Multibody System Dynamics (2023) 60:317–345

https://doi.org/10.1007/s11044-023-09900-0
RESEARCH

Dynamics of double pin caterpillar platform using a

generalized cylindrical contact model

Pingxin Wang1 · Xiaoting Rui1 · Guoping Wang1 · Hailong Yu1 · Bin He1 · Junjie Gu1

Received: 31 October 2022 / Accepted: 15 March 2023 / Published online: 27 March 2023
© The Author(s), under exclusive licence to Springer Nature B.V. 2023

Abstract
Compared with the single pin crawler, a double pin caterpillar platform has lower vibration
and noise, which is widely used in tracked vehicles. However, its dynamical model is more
complex due to its high degrees of freedom and laborious contact detection involved. It is
significant to develop an accurate and efficient contact algorithm between tracks and wheels.
In this paper, the profiles of the track connector and the tooth groove are discretized into
several arc surfaces. The contact between them is characterized as convex and concave. Ac-
cordingly, a general mathematical description of the tooth profile is established to describe
the real tooth groove geometry. Then a generalized cylindrical contact model is developed to
evaluate forces. In the contact detection process, the local frames of the sprocket and tooth
groove are employed, resulting in the vector quantities and coordinate transformations de-
scribing the contact arc surfaces constantly. This approach can facilitate programming and
effectively improve calculation efficiency. Subsequently, a field test was performed to verify
the correctness of the dynamical model. To reduce the wear of the sprocket, different cases
are simulated, and a reasonable radius of the rubber ring is proposed.

Keywords Double pin caterpillar platform · Contact detection · Multibody dynamics ·

Sprocket tooth · Cylindrical contact

1 Introduction

The main contribution of a caterpillar platform is to drive a tracked vehicle moving [1, 2].
Through the meshing between tracks and sprockets, the torques from an engine are trans-
formed into interaction forces between tracks and the ground. Sometimes it also transmits
braking force from the ground to stop the vehicle. Accordingly, the structure of the meshing
pair has a vital influence on the dynamical performance of the caterpillar platform. How-
ever, due to the harsh working condition, sprockets are often required to withstand multi-
directional complex loads. The dynamical analysis has been a popular topic in multibody
research [3, 4].

 J. Gu
[email protected]

1 Nanjing University of Science and Technology, Nanjing, China

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

318 P. Wang et al.

There are two models to connect adjacent track links: single pin connection and double
pin connection. For a track chain with the same total length and number of links, the rev-
olute joints of the double pin track are twice as much as the single pin. Accordingly, the
track pitch is approximately reduced by 50%, which improves the flexibility of the track
chain and is beneficial to reduce impact and noise. But the dynamical model of a double
caterpillar platform is more complex due to its high number of degrees of freedom (DOF)
and laborious contact detection involved. Early studies approximated the track drive to a
belt drive. The track is modeled as a flexible belt, which is driven by friction forces from
the sprocket. This model can effectively simulate the track movement and has an advantage
in calculation time [5–9]. Unfortunately, it fails to analyze the engagement between tracks
and sprockets, resulting in the loss of high-frequency vibration simulation. Then a three-
dimensional multibody dynamical model is established [10, 11]. Adjacent track links are
connected by a revolute joint. The contact between tracks and the sprocket is assumed to oc-
cur between the sprocket teeth and track pins. Three surfaces are used for contact analysis:
the left, right, and seating surfaces. Gilad [12] assumed that the sprocket-tooth-shape profile
is circular and the collision detection is performed analytically.
As is generally known, there are design standards for sprockets of the chain transmission
systems [13]. However, the sprockets of tracked vehicles are still in an empirical design
stage. Considering the similarity of these two structures, we can draw on some existing
analysis approaches in chain drives to study the caterpillar platform. Ambrosio [14, 15]
studied a planar model for the chain drive. A cylindrical contact force model is developed to
analyze the influence of clearance on the system. Pereira [16, 17] developed a generalized
revolute clearance joint formulation to study the dynamics of chain drives. The sprocket
tooth profile is discretized into seven contact regions. This model closely approximates real
geometry. Tian [18] summarized the contact models of clearance joints, in which the part of
spatial revolute joint with clearance can be employed in dynamics analysis of the caterpillar
platform.
The above research is mainly applicable to single pin tracks. The main aim of this re-
search is to establish a dynamic model of the double pin caterpillar platform. Based on the
transfer matrix method for multibody systems (MSTMM) [19], it is modeled as two sub-
systems: a track chain system and a wheel system. They are coupled by contact forces.
The complicated part is the contact analysis between tracks and sprockets. The profiles of
sprocket teeth and track connectors are discretized into several arc surfaces. Then a general-
ized cylindrical contact model is developed to evaluate forces. By means of wireless sensing
technology, a field test is performed to verify the correctness of the proposed dynamical
model. Further, the variation of contact forces in the meshing process is simulated.
The main contributions of this paper are summarized as follows.
(1) Based on MSTMM, a nonlinear dynamical model of the double pin track system is
established.
(2) The contact algorithm between tracks and wheels is developed, in which the track-
tooth engagement is considered.
(3) A field test is performed to collect the acceleration of a selected track link. Through
the comparisons in time domain and frequency domain, the correctness of the proposed
model is verified.
(4) The influence of the sprocket rubber ring on the meshing characteristic is analyzed,
which is useful for the design of sprockets.
The novelties of this paper are:
(1) Using multiple arcs with different curvatures, a general mathematical description of
the tooth profile is developed.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 319

Fig. 1 The double pin track movement system (a) and its dynamical model (b)

(2) The local frames of the sprocket and tooth groove are employed for contact detection,
which can facilitate programming and effectively improve calculation efficiency.
This paper is composed of six sections. The dynamical model of a representative double
pin track system is established in the following section. The contact algorithm between track
connectors and sprocket teeth is detailed in Sect. 2. Section 4 presents the contact algorithm
between a track link and the rubber ring, which is also applicable to that between tracks
and the road wheel or idler. A field test is performed in Sect. 5, and the influence of the
sprocket rubber ring on the system dynamical characteristic is simulated in Sect. 6. Finally,
the conclusions are shown in Sect. 4.

2 Dynamical model

A representative device for dynamical analysis of the double pin track is shown in Fig. 1. It
consists of a base, a sprocket, an idler, a road wheel, and a track chain of 35 track links. The
base is fixed with the ground. The track system and wheels are decoupled in kinematics.
They are coupled by contact forces.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

320 P. Wang et al.

Fig. 2 Schematic diagram of the

double pin track structure

Fig. 3 Topology of the track

chain

2.1 Track model

A portion of the double pin track chain is exhibited in Fig. 2. Adjacent track links are con-
nected by two track pins, and a connector combines these two as a body element. The track
pin with rubber bushing is pressed into the pin hole of a track link according to its configured
location. After assembly, the adjacent track links have a pre-twist angle in a free state.
According to the structural characteristics, a double pin track chain can be modeled as
a closed-loop multibody system, which is alternately connected by multiple spatial rigid
bodies. There are two types of body elements. One is the track link, and the other is a
combination of two pins with a connector and a center guide. The connection between these
two elements is modeled as a revolute joint.
According to MSTMM, the topology of the track chain is sketched in Fig. 3, where a
body element is denoted as “◦” and a joint element and its transfer direction as “→”. The
state vector of a junction of two elements is defined as
 T
z = ẍ ÿ z̈ Ω̇x Ω̇y Ω̇z mx my mz qx qy qz 1 , (1)

where ẍ, ÿ, z̈ are translational accelerations, Ω̇x , Ω̇y , Ω̇z angular accelerations, mx , my , mz
internal moments, and qx , qy , qz internal forces described in the inertial frame. The “1” is
auxiliary to account for external moments and forces [20, 21].
The joint element 70 is cut off, and its role is described as an elastic-damping model.
Then the track system consists of 35 body elements and 34 revolute joints. Its degrees of
freedom are 40. Based on the automatic deduction theorem [22, 23], its overall transfer
equation is obtained as

U all zall = 0, (2)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 321

where

U all = [−I 12 T 1−69 ] (3)

 T
zall = zT69,0 zT1,0 (4)
T 69−1 = U 69 U 68 U 67 · · · U 2 U 1 . (5)

In Eqs. (3)–(5), I 12 is an identity matrix with the dimension of 12×12, U i is the transfer
matrix of element i. The transfer matrices of all elements can be calculated according to
their kinematics information.
The generalized coordinates of the track system consist of the position coordinates and
Euler parameters of element 1 input and relative angles of all joints, i.e.,
 T
y = x1 , y1 , z1 , ε1 , ε2 , ε3 , ε4 , φ2 , φ4 , . . . φj , . . . , φ68 . (6)

The generalized coordinates and velocities [y T ẏ T ]T are explicit quantities at time ti . Solving
Eq. (2), the state vectors of all elements are calculated. Then the angular acceleration of a
joint j is obtained as

φ̈j = H 3 ATj,I Ω̇ j,O − H 3 ATj,I Ω̇ j,I − H 3 Ω̃ j,I Ω r,j , (7)

where Aj,I is the transformation matrix of its local coordinate system with respect to the
inertial frame, Ω̇ j,O , Ω̇ j,I are the angular velocities of its output and input in the inertial
frame, Ω̃ j,I is the skew-symmetric matrix of the angular velocities of its input in the local
coordinate system, H 3 = [ 0 0 1 ], and Ω r,j = [ 0 0 φ̇j ]T . It should be noted that this
paper solves the forward dynamics problem [24]. The forces applied on the elements are cal-
culated according to their kinematic information at the current moment ti . Solving Eq. (2),
the accelerations of elements are acquired, and then the velocities and displacements at the
next moment are given by integrating w.r.t. time.

2.2 Contact model between the sprocket and tracks

The contact between a sprocket and tracks may occur in different locations. One is the
connector and a sprocket tooth groove, as exhibited in Fig. 4a. The other is the track link
and the rubber ring, as exhibited in Fig. 4b.
The contact model between a track link and the rubber is identical to that between driven
wheels and tracks, as will become clear in Sect. 2. The contact between a connector and a
tooth groove is complicated. Figures 5a and 5b illustrate the cross sections of a connector
and a tooth groove, respectively. The connector has two contact arc surfaces: the left and
right. They have the identical radius rC , and the distance between the centers of the two
arcs is LC . The tooth groove profile is divided into ten contact areas, which are bilaterally
symmetrical. These areas are all characterized as arc surfaces with constant radii. The center
of the sprocket is OS . The radius of the sprocket addendum circle is ra , and Oj is the center
of area j with a corresponding radius rj .
An arc surface can be regarded as a generalized cylinder. To simplify the contact model,
this paper only considers the line contact. Therefore, the normal contact force between two
arcs can be evaluated by the formula appropriate for two cylinders [25, 26], which is given
by
      
1 ∗ δ 3 1 − cr2 δ̇
fn = πLE δ 1+ n, (8)
2 2 (R + δ) 4 δ̇ (−)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

322 P. Wang et al.

Fig. 4 Contact locations between a sprocket and tracks: (a) a connector and a tooth groove; (b) a track link
and the rubber ring

Fig. 5 Cross sections of a connector (a) and a tooth groove (b)

where δ is the penetration depth, δ̇ is the contact velocity, δ̇ (−) is the contact velocity at
the start of the contact, L is the contact length in the axial direction, cr is the restitution
coefficient, and n is the unit normal vector [27, 28]. E ∗ represents the composite Young’s
modulus and is evaluated as

1 1 − v12 1 − v22
∗
= + , (9)
E E1 E2

where E1 , E2 denote elastic moduli and v1 , v2 Poisson coefficients [29, 30].

Equation (8) can be employed to internal and external contacts depending on what the
parameter R represents. For internal contact, R represents the difference between the

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 323

cylinders’ radii, |R1 − R2 |. Otherwise, for external contact, R is quantified by the sum of
the cylinders’ radii, (R1 + R2 ) [17, 31].
To accurately describe the friction phenomena during contact, a continuous velocity-
based friction model [32–34] is adopted in this paper. It avoids numerical instability and
improves calculation efficiency. Neglecting the viscous friction term, the tangential friction
force is evaluated as
⎛ ⎞
⎜ vτ ⎟
⎜ vτ vt ⎟
f τ = ⎜fn μd tanh 4 + fn (μs − μd )  2 2⎟
sign(vτ )τ , (10)
⎝ vt ⎠
1
4
vτ
vt
+ 3
4

in which fn is the magnitude of normal force, vτ is the relative tangential velocity, vt is the
transition velocity, μs , μd are the coefficients of static and dynamic friction, and τ is the
unit tangential vector.
Similarly, the axial friction force is evaluated as
⎛ ⎞
⎜ vκ ⎟
⎜ vκ vt ⎟
f κ = ⎜fn μd tanh 4 + fn (μs − μd )  2 2⎟
sign(vκ )κ, (11)
⎝ vt ⎠
1
4
vκ
vt
+ 3
4

where vκ is the relative axial velocity and κ is the unit axial vector.

3 Contact algorithm between a connector and sprocket teeth

The track chain engagement exhibits plenty of contact areas. A high-efficiency geometric
contact algorithm is a prerequisite for contact force evaluation. First, the potential meshing
groove needs to be detected. Then the potential contact area of this groove is confirmed. Ul-
timately, the penetration depth is calculated according to the relative positions and velocities
between the connector and the area.

3.1 Potential meshing groove detection

The engagement of a connector and the sprocket is highlighted in Fig. 6. The left arc center
of the connector is point C. The position and velocity coordinates of point C and OS are R C ,
Ṙ C and R OS , Ṙ OS , respectively. They are all described in the inertial frame {O, X, Y, Z}.
The sprocket-fixed coordinate frame {OS , xS , yS , zS } has the OS xS axis pointing to the high-
est part of its one tooth. Due to the displacement limitation of center guides and track pins,
the relative movement between tracks and the sprocket can be expressed as motions in the
plane {OS , xS , yS }, a rotational and a translational motion about zS -axis. 
Grooves
 are indexed successively, and the local coordinate frame of groove k OS , xk ,
yk , zk is obtained by rotating the sprocket-fixed coordinate frame βk about the zS -axis. Its
OS yk axis points to the lowest part of groove k. The relative angle of these two frames βk is
expressed as

βk = (2k − 1)π/m − π/2, (12)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

324 P. Wang et al.

Fig. 6 The engagement of a

connector and the sprocket

where m = 13 corresponds to the number of sprocket teeth. The transformation matrix from
{O, X, Y, Z} to {OS , xk , yk , zk } is written as

Ak = AS Az (βk ), (13)

where AS is the orientation of {OS , xS , yS , zS } w.r.t. {O, X, Y, Z}, and the orientation of
{OS , xk , yk , zk } w.r.t. {OS , xS , yS , zS } is expressed as
⎡ ⎤
cos βk − sin βk 0
Az (βk ) = ⎣ sin βk cos βk 0 ⎦ . (14)
0 0 1

The relative position vector from OS to C in {OS , xS , yS , zS } is gained as

 T  
y
l S,OS C = lS,O
x
S C , lS,O S C , l z
S,O S C = ATS R C − R OS . (15)

The horizontal rotational angle of point C is

 y x
atan 2(lS,OS C , lS,O ) θC ≥ 0
θC = y
SC . (16)
x
atan 2(lS,OS C , lS,O SC
) + 2π θC < 0

The groove index of potential meshing for this connector is yielded as

k = θC /(2π/m) + 1. (17)

To ensure contact, the arc surface of the connector must not exceed the highest of the
tooth, i.e.,

y
x
(lS,O SC
)2 + (lS,OS C )2 ≤ ra + rC . (18)

Furthermore, the contact length in the axial direction must exist, i.e.,
 
z 
L = (tS + tC )/2 − lS,O SC
 > 0, (19)

where tC is the axial length of a connector and tS is the thickness of a tooth.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 325

Equations (18) and (19) are two necessary but not sufficient conditions for the engage-
ment. They should be checked first to avoid the later useless calculation if one of them is not
satisfied.

3.2 Potential contact area detection of a groove

A connector has two contact arc surfaces. In normal working conditions, the left arc will
only contact with areas 1–5 of a groove and the right arc contacts with areas 6–10. However,
with abnormal operation switching, the left or right arc may contact with any areas of the
groove. The contact analysis of left and right arcs needs to be proceeded, respectively. In the
following, the left arc contact is illustrated, and the right one can copy it.
A local frame {OS , xk , yk , zk } is attached to groove k to characterize its contact areas. It
makes that geometric quantities of the groove profile are invariant, resulting in a dramatical
reduction of the computational cost associated with potential contact arc determination. The
relative position vector from OS to C in {OS , xk , yk , zk } is yielded as
 T  
y
l k,OS C = lk,O
x
, l , l z
S C k,OS C k,OS C
= ATk R C − R OS . (20)

x
If lk,O SC
≤ 0, then the contact occurs in areas 1–5. And then the contact search follows a
sequence from area 1 to 5 one by one. Otherwise, it occurs in areas 6–10 and the search
sequence is from 10 to 6.
In case of the cylinder in which the left arc located rolls on the tooth surface of groove
k, the trajectory of its center is highlighted by the dotted line in Fig. 7a. Assuming that
the maximum penetration is rC , the contact between the left arc and groove k only occurs
when the center is located within the range of the groove profile and the center trajectory.
According to the indexes of groove areas, the located area of the center for potential contact
is discretized into ten annular sectors, as illustrated in Fig. 7b. After determining the sign of
x
lk,O SC
, the index of the annular sector for contact is calculated in order.
Although a groove has ten areas, they can be sorted into two categories: convex and
concave. Among them, areas 1, 4, 7, 10 are convex and areas 2, 3, 5, 6, 8, 9 are concave.
The detection for a center point C located in an annular sector j is illustrated as follows.
As shown in Fig. 8, Oj is the center of annular sector j . The position vectors of Oj and
xy xy xy
C w.r.t. OS in {OS , xk , yk } are l k,OS Oj and l k,OS C , respectively. The quantities of l k,OS Oj are
 T
xy y
listed in Table 1, and l k,OS C = lk,O x
,l
S C k,OS C
. For convex surfaces, the radius of the small
sector is rj and the radius of the large sector is (rj + rC ). For concave surfaces, the radius
of the large sector is rj and the radius of the small sector is (rj − rC ). The annular sector
expands by angle θ from both sides of the direction vector u. First, the distance between Oj
and C should be detected within the radius of the annular sector, i.e.,
⎧  
⎨ rj ≤ l xy xy 
k,OS C − l k,OS Oj  ≤ rj + rC j = 1, 4, 7, 10
  . (21)
⎩ rj − rC ≤ l xy xy 
k,OS C − l k,OS Oj  ≤ rj j = 2, 3, 5, 6, 8, 9

−−→
Subsequently, the angle between Oj C and u need to be satisfied as
⎛ ⎞
xy xy
l k,C − l k,Oj
cos−1 ⎝   · u⎠ ≤ θ. (22)
 xy xy 
l k,C − l k,Oj 

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

326 P. Wang et al.

Fig. 7 Trajectory of the left arc

center (a) and the discretization
(b) of its located area for
potential contact

Fig. 8 Characteristics of the convex (a) and concave (b) surfaces

Since division is generally much slower than multiplication in a computer implementation,

and cos θ decreases monotonously within θ ∈ [0, π], Eq. (22) reads as
   
xy xy  xy xy 
l k,C − l k,Oj · u ≥ l k,C − l k,Oj  cos θ. (23)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 327

Fig. 9 Contact between the left arc and area 1

For annular sector j , its spread angle θ is constant, leading to the pre-calculation of cos θ .
These are all operations for improving computational efficiency. Consequently, if Eqs. (21)
and (23) are all satisfied, the contact between the left arc and area j of groove k occurs.

3.3 Contact force evaluation

After detecting the area index, the contact forces can be evaluated in terms of the penetration
depth and velocity between the contacts. The evaluation process is different for convex and
concave surfaces.

3.3.1 Contact force evaluation for the convex surface

Figure 9 presents the contact between the left arc and area 1, where O1 is the arc center of
area 1 and C the connector left arc center. Take this as an example to give the contact force
calculation process for the convex surface.
In the inertial frame, the position coordinate of O1 is

R O1 = R OS + Ak l OS O1 . (24)

Its inertial velocity vector is expressed as

Ṙ O1 = Ṙ OS + Ω̃ S Ak l OS O1 . (25)

The relative position vector of C w.r.t. O1 is defined as the eccentricity vector e1 , which is
described in {OS , xk , yk , zk } as
 y T  
e1 = e1x , e1 , e1z = ATk R C − R O1 . (26)

By differentiation w.r.t. time, we obtain

 y T    
ė1 = ė1x , ė1 , ė1z = ATk Ṙ C − Ṙ O1 − ATk Ω̃ S R C − R O1 , (27)

where l k,OS O1 is the position vectors of O1 w.r.t. OS in {OS , xk , yk }.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

328 P. Wang et al.

Fig. 10 Contact between the left arc and area 2

The magnitude of e1 in the plane {OS , xk , yk } is

 xy   x 2  y 2
e  = e + e . (28)
1 1 1

Accordingly, the penetration depth is yielded as

 xy 
δ1 = rC + r1 − e1  (29)

in which rc and r1 are constant quantities. By differentiation w.r.t. time of Eq. (29) and
combining Eq. (28), the penetration velocity is
y y
e1x ė1x + e1 ė1
δ̇1 = −  xy  . (30)
e 
1

The normal forcecan be evaluated xy  by substituting δ1 and δ̇1 into Eq. (8), in which R =
rC + r1 , n = [e1x / e1  , e1 / e1  , 0]T .
xy y

The tangential velocity is

v1τ = ė1 · τ . (31)

 xy 
into Eq. (10) yields the tangential friction force, where τ = [−e1 / e1  ,
y
Substitution
 
e1x / e1  , 0]T .
xy

The axial velocity is

v1κ = ė1 · κ. (32)

Substitution into Eq. (11) yields the axial friction force, where κ = [0, 0, 1]T .

3.4 Contact force evaluation for the concave surface

Figure 10 presents the contact between the left arc and area 2, where O2 is the arc center of
area 2. Take this as an example to give the calculation process for the concave surface.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 329

Likewise, the eccentricity vector e2 and its derivative ė2 are acquired by using Eqs. (26)
and (27), in which all quantities with index 1 now refer to 2. Nevertheless, the evaluation of
the penetration depth is different, which is expressed as
 xy 
δ2 = e2  + rC − r2 . (33)

By differentiation w.r.t. time of Eq. (33), the penetration velocity is

y y
e2x ė2x + e2 ė2
δ̇2 =  xy  . (34)
e 
2

The normal force can  xybe  xy  by substituting δ2 and δ̇2 into Eq. (8), in which R =
 evaluated
r2 − rC , n = [−e1x / e2  , −e1 / e2  , 0]T . The calculations
y
 xy  of
 tangential
 and axial forces are
borrowed from that of area 1, in which τ = [e2 / e2  , −e2x / e2  , 0]T and all quantities with
y xy

index 1 now refer to 2.

3.5 Computational procedure of contact algorithm

The contact algorithm consists of geometric contact criteria and force evaluation. Its compu-
tational procedure is illustrated in Fig. 11. First, a sprocket-fixed frame is attached to detect
the index of the potential groove. Then contact analysis of the connector’s left and right arcs
is carried out independently. Strikingly, a local frame of the groove is attached to improve
the computational efficiency owing to the invariant of geometric quantities. These quantities
are pre-calculated in accordance with the tooth profile, as detailed in Table 1. Subsequently,
the penetration depth and velocity are assessed to evaluate the contact force. Eventually, the
forces of left and right arcs are integrated as the external force applying to the connector.

4 Contact algorithm between a track link and the rubber ring

As presented in Fig. 12, there are three contact cases between a track link and the rubber
ring: vertical contact, left end contact, and right end contact. The left and right ends of a link
are labeled A and B. In the plane {OS , xS , yS }, draw a vertical line from the center of the
sprocket OS to line AB, and the perpendicular is recorded as Q. The vertical line intersects
the outer contour of the rubber ring at P .
Take the vertical contact condition as an example to give the calculation process of con-
tact forces. A local frame Q, xf , yf , zf is located in Q, which has its yf axis going
through the sprocket center OS . Its orientation w.r.t. the inertial frame is

Af = AS Az (ϕ), (35)

where ϕ is the rotational angle from the OS yS axis to the Qyf axis. The transformation
matrix Az (ϕ) is calculated using Eq. (14).
In three cases, the penetration depth of the track link to the rubber ring is as follows:
⎧  
 
⎨ D/2 − H 1−2 AS (R Q − R P ) vertical contact
T
⎪
δr = D/2 − H 1−2 ATS (R A − R P ) left contact , (36)
⎪
⎩  
D/2 − H 1−2 AT (R B − R P )
S right contact

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

330 P. Wang et al.

Fig. 11 Flowchart of the contact algorithm between a connector and sprocket teeth

where H 1−2 = [H 1 H 2 ]T , H 1 = [1 0 0]T , H 2 = [0 1 0]T . R Q , R P , R A , R B are

position vectors in the inertial frame, respectively. The thickness of a link is D. Particularly,
H 1−2 ATS (R Q − R P ) represents the x- and y-direction components of the relative distance
between Q and P in {OS , xS , yS , zS }. The normal force is evaluated as
" #
4E ∗ R 1/2 3(1 − e2 ) δ̇r
fn = δr 1 + , δr ≥ 0, (37)
3 4 δ̇r(−)

where δr is the penetration depth, δ̇r is the penetration velocity, δ̇r(−) is the penetration veloc-
ity at the start of the contact, R refers to the radius of the rubber ring, and e is the recovery
coefficient.
Additionally, the tangential slip velocity between the two is
⎧ T T
⎨ H 1 Af Ṙ Q − (rr + D/2)ω
⎪ vertical contact
vτ = H T1 ATf Ṙ A − (rr + D/2)ω left contact , (38)
⎪
⎩ T T
H 1 Af Ṙ B − (rr + D/2)ω right contact

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 331

Table 1 Geometry of the sprocket and connector

Variable Description Value

ra Radius of the sprocket addendum circle 0.3384 m

rp Radius of the sprocket pitch circle 0.308 m
rC Radius of the connector left (right) arc 0.02025 m
LC Distance between the centers of the two arcs 0.06 m
of a connector
tS Thickness of the sprocket tooth 0.04 m
tC Axial length of the connector 0.048 m
y
(l xk,O O , l k,O O , l zk,O O ) Position vector from O1 to OS in (−0.0905, 0.2956, 0) m
S 1 S 1 S 1
{OS , xk , yk , zk }
y
(l xk,O O , l k,O O , l zk,O O ) Position vector from O2 to OS in (−0.0038, 0.3327, 0) m
S 2 S 2 S 2
{OS , xk , yk , zk }
y
(l xk,O O , l k,O O , l zk,O O ) Position vector from O3 to OS in (−0.03, 0.3065, 0) m
S 3 S 3 S 3
{OS , xk , yk , zk }
y
(l xk,O O , l k,O O , l zk,O O ) Position vector from O4 to OS in (−0.03, 0.2795, 0) m
S 4 S 4 S 4
{OS , xk , yk , zk }
y
(l xk,O O , l k,O O , l zk,O O ) Position vector from O5 to OS in (0, 0.2773, 0) m
S 5 S 5 S 5
{OS , xk , yk , zk }
r1 Radius of area 1 0.0343 m
r2 Radius of area 2 0.06 m
r3 Radius of area 3 0.023 m
r4 Radius of area 4 0.0041 m
r5 Radius of area 5 0.026 m
u1 Direction vector of area 1 in {OS , xk , yk } (0.6628, 0.7488)
u2 Direction vector of area 2 in {OS , xk , yk } (−0.8283, −0.5603)
u3 Direction vector of area 3 in {OS , xk , yk } (−0.3825, −0.9239)
u4 Direction vector of area 4 in {OS , xk , yk } (0.7316, 0.6817)
u5 Direction vector of area 5 in {OS , xk , yk } (−0.7316, −0.6817)
θ1 Spread angle of area 1 0.4421 rad
θ2 Spread angle of area 2 0.1907 rad
θ3 Spread angle of area 3 0.3925 rad
θ4 Spread angle of area 4 0.8207 rad
θ5 Spread angle of area 5 0.8207 rad

where rr is the radius of the rubber ring and ω is the rotational speed of the sprocket. Sub-
stitution into Eq. (10) yields the tangential friction force.
The axial slip velocity between the two is

⎧  
⎪
⎪ H T
A T
(Ṙ − Ṙ ) − Ω̃A T
(R − R ) vertical contact
⎪
⎪ 3 S Q P S Q P
⎨  
vκ = H T3 ATS (Ṙ A − Ṙ P ) − Ω̃ATS (R A − R P ) left contact . (39)
⎪
⎪  
⎪
⎪
⎩ H T AT (Ṙ B − Ṙ P ) − Ω̃AT (R B − R P ) right contact
3 S S

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

332 P. Wang et al.

Fig. 12 Contact cases between a track link and the rubber ring

Substitution into Eq. (11) yields the axial friction force. Moreover, the contact algorithm
between tracks and the rubber ring is also applicable to that between tracks and the road
wheel or idler.

5 Validation of the proposed approach

In this section, the proposed model is simulated by commercial software Recurdyn for com-
parison. Then the simulation and experimental results are analyzed to verify the correctness
of the proposed dynamic model.

5.1 Comparison with the simulation by Recurdyn

Using the high mobility tracked vehicle (HMTV) of commercial software Recurdyn, the
dynamical model of the double pin track in Fig. 1a is established, as shown in Fig. 13a.
The sprocket tooth profile is slightly different from the actual one due to the limitation of
this software. It is discretized into eight arcs, where the shape between points 2 and 8 is
the same as the actual shape, as detailed in Fig. 13b. The contacts between the sprocket
and tracks mainly occur in this area. Three revolute joints are applied at the sprocket, road
wheel, and idler. Meanwhile, driving torque is applied at the sprocket to keep the speed of
180 r/min.
The simulated vertical displacement of the selected track link is compared with the sim-
ulation result by the proposed approach, as exhibited in Fig. 14. The two are in good agree-
ment and the maximum difference is 6.03%.
The computational times of the proposed method and Recurdyn are shown in Table 2.
The (central processing unit) CPU used is an i7-8565u processor with maximum frequency
of 1.8 GHz. The calculations are all single-threaded. It can be seen that the calculation speed
of the proposed method is slightly faster than that of Recurdyn.
The joint constraint violation results in the closed-loop system are shown in Fig. 15. The
maximum distance between the two ends of the closed-loop system is 1.1023 × 10−10 m.
Therefore, the dynamics treatment for the closed-loop track system is appropriate.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 333

Fig. 13 The dynamical model established by Recurdyn (a) and its tooth profile (b)

Fig. 14 Cure of the vertical displacement

Table 2 Computational times of

the proposed method and Proposed method Recurdyn
Recurdyn
Time (s) 662 973

5.2 Comparative analysis of simulation and experimental results

The experimental framework is performed in Fig. 16. A motor drives the sprocket to rotate at
different speeds and the transmission ratio is 5.57. A wireless three-axis acceleration sensor

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

334 P. Wang et al.

Fig. 15 Constraint violation results in the closed-loop system

Fig. 16 Experimental framework

is installed on the selected track link. Its sampling rate is 1000 Hz and the range is ±20 g. A
signal collection collects the acceleration signal and transmits it to a computer. During the
rotation process, the centers of the three wheels are motionless, which can be regarded as
being connected to the base by three revolute joints. Until the speed stabilizes, the data is
carried out for processing and analysis, and the stable operation time is set to 90 s.
There is a constant acceleration of +1 g in the vertical direction when the wireless accel-
eration sensor is stationary horizontally. Therefore, to compare the simulation results with
the experimental results, it is necessary to add the projection of the constant acceleration on
the link-fixed frame to the simulated acceleration, i.e.,

a t = a s + ATt G, (40)

where a t is simulated acceleration for comparison, a s is the raw simulated acceleration, At is

the link-fixed frame, and G is the constant acceleration in the inertial frame, G = [0, g, 0]T .
Clearly, a t and a s are described in the link-fixed frame.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 335

Fig. 17 Time history curves of the y-direction components of simulation and experimental results in case of
different motor speeds: (a) 115 rpm; (b) 150 rpm; (c) 180 rpm; (d) 210 rpm

Table 3 Statistical values of the y-direction components of simulation and experimental results

Speed RMS SD
Simulation Experiment Value difference Simulation Experiment Value difference

115 rpm 1.0062 g 1.0587 g 4.96% 1.0044 g 1.0588 g 5.14%

150 rpm 1.0789 g 1.1795 g 8.53% 1.0758 g 1.1744 g 8.40%
180 rpm 1.1144 g 1.2770 g 12.73% 3.8161 g 1.2601 g 12.09%
210 rpm 1.5056 g 1.4069 g 7.02% 1.4964 g 1.3849 g 8.05%

The y-direction component of the simulation result is compared with the experimental
result in case of different motor speed: 115 rpm, 150 rpm, 180 rpm, and 210 rpm, respec-
tively. Figure 17 details their time history curves in about two adjacent periods for clarity.
This acceleration component during all the stable operation is statistically analyzed, and the
root mean square (RMS) and standard deviation (SD) are highlighted in Table 3. Moreover,
Fig. 18 presents their corresponding frequency spectrums of these acceleration components.
The sample rate of fast Fourier transform (FFT) is 1000 Hz, and the analysis point is 90,000.
The comparison shows that the influence of the speed on the simulated and experimental
track vertical acceleration is consistent, as reflected in the following:
(1) The vertical acceleration of the track link increases with the increase of the motor
speed.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

336 P. Wang et al.

Fig. 17 (Continued)

Table 4 RMS of the vertical accelerations simulated by the contact algorithm in Ref. [2]

Speed Simulation Experiment Value difference

115 rpm 1.9369 g 1.0587 g 82.95%

150 rpm 2.4070 g 1.1795 g 100.07%
180 rpm 2.6772 g 1.2770 g 109.65%
210 rpm 3.3915 g 1.4069 g 141.06%

(2) At different speeds, the RMS differences between the simulation and the experiment
are 4.96%, 8.53%, 12.73%, and 7.02%, respectively. The SD differences are 5.14%, 8.40%,
12.09%, and 8.05%, respectively. They are all less than 15%.
Meanwhile, we compared the experimental results with the simulation results using the
contact algorithm in Ref. [2]. Their RMS are highlighted in Table 4. The algorithm proposed
in this paper has better computational accuracy because it takes into account the tooth profile
more finely.
Although the simulation and experimental results have the same variation tendency, the
experimental acceleration changes more dramatically. Three reasons might account for this
phenomenon. First, the actual contact duration is very short, but the simulation step size
in this paper is larger for efficiency. The second one is about the connection between the
sensor and the track. In the simulation, they are ideally consolidated. But in the experiment,
they stick together with glue, so the two may have a relative impact resulting in a large

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 337

Fig. 18 Frequency spectrums of the y-direction components of simulation and experimental results in case
of different motor speeds: (a) 115 rpm; (b) 150 rpm; (c) 180 rpm; (d) 210 rpm

acceleration. The third one is about the sprocket speed. The simulated sprocket speed is
constant, but the experimental speed is not steady. It fluctuates between the expected speed.
Through comparative analysis of acceleration spectrums at different speeds, the main
conclusions are drawn below:
(1) The main frequencies and amplitudes of simulation results at 115 rpm, 150 rpm,
180 rpm, and 210 rpm are 0.122 Hz and 0.793 g, 0.167 Hz and 1.174 g, 0.200 Hz and
1.167 g, 0.233 Hz and 1.162 g. The corresponding experimental results are 0.131 Hz and
1.083 g, 0.176 Hz and 1.1200 g, 0.2090 Hz and 1.1720 g, 0.253 Hz and 0.7698 g. This
suggests that the main frequencies and amplitudes of simulation results are consistent with
the experimental results at different speeds. The main frequency corresponds to the time
required for one revolution of the track ring. However, the main frequencies of simulation
are always slightly less than the experimental values. The possible reason is that the sprocket
does not maintain the expected speed on account of the track vibration.
(2) The collisions between tracks and wheels cause the high-frequency track vibration.
With the speed increasing, the frequency in the high range of the acceleration spectrum
increases significantly.

6 Influence of the sprocket rubber ring

In Sect. 5, experimental results verify the accuracy of the dynamic model established in
this paper. Therefore, this section analyzes the influence of the sprocket rubber ring on the
system dynamical characteristics based on numerical simulation.

6.1 Contact force variation with and without the rubber ring

Early sprockets were all metal without rubber rings. The contact between tracks and the
sprocket only occurs in tooth grooves. Two demonstration cases are compared to illustrate
the importance of rubber rings. These two cases have the same simulation conditions, except
for one with a rubber ring and one without. The motor speed is set to 115 rpm.
First, the case without the rubber ring is carried out, and Fig. 19a depicts the normal
contact forces of the selected connector during the meshing process. Strikingly, both the
left and right arc surfaces are in contact with the groove. Figure 19b shows the groove area
indexes corresponding to the left and right arc contacts of the connector. A screenshot of

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

338 P. Wang et al.

Fig. 18 (Continued)

animation of the connector meshing with the sprocket is detailed in Fig. 20. It should be
noted that the track center guides are not displayed for focusing on the engagement. At
t = 4.2 s, the connector is about to go into engagement. At t = 4.29 s, the contact between
area 7 of groove 1 and the right arc surface occurs firstly, and the normal force reaches
its maximum value at t = 5.36 s. At t = 4.29 s, the contact between area 2 and the left arc
occurs, and its maximum appears at t = 4.46 s. At t = 5.58 s, only the left arc contact exists,
and this connector is out of engagement at t = 5.6 s.
Then the case with the rubber ring is carried out. Figure 21a depicts the normal contact
forces between the same connector and the sprocket. Meanwhile, the normal force between
its adjacent track link and the rubber ring is also depicted. Figure 21b shows the groove area
indexes corresponding to the left and right arc contacts of the connector. Obviously, the right

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 339

Fig. 19 Normal contact forces (a) of the selected connector and their corresponding groove area index (b)

Fig. 20 Screenshots of the selected connector meshing with the sprocket without the rubber ring

arc is not in contact with the sprocket, and the contact duration and force magnitude of the
left arc are distinctly less than the first case. Similarly, several screenshots at crucial times

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

340 P. Wang et al.

Fig. 21 Normal contact forces (a) between the connector, its adjacent link, and the sprocket, and their corre-
sponding groove area index (b)

are presented in Fig. 22. At t = 4.05 s, the contact between the link and the rubber ring
occurs firstly, and the normal force reaches its maximum value at t = 4.27 s. At t = 5.3 s,
the contact between area 2 and the left arc occurs, and its maximum appears at t = 5.47 s.
After t = 5.48 s, the link is separated from the rubber ring, and then the connector is out of
the tooth at t = 5.56 s.
In view of the whole process, the contact force is mainly borne by the rubber ring, and
only area 2 has contact. The force variation is also smaller than in the first case. In this way,
the deformation of sprocket teeth can be effectively reduced, thereby reducing the wear of
the tooth profile and improving its service life.

6.2 Appropriate radius of the rubber ring

The rubber ring effectively alleviates the meshing forces on the sprocket teeth. Previous
studies have reported that the contact radius between the track ring and the sprocket is im-
portant, which pertains to the rubber radius [4]. For comparison, different radii from 0.303 m
to 0.313 m are simulated, among which 0.308 m corresponds to the equal-pitch.
The RMSs and SDs of normal contact forces between an appointed connector and the
rubber ring, the left and right arcs of its meshing tooth groove are listed in Fig. 23. Several
conclusions can be drawn from our findings:
(1) Faster speed results in larger force magnitude and variation, especially in the left arc
contact. However, this influence is limited.
(2) The force magnitude and variation increase significantly with the radius of rubber
ring.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 341

Fig. 22 Screenshots of the same connector and its adjacent link connecting with the sprocket

(3) When the radius is greater than or equal to 0.305 m, there is no right arc contact.
(4) The radius has practically no effect on the left arc contact.
The percentage of contact duration in one turn of the sprocket is also important. Pro-
longed left and right arc contact may aggravate tooth wear. When the radius is greater than
or equal to 0.307 m, the left contact duration percentage remains near 10%, as shown in
Fig. 24. Since the rubber ring is elastic and replaceable, the main contact force should be
applied on it.
Finally, the above three factors are comprehensively considered to reasonably select the
radius of rubber ring as 0.308 m.

7 Conclusions

Based on MSTMM, a dynamical model of a double pin caterpillar platform is developed,

which consists of two kinematically decoupled subsystems: tracks and wheels. These two
are coupled by contact forces. By using a generalized cylindrical contact model, the geo-
metric contact criteria and force evaluation between double pin tracks and a sprocket are
developed. A tooth groove profile of the sprocket is divided into ten contact areas, which are
sorted into two categories: convex and concave. They correspond to external and internal
contact, respectively. Accordingly, a general mathematical description of the tooth profile is
established to describe the real tooth groove geometry. This description is also suitable for
single-pin track meshing pairs and chain drives and has a wide application prospect. Taking
into account the supporting effect of the rubber ring, the contact between tracks and the rub-

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

342 P. Wang et al.

Fig. 23 RMSs (a) and SDs (b) of normal contact forces for different contact cases

Fig. 24 Percentages of contact duration in one turn of the sprocket for different contact cases

ber ring is treated as a cylindrical contact with a plane. It is also applicable to that between
tracks and the road wheel or idler.
In this paper, a sprocket-fixed frame is employed in the groove index. Depending on the
position coordinates of a connector in this frame, the groove meshing with it is determined.
Then a local frame of this groove is employed to search the potential contact area index.
By using these two frames, the vector quantities and coordinate transformations describing
the contact arc surfaces are constant, which can be precalculated to improve calculation
efficiency. Lastly, the penetration depth and velocity are calculated depending on the relative

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 343

positions and velocities of the connector and contact surface. The force is expressed as an
explicit function of kinematic information, avoiding the iterative process.
Subsequently, a field test is performed to verify the correctness of the dynamical model.
The vertical acceleration of a track plate is measured by a wireless sensor. Then simulation
results are compared with this test data in time domain and frequency domain, leading to
the validity of the proposed model. Ultimately, this model is used to simulate and analyze
the influence of the sprocket rubber ring on the meshing characteristic, and a reasonable
rubber ring radius is given. Our investigations can be used in the virtual prototype tech-
nology for tracked vehicles to improve simulation accuracy and efficiency. In the future,
dynamic simulation is important in the design period of tracked vehicles, which can shorten
the development period and reduce the development cost.
However, our work has some limitations. The main limitation of the methodology pro-
posed in this paper relies on the assumption of the tooth groove consisting of arc surfaces. It
is inappropriate for the groove comprised of flat surfaces. Future studies will be performed
to develop a more general mathematical model of the tooth profile.

Nomenclature
Variable Description Unit
ẍ, ÿ, z̈ Translational accelerations in the inertial frame m/s2
Ωx , Ωy , Ωz Angular accelerations in the inertial frame rad/ s2
mx , my , mz Internal moments in the inertial frame N·m
qx , q y , q z Internal forces in the inertial frame N
In An identity matrix with the dimension of n × n /
Ui Transfer matrix of element i /
φi Relative angle about the z-axis of a joint i rad
A Transformation matrix /
fn Normal contact force vector N
fτ Tangential friction force vector N
fκ Axial friction force vector N
δ Penetration depth m
δ̇ Contact velocity m/s
R Position vector in the inertial frame m
Ṙ Velocity vector in the inertial frame m/s
l k,OS C Position vector from OS to C in {OS , xk , yk , zk } m
ra Radius of the sprocket addendum circle m
rp Radius of the sprocket pitch circle m
rC Radius of the connector left (right) arc m
LC Distance between the centers of the two arcs of a connector m
tS Thickness of the sprocket tooth m
tC Axial length of the connector m
rj Radius of area j m
uj Direction vector of area j in {OS , xk , yk } /
θj Spread angle of area j rad
a Acceleration m/s2
G Acceleration of gravity in the inertial frame m/s2
Author contributions Assoc. Prof. Pingxin Wang wrote the main manuscript text and provided the funding
acquisition. Prof. Xiaoting Rui and Prof. Guoping Wang contributed to the conception of the study. Prof.
Hailong Yu and Prof. Bin He helped perform the analysis with constructive discussions. Assoc. Prof. Junjie
Gu provided the funding acquisition.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

344 P. Wang et al.

Funding This work was supported by the Natural Science Foundation of Jiangsu Province (No. BK20210321),
the National Natural Science Foundation of China (No. 11972193), and the National Natural Science Foun-
dation of China (No. 12202196).

Data Availability The datasets generated during and/or analyzed during the current study are available from
the corresponding author on reasonable request.

Declarations

Competing interests The authors declare no competing interests.

References
1. Zhao, Z., Liu, H., Chen, H., Hu, J., Guo, H.: Kinematics-aware model predictive control for autonomous
high-speed tracked vehicles under the off-road conditions. Mech. Syst. Signal Process. 123, 333–350
(2019)
2. Wang, P., Yu, H., Rui, X., Zhang, J., Gu, J.: Transversal vibration analysis of the upper span of nonlinear
closed-loop track systems. Appl. Math. Model. 78, 249–267 (2020)
3. Wallin, M., Aboubakr, A.K., Jayakumar, P., Letherwood, M.D., Gorsich, D.J., Hamed, A., Shabana,
A.A.: A comparative study of joint formulations: application to multibody system tracked vehicles. Non-
linear Dyn. 74(3), 783–800 (2013)
4. Wang, P., Wang, G., Rui, X., Yu, H., Zhang, S.: Contact dynamics analysis of the single-pin meshing
pair of a tracked vehicle. Nonlinear Dyn. 104(2), 1139–1155 (2021)
5. Scholar, C., Perkins, N.C.: Efficient vibration modelling of elastic vehicle track systems. J. Sound Vib.
228(5), 1057–1078 (1999)
6. Huh, K., Hong, D.: Track tension estimation in tacked vehicles under various maneuvering tasks. J. Dyn.
Syst. Meas. Control 123(2), 179–185 (2001)
7. Sandu, C., Freeman, J.S.: Military tracked vehicle model, part I: multibody dynamics formulation. Int. J.
Veh. Syst. Model. Test. 1(1), 48–67 (2005)
8. Sandu, C., Freeman, J.S.: Military tracked vehicle model. Part II: case study. Int. J. Veh. Syst. Model.
Test. 1(1), 216–231 (2005)
9. Ma, Z., Perkins, N.C.: A super-element of track-wheel-terrain interaction for dynamic simulation of
tracked vehicles. Multibody Syst. Dyn. 15(4), 347–368 (2006)
10. Lee, J.H., Choi, J.H., Shabana, A.A.: Spatial dynamics of multibody tracked vehicles part II: contact
forces and simulation results. Veh. Syst. Dyn. 29(2), 113–137 (1998)
11. Nakanishi, T., Shabana, A.A.: Contact forces in the non-linear dynamic analysis of tracked vehicles. Int.
J. Numer. Methods Eng. 37(8), 1251–1275 (1994)
12. Gat, G., Franco, Y., Shmulevich, I.: Fast dynamic modeling for off-road track vehicles. J. Terramech. 92,
1–12 (2020)
13. American, N.: Standards Institute, Precision Power Transmission Roller Chains, Attachments, and
Sprockets (1975). ANSI Standard B29.1-1975, New York
14. Ambrósio, J., Pombo, J.: A unified formulation for mechanical joints with and without clearances/bush-
ings and/or stops in the framework of multibody systems. Multibody Syst. Dyn. 42(3), 317–345 (2018)
15. Ambrosio, J., Malça, C., Ramalho, A., Ramalho, A.: Planar roller chain drive dynamics using a cylindri-
cal contact force model. Mech. Based Des. Struct. Mach. 44(1–2), 109–122 (2016)
16. Pereira, C., Ambrósio, J., Ramalho, A.: Dynamics of chain drives using a generalized revolute clearance
joint formulation. Mech. Mach. Theory 92, 64–85 (2015)
17. Pereira, C.M., Ramalho, A.L., Ambrósio, J.A.: A critical overview of internal and external cylinder
contact force models. Nonlinear Dyn. 63(4), 681–697 (2011)
18. Tian, Q., Flores, P., Lankarani, H.M.: A comprehensive survey of the analytical, numerical and experi-
mental methodologies for dynamics of multibody mechanical systems with clearance or imperfect joints.
Mech. Mach. Theory 122, 1–57 (2018)
19. Rui, X., Wang, X., Zhou, Q., Zhang, J.: Transfer matrix method for multibody systems (Rui method) and
its applications. Sci. China, Technol. Sci. 62(5), 712–720 (2019)
20. Wang, P., Rui, X., Yu, H.: Study on dynamic track tension control for high-speed tracked vehicles. Mech.
Syst. Signal Process. 132, 277–292 (2019)
21. Rui, X., Bestle, D.: Reduced multibody system transfer matrix method using decoupled hinge equations.
Int. J. Mech. Syst. Dyn. 1(2), 182–193 (2021)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Dynamics of double pin caterpillar platform using a generalized. . . 345

22. Rui, X., Zhang, J., Zhou, Q.: Automatic deduction theorem of overall transfer equation of multibody
system. Adv. Mech. Eng. 6, 1–12 (2015)
23. Rui, X., Zhang, J., Wang, X., Rong, B., He, B., Jin, Z.: Multibody system transfer matrix method: the
past, the present, and the future. Int. J. Mech. Syst. Dyn. 2(1), 3–26 (2022)
24. Corral, E., Meneses, J., Castejón, C., García-Prada, J.C.: Forward and inverse dynamics of the biped
PASIBOT. Int. J. Adv. Robot. Syst. 11(7), 109 (2014)
25. Wang, G., Liu, C.: Further investigation on improved viscoelastic contact force model extended based
on Hertz’s law in multibody system. Mech. Mach. Theory 153, 103986 (2020)
26. Liu, C., Zhang, K., Yang, R.: The FEM analysis and approximate model for cylindrical joints with
clearances. Mech. Mach. Theory 42(2), 183–197 (2007)
27. Rodrigues Da Silva, M., Marques, F., Tavares Da Silva, M., Flores, P.: A compendium of contact force
models inspired by Hunt and Crossley’s cornerstone work. Mech. Mach. Theory 167, 104501 (2022)
28. Corral, E., Moreno, R.G., García, M.J.G., Castejón, C.: Nonlinear phenomena of contact in multibody
systems dynamics: a review. Nonlinear Dyn. 104(2), 1269–1295 (2021)
29. Bai, Z., Xu, F., Zhao, J.: Numerical and experimental study on dynamics of the planar mechanical system
considering two revolute clearance joints. Int. J. Mech. Syst. Dyn. 1(2), 256–266 (2021)
30. Liu, J., Zhao, F., Xiao, Z., Wang, Y., Liu, Z., Zheng, H.: The establishment of the normal contact
force model for a one-dimensional sphere chain subjected to impact load. Int. J. Mech. Syst. Dyn. 2(1),
131–142 (2022)
31. Corral, E., Gismeros Moreno, R., Meneses, J., García, M.J.G., Castejón, C.: Spatial algorithms for geo-
metric contact detection in multibody system dynamics. Mathematics 9(12), 1359 (2021)
32. Marques, F., Flores, P., Claro, J.C.P., Lankarani, H.M.: Modeling and analysis of friction including rolling
effects in multibody dynamics: a review. Multibody Syst. Dyn. 45(2), 223–244 (2019)
33. Pennestrì, E., Rossi, V., Salvini, P., Valentini, P.P.: Review and comparison of dry friction force models.
Nonlinear Dyn. 83(4), 1785–1801 (2016)
34. Brown, P., McPhee, J.: A continuous velocity-based friction model for dynamics and control with phys-
ically meaningful parameters. J. Comput. Nonlinear Dyn. 11(5), 54502 (2016)

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a pub-
lishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript
version of this article is solely governed by the terms of such publishing agreement and applicable law.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center
GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers
and authorised users (“Users”), for small-scale personal, non-commercial use provided that all
copyright, trade and service marks and other proprietary notices are maintained. By accessing,
sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of
use (“Terms”). For these purposes, Springer Nature considers academic use (by researchers and
students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and
conditions, a relevant site licence or a personal subscription. These Terms will prevail over any
conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription (to
the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of
the Creative Commons license used will apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may
also use these personal data internally within ResearchGate and Springer Nature and as agreed share
it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not otherwise
disclose your personal data outside the ResearchGate or the Springer Nature group of companies
unless we have your permission as detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial
use, it is important to note that Users may not:

1. use such content for the purpose of providing other users with access on a regular or large scale
basis or as a means to circumvent access control;
2. use such content where to do so would be considered a criminal or statutory offence in any
jurisdiction, or gives rise to civil liability, or is otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association
unless explicitly agreed to by Springer Nature in writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a
systematic database of Springer Nature journal content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a
product or service that creates revenue, royalties, rent or income from our content or its inclusion as
part of a paid for service or for other commercial gain. Springer Nature journal content cannot be
used for inter-library loans and librarians may not upload Springer Nature journal content on a large
scale into their, or any other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not
obligated to publish any information or content on this website and may remove it or features or
functionality at our sole discretion, at any time with or without notice. Springer Nature may revoke
this licence to you at any time and remove access to any copies of the Springer Nature journal content
which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or
guarantees to Users, either express or implied with respect to the Springer nature journal content and
all parties disclaim and waive any implied warranties or warranties imposed by law, including
merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published
by Springer Nature that may be licensed from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a
regular basis or in any other manner not expressly permitted by these Terms, please contact Springer
Nature at

[email protected]