Motion sensing study on a mobile robot through simulation model and
experimental tests
PIERANGELO MALFI, ARMANDO NICOLELLA, MARIO SPIRTO, CHIARA COSENZA,
VINCENZO NIOLA*, SERGIO SAVINO
Department of Industrial Engineering
University of Naples Federico II
Via Claudio 21 – 80125, Naples, Italy
ITALY
Abstract: - The employment of mobile robots, remotely or autonomously controlled for field operations, is an
emerging topic. The development of suitable simulation models can aid the design and testing of such mobile
robots. Here a small commercial prototype of a Rocker-Bogie Rover has been employed in experimental tests
compared to simulations on a multibody model. In particular, the tests have dealt with the acquisition and
processing of accelerometer signals both on experimental prototype and in simulation environment. The paper
proposes a real time signal processing approach that could aid the motion planning and rover navigation control.
Key-Words: - Rover, Rocker-Bogie, signal processing, FFT, IMU
Received: May 15, 2021. Revised: April 17, 2022. Accepted: May 16, 2022. Published: July 6, 2022.
1 Introduction
In last years, academic community and industry
have been showing a growth of interest in
unmanned service robotics and vehicles for field
operations [1],[2],[3],[4]. For instance, some
autonomous vehicles have been designed for
agricultural tasks [5],[6],[7]. Field operation
rovers, similarly, to space exploration rovers
such as Curiosity and Perseverance, are
autonomous or remotely controlled vehicles able
to tread on uneven terrain [8] and perform
several operations [9],[10].
The object of this paper is a small rover prototype
that is equipped with a suspension system named
Rocker-Bogie [11], employed in some of most
recent Mars exploration rovers.
The Rocker-Bogie [12],[13],[14],[15] is an
articulated, passive suspension system whose
purpose is to provide stability to a six-wheeled
rover over rough terrain in a slow speed mode. The
adoption of this articulated suspension system
allows all wheels to keep contact on the ground and
be independently actuated by six motors. The
geometry of the suspension is designed so that the
weight of the vehicle can be assumed equally
distributed among the wheels of the rover. A
schematic representation of this suspension system
is shown in Fig. 1.
Figure 1 –Rocker-Bogie suspension system: sketch
and schematic representation of the main parts.
One of the main challenges with this type of system
is the control of the trajectory followed by the rover.
To face this problem, the employment of vision
WSEAS TRANSACTIONS on APPLIED and THEORETICAL MECHANICS
DOI: 10.37394/232011.2022.17.11
Pierangelo Malfi, Armando Nicolella,
Mario Spirto, Chiara Cosenza,
Vincenzo Niola, Sergio Savino
E-ISSN: 2224-3429
79
Volume 17, 2022
system can enhance robot trajectory control [16],
[17]. C. Chen et al. [18] developed a method to
control a planetary rover trajectory based on both
kinematic and dynamic control, using rover motion
data and the forces acting on it. The acquired data
from such controls are angular velocities which,
appropriately processed by the controller, serve as
inputs to the motors. Such control system makes
possible to decrease the internal forces acting
between the same side wheels at least 60 percent and,
at the same time, obtain the pursuit of the desired
trajectory.
S. Wu et al. [19] developed a control algorithm, based
on exponential type law, on the traction coordination
of a lunar rover traveling over rough terrain. The
algorithm is designed to optimize the traction force
and minimize the slip ratio, control the velocity and
normal force. The validity of the proposed algorithm
has been demonstrated through simulations for a
rover moving along slopes and ditches. The proposed
approach can minimize slip ratios and mechanical
wear and tear, improve mobility, and increase the
overall durability of the rover. The enrichment of
multibody simulation model with experimental data
such as vision data has been showed in previous
works in robotics context [20],[21],[22]. M.
Thianwiboon et al [23], starting from the study of the
kinematics of a six-wheeled rover with Rocker-
Bogie suspension, developed a traction control
system (TCS) based on the slip angle that is
estimated from the experimental measurement of the
rolling speed of the wheels and the rover velocity,
through both simulation and experimental tests. In
a previous work [24], the study of the variation in the
distribution of the load among the wheels has been
presented for a rover. In addition, the rover stability
has been studied for different ground geometries
through a kinematic analysis to define the
suspension configuration as a function of ground
geometry. Moreover, it has been evaluated the
possibility of equipping the passive hinge, that
connects the rocker with the bogie, by a torsional
adjustable preloaded spring [25].
In this paper, a signal processing approach has been
proposed through the introduction of a new
parameter that can be used as an additional index for
motion planning of a six-wheeled rover featured by
Rocker-Bogie suspension system. The adoption
sophisticated signal processing has allowed easier
control of complex systems [26].
2 Prototype and Multibody Model
A prototype rover with Rocker-Bogie suspension
(mentioned above) was built to make a comparison
between experimental analysis and computer
simulation results. Several software have been
employed to create a multibody model for
the simulations. The Rover prototype is presented in
Fig. 2.
Figure 2 – Rover prototype.
The prototype (manufactured by Actobotics®) is
composed by: articulated chassis, six wheels, and six
motors. Simplicity, low cost, and low weight led to
this choice. The chassis is ABS 3D printed and is
sized in order to have similar vertical loads on the six
wheels while their axes are on the same horizontal
plane. The six non-steered wheels rover changes
trajectory by skid steering [27]. On the tire tread there
are four lines of twenty-four tire blocks
circumferentially organized. Each DC electric motors
(TT Right Angle Gear Motor type) is coupled to a
gearbox (48:1 gear ratio) that slows down the rotation
and changes the direction of output shaft.
The hardware for rover control consists of an
Arduino Uno WiFi Rev 2 controller board and two
Adafruit Motor Shield V2 boards for the six motors
drive. To avoid voltage drops, two different power
systems were used: a battery pack consisting of 4 AA
batteries (1.5 V each) used to power the controller,
and a 5 V, 2A USB powerbank to power the motor
shields. A communication logic based on the User
Datagram Protocol (UDP) was used for WiFi remote
control. For this purpose, the Osoyoo WiFi UDP
Robot Car Controller App was chosen: this App is
preset to control a mobile robot by sending
commands by sending alphanumeric characters. The
CAD was modelled in Solidworks®, which was then
imported into the MatLab®/Simulink® environment
for multibody modeling.
WSEAS TRANSACTIONS on APPLIED and THEORETICAL MECHANICS
DOI: 10.37394/232011.2022.17.11
Pierangelo Malfi, Armando Nicolella,
Mario Spirto, Chiara Cosenza,
Vincenzo Niola, Sergio Savino
E-ISSN: 2224-3429
80
Volume 17, 2022
Figure 3 – Multibody model scheme.
Fig. 3 shows the first level of modeling in Simulink,
which consists of three macroblocks:
World Frame, System Config. and Solver;
Environment Multibody;
Rover Multibody.
In the second macroblock, the ground and obstacles
were modeled.
The third macroblock is composed of the multibody
models of the chassis, the six wheels with their own
wheel-to-ground contact. Moreover, there are the six
Simscape® DC-motors and gearboxes models.
3 Experimental Phase
The embedded LSM6DS3TR IMU on Arduino board
was used as sensor to compare the obtained
experimental results and simulations output. Through
the IMU, It is possible to acquire translational
accelerations and rotational velocities along and
around the three fundamental axes. The reference
system used, and shown in Fig. 4, is the same as that
shown on the Arduino board.
Figure 4 – IMU sensor reference frame.
The obtained accelerometer signals from the
prototype and the model, were acquired in front-
rocker and front-bogie configurations in a
straight-line motion. Table 1 summarizes the
acquired signals.
Table 1 – Signal acquisition parameters
SIGNAL
FRONT
CONFIGURATION
SOURCE
SAMPLING
FREQUENCY
[Hz]
SUPPLY
VOLTAGE
[V]
NO-LOAD
SPEED
[RPM]
1
Bogie
Model
90
6
230
2
Rocker
Model
90
6
230
3
Bogie
Prototype
400
6
230
4
Rocker
Prototype
400
6
230
WSEAS TRANSACTIONS on APPLIED and THEORETICAL MECHANICS
DOI: 10.37394/232011.2022.17.11
Pierangelo Malfi, Armando Nicolella,
Mario Spirto, Chiara Cosenza,
Vincenzo Niola, Sergio Savino
E-ISSN: 2224-3429
81
Volume 17, 2022
The no-load speed used in the last column of the
Table 1 are available in the data sheet of the DC
electric motor. Signals three and four were
sampled by the IMU and the FFT was performed
on the Arduino board in Real-Time and then
saved on an SD card connected to the controller.
4 Results
The sampled accelerometer signals were analyzed
by the Fast Fourier Transform (FFT) with the
aim of showing the presence of isolated peaks
in the Transform, which were found to be the
rotational frequencies of the wheels. The
analysis of the accelerometer signals was carried
out for all three directions in space, but only y-
axis will be shown below because more significant.
Fig. 5 shows the FFT of signals 1 and 2 from Table
1. In both FFTs, despite the low sampling rate and
the presence of noise, two peaks corresponding
to a frequency of 3.7 Hz are clear in both
configurations. In the Fig. 5 (left) the rover
advances with the bogie
at the front; in the Fig. 5 (right) it advances with the
rocker at the front.
The is obtained by no-load speed at 6 DC
Voltage as in the electric motor datasheet. The
corresponds to the real peak frequency obtained by
FFT analysis. The difference between the theoretical
value obtained from the motor data sheet and that of
the simulation is only 0.1 Hz, that leads to a
percentage error of 2.6 %. The achieved result
provides insight to validate the multibody model.
Figure 5 – FFTs calculated on accelerometer signals of model in straight line motion: left) bogie front; right)
rocker front.
In Fig. 6 is shown the FFT of signals 3 and 4 from
Table 1. The two peaks are clearly recognizable,
respectively at 87.8 Hz ( for bogie at the
front Fig. 6 (left) and 83.1 Hz ( for
rocker at the front Fig. 6 (right).
These two values are significantly higher than the
previously nominal values of 3.8 Hz. This
discrepancy can be addressed to the fact that on the
generic rover wheel tire tread there are four lines of
twenty-four tire blocks circumferentially organized
(Fig. 7).
WSEAS TRANSACTIONS on APPLIED and THEORETICAL MECHANICS
DOI: 10.37394/232011.2022.17.11
Pierangelo Malfi, Armando Nicolella,
Mario Spirto, Chiara Cosenza,
Vincenzo Niola, Sergio Savino
E-ISSN: 2224-3429
82
Volume 17, 2022
Figure 6 – FFTs calculated on accelerometer signals of prototype in straight line motion: left) bogie at the front;
right) rocker at the front.
It can be said that the sensor (IMU), during only one
wheel rotation, can detect the contact between the
individual block and the ground, an event that is
repeated 24 times in one revolution of the tire.
Indeed, in case of bogie at the front:
In case of rocker at the front:
Figure 7 – Rover wheel tire tread with four lines of
twenty-four tire blocks circumferentially organized.
The differences between the real and nominal values,
0.2 Hz in the first case and 0.3 Hz in the second, could
be due to the backlash between the Rocker-Bogie
suspension components, the deformability of the
parts (in the model they were assumed rigid), and the
non-uniform distribution of grease within the
gearbox in each of the six motor groups.
The percentage errors in the experimental tests are
higher than the simulation case. The reason of such
difference could be addressed to a lower modelled
friction than the actual friction in the experimental
setup.
This analysis is still valid in the case of the rover
performs a skid steering where the two sides are at
different speeds. For this purpose, it was decided to
acquire from the embedded IMU, by frequency of 90
Hz, a signal, post-processed by FFT in MatLab®,
during a left-hand steering in a configuration with
bogie at the front.
In this case, the motors on the right side were
supplied by 2.6 V, that corresponds to 93 rpm, and
WSEAS TRANSACTIONS on APPLIED and THEORETICAL MECHANICS
DOI: 10.37394/232011.2022.17.11
Pierangelo Malfi, Armando Nicolella,
Mario Spirto, Chiara Cosenza,
Vincenzo Niola, Sergio Savino
E-ISSN: 2224-3429
83
Volume 17, 2022
those on the left by 1.6 V, that are equal to 57 rpm.
Such wheel speeds correspond to a frequency of 1.5
Hz and 0.95 Hz, respectively. Once the accelerometer
signal acquisition process was completed, the FFT
shown in Fig. 8 was computed.
Figure 8 – FFT calculated on accelerometer signal
of prototype in skid steering maneuver with bogie at
the front.
In Fig.8, two peaks are clearly distinguishable: the
first peak corresponds to 23.2 Hz, while the second
one corresponds to 37.3 Hz. Such test is still
characterized by the presence of the multiplicative
factor due to the number of the tire tread blocks.
Thus, dividing these values by 24, the real frequency
values are 0.96 and 1.55 Hz. Both differs by only 0.01
Hz from expected values, which the first is 0.95 Hz,
while the second is 1.54 Hz; that leads to percentage
error calculation for both Left and Right sides.
5 Conclusions
This work dealt with the study of the motion sensing
of a mobile robot through simulation and
experimental combined approach.
The analysis on the accelerometer signals made it
possible to show, both experimentally and in
simulation, that by means of FFT analysis, it was
possible to distinguish sharply peaks in the
Transform. These frequencies corresponded to the
rover wheel rotations. In simulation, these
frequencies were completely coincident with those
obtained theoretically from the data sheet of the
electric DC motors, while in the experimental tests
they were multiplied by 24, which is the number of
the tire tread blocks.
These results were still valid both in the case of bogie
and rocker at the front. In addition, during the skid
steering test, it was possible to distinguish two peaks
in the FFT whose frequencies corresponded to the
two different rotational speeds of the wheels on the
two sides.
The computed parameters through the proposed
Real-Time signal processing approach could aid the
motion planning and rover navigation control in a
future perspective.
References:
[1] Liu, Y.; Ma, X.; Shu, L.; Hancke, G.P.; Abu-
Mahfouz, A.M. From Industry 4.0 to Agriculture 4.0:
Current Status, Enabling Technologies, and Research
Challenges. IEEE Trans. Ind. Informatics 2021, 17,
4322–4334, doi:10.1109/TII.2020.3003910.
[2] Navas, E.; Fernández, R.; Sepúlveda, D.; Armada, M.;
Gonzalez-de-Santos, P. Soft Grippers for Automatic
Crop Harvesting, A Review; Sensors 2021, 21, 2689.
[3] Friha, O.; Ferrag, M.A.; Shu, L.; Maglaras, L.; Wang,
X. Internet of Things for the Future of Smart
Agriculture: A Comprehensive Survey of Emerging
Technologies. IEEE/CAA J. Autom. Sin. 2021, 8,
718–752.
[4] Califano F, Cosenza C, Niola V, Savino S. Multibody
Model for the Design of a Rover for Agricultural
Applications: A Preliminary Study. Machines. 2022;
10(4):235.
https://doi.org/10.3390/machines10040235
[5] Bechar, A.; Vigneault, C. Agricultural robots for field
operations: Concepts and components. Biosyst. Eng.
2016, 149, 94–111.
[6] Kan, X.; Thayer, T.C.; Carpin, S.; Karydis, K. Task
Planning on Stochastic Aisle Graphs for Precision
Agriculture. IEEE Robot. Autom. Lett. 2021, 6, 3287–
3294.
[7] Bac, C.W.; Hemming, J.; van Tuijl, B.A.J.; Barth, R.;
Wais, E.; van Henten, E.J. Performance Evaluation of
a Harvesting Robot for Sweet Pepper. J. F. Robot.
2017, 34, 1123–1139, doi:10.1002/rob.21709.
[8] SILES, IVAN, and IAND WALKER, "Continuum
Robotic Elements for Enabling Negotiation of
Uneven Terrain in Unstructured Environments." Issue
WSEAS TRANSACTIONS on APPLIED and THEORETICAL MECHANICS
DOI: 10.37394/232011.2022.17.11
Pierangelo Malfi, Armando Nicolella,
Mario Spirto, Chiara Cosenza,
Vincenzo Niola, Sergio Savino
E-ISSN: 2224-3429
84
Volume 17, 2022
4, Volume 8, October 2013, E-ISSN: 2224-3429,
WSEAS TRANSACTIONS on APPLIED and
THEORETICAL MECHANICS
[9] Toupet, Olivier. "An Overview of the Mars 2020
Perseverance Rover's Enhanced Path-
Planner." (2020).
[10] Jacobstein, Neil. "NASA’s Perseverance: Robot
laboratory on Mars." Science Robotics 6.52 (2021):
eabh3167.
[11] Bickler, D.B. Articulated Suspension System, Patent:
4840394, USA 1989, 1–6.
[12] Bruzzone, Luca, and Giuseppe Quaglia. "Locomotion
systems for ground mobile robots in unstructured
environments." Mechanical sciences 3.2 (2012):
49-62
[13] D. S. Chinchkar, S. S. Gajghate, R. N. Panchal, R. M.
Shetenawar, P. S. Mulik, Design of Rocker Bogie
Mechanism, International Advanced Research Journal
in Science, Engineering and Technology, National
Conference on Design, Manufacturing, Energy &
Thermal Engineering (NCDMETE-2017), Vol. 4,
Special Issue 1, 2017, pp. 46-50.
[14] N. Yadav , B. Bhardwaj, S. Bhardwaj, Design analysis
of Rocker Bogie Suspension System and Access the
possibility to implement in Front Loading Vehicles,
IOSR Journal of Mechanical and Civil Engineering
(IOSR-JMCE), Vol. 12, N. 3, 2015, pp. 64-67.
[15] K. Harish Chandu, P. Hari Narayana, K. C. Charan
Teja, B. Sai, Y. Murali Mohan, Design and
Fabrication of Rocker Bogie Mechanism,
International Journal of Scientific Engineering and
Technology Research, Vol. 7, N. 4, 2018, pp. 781-84.
[16] Cosenza C, Nicolella A, Esposito D, Niola V, Savino
S. Mechanical System Control by RGB-D Device.
Machines. 2021; 9(1):3.
https://doi.org/10.3390/machines9010003
[17] Cumani, Aldo, and Antonio Guiducci. "Fast stereo-
based visual odometry for rover navigation." WSEAS
Trans Circuits Syst 7.7 (2008): 648-657.
[18] C. Chen, M. Shu, Y. Wang, L. Ding, H. Gao, H. Liu,
S. Zhou, Simultaneous control of trajectory tracking
and coordinated allocation of rocker-bogie planetary
rovers, Mechanical Systems and Signal Processing,
Vol. 151, 2021
[19] S. Wu, L. Li, Y. Zhao, M. Li, Slip ratio based traction
coordinating control of wheeled lunar rover with
rocker bogie, Advanced in Control Engineeringand
Information Science, Vol. 15, 2011, pp. 510-15.
[20] Cosenza, Chiara, Vincenzo Niola, and Sergio
Savino. "Modelling friction phenomena in an
underactuated tendon driven finger by means of
vision system device data." The International
Conference of IFToMM ITALY. Springer, Cham,
2018.
[21] Cosenza, Chiara, Vincenzo Niola, and Sergio
Savino. "Underactuated finger behavior correlation
between vision system based experimental tests and
multibody simulations." IFToMM Symposium on
Mechanism Design for Robotics. Springer, Cham,
2018.
[20]
[21]
[22] Niola V., Rossi C., Savino S., Troncone S., An
underactuated mechanical hand: A first prototype,
23rd International Conference on Robotics in Alpe-
Adria-Danube Region, 2014.
[23] M. Thianwiboon, V. Sangveraphunsir, Traction
Control of a Rocker-Bogie Field Mobile Robot,
Thammasat Int. J. Sc. Tech., Vol. 10, No. 4, 2005, pp.
48-59.
[24] A. Nicolella, V. Niola, S. Pagano, S. Savino,
M.Spirito An overview on the kinematic analysis of
the rocker-bogie suspension for six wheeled rovers
approaching an obstacle. Advances in Italian
Mechanism Science in Proceedings of the 4th
International Conference of IFToMM Italy
Mechanisms and Machine Science vol. 122
[25] C. Cosenza, V. Niola, S. Pagano, S. Savino. Spring-
loaded rocker-bogie suspension for six wheeled
rovers. Advances in Italian Mechanism Science in
Proceedings of the 4th International Conference of
IFToMM Italy Mechanisms and Machine Science vol.
122
[26] Amoresano, A., Niola, V., Quaremba, A. (2012). A
sensitive methodology for the EGR optimization: A
perspective study. International Review of Mechani-
cal Engineering, 6(5), 1082-1088.
[27] O.Elshazly et al. “Skid steering mobile robot
modeling and control” UKACC International
Conference on Control. Loughborough, U. – 2014
WSEAS TRANSACTIONS on APPLIED and THEORETICAL MECHANICS
DOI: 10.37394/232011.2022.17.11
Pierangelo Malfi, Armando Nicolella,
Mario Spirto, Chiara Cosenza,
Vincenzo Niola, Sergio Savino
E-ISSN: 2224-3429
85
Volume 17, 2022
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
The author(s) contributed in the present research,
at all stages from the formulation of the problem
to the final findings and solution.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The author(s) declare no potential conflicts of
interest concerning the research, authorship, or
publication of this article.
Creative Commons Attribution
License 4.0 (Attribution 4.0
International, CC BY 4.0)
This article is published under the terms of
the Creative Commons Attribution License
4.0
https://creativecommons.org/licenses/by/4.0/deed.en_US
