)
ξˆ θ ξˆ6θ
Applying sub-problem 1 with
p = e 5 5e 6
g st (0) g d−1 pw and q = pw to find θ 4 .
Step 3 (solve for θ1 and θ 2 ).
Since
θ 4 , θ5 and θ6 are known, rearranging equation (1)
ˆ ˆ ˆ ˆ ˆ ˆ
eξ1θ1 eξ2θ2 eξ3θ3 = g d g st−1 (0)e −ξ θ6 e−ξ5θ5 e−ξ4θ4 6
(9)
Applying both sides of equation (9) to a point which is on the axis of
ξ3 but not on the ξ1 and ξ 4
axes. This gives
ˆ ˆ ˆ6 ˆ ˆ
eξ1θ1 eξ2θ2 p3 = g d g st−1 (0)e−ξ θ6 e −ξ5θ5 e −ξ4θ4 p3 (10)
−ξˆ6θ ˆ −ξˆ θ
Equation (10) is in the form of sub-problem 2 with
p = p3
and
q = g d g st−1 (0)e 6
e−ξ5θ5 e 4 4
p3
,
we solve for 1 and
θ θ2 .
Step 4 (solve for θ3 ).
The remaining unknown is
θ3 . Rearranging equation (9) and applying both sides any point p5

which is not on the axis of

ξ3 ,
ˆ ˆ ˆ ˆ6 ˆ ˆ
eξ3θ3 p5 = e −ξ2θ2 e−ξ1θ1 g d g st−1 (0)e−ξ θ6 e −ξ5θ5 e −ξ4θ4 p5 (11)
−ξˆ θ −ξˆ θ −ξˆ6θ −ξˆ θ −ξˆ θ
Apply sub-problem 1 with
p = p5
and
q=e 2 2
e 1 1
g d g st−1 (0)e 6
e 5 5
e 4 4
p5
to get
θ3 .

At the end,
θ1 through θ6 are determined.

Experiments
To verify the correctness of the algorithm, we construct such a manipulator whose configuration is
same as that shown in Fig.1 and its’ structure parameter can be determined as wish, give an arbitrary
 Advanced Materials Research Vol. 216 253

group of joint angles, and get the positions and orientations of the end-effector by applying the
forward kinematics. Taking the computed positions and orientations of the end-effector as a known
condition, we solve for the joint angles of the manipulator by applying the inverse kinematics.
Comparing the given joint angles with the computed joint angle, we found that they are identical, with
which the correctness of the algorithm is verified.

Summary
Based on four basic sub-problems and two extended sub-problems, solution for a new special
configuration manipulator is presented. The inverse kinematics process based on the screw theory is
more explicit and direct than that of D-H parameters method, and provides a new approach for the
inverse kinematics of manipulator. The result and solving procedure presented in this paper is
beneficial for solving the inverse kinematics of a reconfigurable module manipulator, with which it is
possible to found a general mathematic model for any configuration manipulator.

References
[1] KELMAR L, KHOSLA P K. Automatic generation of kinematics for a reconfigurable modular
manipulator system [C]. Proceedings of IEEE Conference on Robotics and Automation,
Philadelphia, PA, 1988: 663-668.
[2] CHEN I M, YANG G L. Inverse kinematics for modular reconfigurable robots [C].
Proceedings of IEEE Conference on Robotics and Automation. Leuven, Belgium, 1998:
1647-1652
[3] TEJOMURTULA S, KAK S. Inverse kinematics in robotics using neural networks[J].
Information Sciences, 1999, 116(2-4):147-164
[4] Z. Bingul, H.M. Ertunc and C. Oysu, Comparison of Inverse Kinematics Solutions Using
Neural Network for 6R Robot Manipulator with Offset, 2005 IEEE
[5] Eimei OYAMA, Susumu TACHI, Inverse Kinematics Learning by Modular Architecture
Neural Networks,1999 IEEE, P2065~2070
[6] F. Chapelle, P. Bidaud,A Closed Form for Inverse Kinematics Approximation of General 6R
Manipulators using Genetic Programming
[7] Richard M.Murray, Ze xiang Li, S.ShankarSastry, A Mathematical Introduction to Robotic
Manipulation [M ]. 2005
[8] I-Ming Chen,Yan Gao,Closed-Form Inverse Kinematics Solver for Reconfigurable
Robots,Proceedings of the 2001 IEEE,International Conference on Robotics B Automation
Seoul, Korea May 21-26, 2001
[9] Zhao Jie, Wang Weizhong, Cai Hegao, Generation of Closed-Form Inverse Kinematics for
Reconfigurable Robots[J].Chinese Journal of Mechanical Engineering, 2006:210-214
[10] Zhao Jie, Wang Weizhong, Cai Hegao, Solution for One Type of Inverse Kinematics
Sub-problem in Screw Theory and Its Applica tion[J]. ROBO T, 2005:163~167
[11] Yue-Sheng Tan,Ai-Ping, Xiao,Extension of the second paden-kahan sub-problem and its'
application in the inverse kinematics of a manipulator[C].RAM 2008,p 379-381
[12] Tan Yue-sheng,Cheng Peng-le,Xiao Ai-ping,Solution for a new Sub-problem in Screw Theory
and its’ Application in the Inverse Kinematics of a Manipulator,Periodical of Applied
Mechanics and Materials Vols. 34-35,2010
Optical, Electronic Materials and Applications
10.4028/www.scientific.net/AMR.216

Inverse Kinematics Solution for a 6R Special Configuration Manipulators Based on Screw Theory
10.4028/www.scientific.net/AMR.216.250

DOI References
[3] TEJOMURTULA S, KAK S. Inverse kinematics in robotics using neural networks[J]. nformation
Sciences, 1999, 116(2-4):147-164
doi:10.1016/S0020-0255(98)10098-1
[3] TEJOMURTULA S, KAK S. Inverse kinematics in robotics using neural networks[J]. Information
Sciences, 1999, 116(2-4):147-164
doi:10.1016/S0020-0255(98)10098-1