PAMM · Proc. Appl. Math. Mech. 9, 119 – 120 (2009) / DOI 10.1002/pamm.

200910034

Inverse dynamics of underactuated multibody systems

Mahmud Quasem∗1 , Stefan Uhlar1 , and Peter Betsch1
1
Chair of Computational Mechanics, Department of Mechanical Engineering, University of Siegen, Paul-Bonatz-Str. 9-11,
D-57068 Siegen, Germany

The present work deals with controlled mechanical systems subject to holonomic constraints. In particular, we focus on
underactuated systems, defined as systems in which the number of degrees of freedom exceeds the number of inputs. The
governing equations of motion can be written in the form of differential-algebraic equations (DAEs) with a mixed set of
holonomic and control constraints. The rotationless formulation of multibody dynamics will be considered [1]. To this end,
we apply a specific projection method to the DAEs in terms of redundant coordinates. A similar projection approach has been
previously developed in the framework of generalized coordinates by Blajer & Kołodziejczyk [2].


c 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

1 Equations of motion
The motion of the discrete mechanical systems with mixed holonomic and servo constraints is governed by DAEs in the
following form
q̇ − v = 0
M v̇ + ∇V (q) + G (q)λ + B T u
T
= 0
(1)
g(q) = 0
c(q, t) = 0
Here, q ∈ Rn is the vector of redundant coordinates, λ ∈ Rmh is a vector of multipliers, V (q) ∈ R is the potential energy
function, g(q) ∈ Rmh is a vector of geometric (or holonomic) constraint functions, G = Dg(q) ∈ Rmh ×n is the constraint
Jacobian. Consequently, the constrained mechanical system has ñ = n − mh degrees of freedom. Furthermore, c(q, t) ∈ Rm̃
is a vector of servo (or control) constraint functions which may be written in the form
c(q, t) = Φ(q) − γ(t) (2)
The servo constraints serve the purpose of partially specifying the motion of underactuated mechanical systems (m̃ < ñ).
Servo constraints may be used for trajectory tracking problems where the vector-valued function γ(t) ∈ Rm̃ together with
Φ(q) ∈ Rm̃ specify the output. The corresponding actuator forces are determined by the control inputs u ∈ Rm̃ in conjunction
with the input transformation matrix B ∈ Rm̃×n .

2 3-link planar manipulator in redundant coordinates

Here we will consider a 3-link planar manipulator [3] with three revolute joints as a prototypical example of an underactuated
mechanical system. For i = 1, 2, 3, αi , mi , Ji , li (= 2ri ) denotes the absolute angle, the mass, the moment of inertia and the
length of the i−th link and the actuating torques τ1 and τ2 are applied on the first and second link but the third link is free to
rotate. In the 2D case the orientation of the i−th body is specified by means of two mutually orthogonal vectors dij ∈ R2 ,


j = 1, 2 which form the director frame di1 , di2 located at the center of mass ϕi ∈ R2 of the i−th (rigid) body, i = 1, 2, 3.
In this example the total number of redundant coordinates, (n = 21) is given by
 T
q = q1 q2 q3 θ1 θ2 θ3 (3)

where θ1 , θ2 and θ3 are relative angles of the link used as augmented coordinates and
q 1 , q 2 , q 3 are defined as
α3
 T
q i = ϕi di1 di2 i = 1, 2, 3 (4)
γ(t)
Furthermore, the motion of the center of mass of the last link is considered as a pre- α2
e2 d12
scribed trajectory in the form d11 τ2
α1
e1
γ(t) = ϕ3 (5) τ1
∗ Corresponding author: e-mail: [email protected], Phone: +49 271 740 4637, Fax: +49 271 740 2436

c 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 16177061, 2009, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pamm.200910034 by Readcube (Labtiva Inc.), Wiley Online Library on [25/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
120 Short Communications 1: Multibody Dynamics

3 Projected formulation in terms of redundant coordinates

Here the projection method originally proposed in terms of generalized coordinates by Blajer & Kołodziejczyk [2] is applied
to the present formulation by differentiating the servo constraints twice with respect to time:
c̈(q, t) = C v̇ + ξ = 0 (6)
By satisfying the complementary condition CD = 0, we obtain the projected formulation of the DAEs by multiplying (1)2
with D T and CM −1
 
D T M v̇ + D T ∇V (q) + GT (q)λ + B T u = 0
  (7)
−1
CM ∇V (q) + G (q)λ + B u − ξ = 0
T T

Finally the backward Euler method is used for a numerical discretization:

 q n+1 − q n − ∆t vn+1 = 0
D M (v n+1 − v n ) + D ∇V (q n+1 ) + G (q n+1 )λn+1 + B un+1 ∆t
T T T T
= 0
 
CM −1 ∇V (q n+1 ) + GT (q n+1 )λn+1 + B T un+1 − ξn+1 = 0 (8)
g(q n+1 ) = 0
c(q n+1 , tn+1 ) = 0

4 Numerical Results
We parametrize the ñ dimensional configuration manifold with generalized coordinates, µ = (θ1 , θ2 , θ3 ), where θ1 , θ2 and
θ3 are the relative angles of the link. Comparison is made between the redundant coordinates (RED) and the generalized
coordinates (GEN). Here, the control inputs and the configuration variables are shown in the figures below using a time step
size of ∆t = 0.01. Additionally, the initial (upper left) and the final configuration (lower left) of the rotary crane are shown.

15 time = 0s
200
θ1

100
10

0
0 0.5 1 1.5 2 2.5 3
5 time

0
0 RED
2

−100 GEN
θ

−5 −200
0 0.5 1 1.5 2 2.5 3
time
relative angles
−10 200
2

0
θ

−15

−15 −10 −5 0 5 10 15 −200

0 0.5 1 1.5 2 2.5 3
time

4 input profiles
x 10
1.5

time = 3s 1
15
0.5
τ1

10 0

−0.5
5
−1
0 0.5 1 1.5 2 2.5 3
time
0 input vectors
4000
RED
GEN
−5 2000
τ2

0
−10
−2000

−15 −4000
0 0.5 1 1.5 2 2.5 3
−15 −10 −5 0 5 10 15 time

References
[1] Betsch, P. and Uhlar, S., Energy-momentum conserving integration of multibody dynamics. Multibody System Dynamics, 17(4),
243-289, 2007
[2] W. Blajer, K. & Kołodziejczyk, A geometric approach to solving problems of control constraints: Theory and a DAE framework,
Multibody System Dynamics, 11(4), 343-364, 2004.
[3] De Luca, A. and Oriolo, G., Trajectory Planning and Control for Planar Robots with Passive Last Joint. The International Journal of
Robotics Research, 21(5-6), 575-590, 2002

c 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.gamm-proceedings.com