압둘아리스우마
(Abdul Aris Umar)
1iD
김정수
(Jung-Su Kim)
†iD
-
(Dept. of Electrical and Information Engineering, Seoul National University of Science
and Technology, Korea)
Copyright © The Korean Institute of Electrical Engineers(KIEE)
Key words
Mecanum-wheeled omnidirectional mobile robot, path-following, nonlinear model predictive control, ROS
1. Introduction
A mecanum-wheeled omnidirectional mobile robot (MWOMR) is one type of omnidirectional
mobile robot which is suitable for application in a narrow environment where a nonholonomic
mobile robot movement might be limited. On the other hand, MWOMR can work with that
environment because mecanum wheels have a different wheel shape compared to common
wheels. They are made up of rollers which are augmented around the wheel with angle
$+-45^{\circ}$. These kinds of wheels allow a mobile robot with mecanum wheels to
remove steering systems.
A mecanum-wheeled omnidirectional mobile robot is expected to be able to follow varying
references with a specified goal. There are two kinds of problems that track varying
references: trajectory tracking and path following. Trajectory tracking and path-following
problems can track varying references using a different approach. In trajectory tracking,
the robot has to follow a time-dependent signal while path-following’s main objective
is to track a geometric curve with the timing being a secondary interest (1). The performance difference between trajectory tracking and path-following is presented
in (2).
For a mecanum-wheeled omnidirectional mobile robot, several efforts have been taken
to control the movement of the robot. In (3)-(4), an approach adapting sliding mode is presented to track a trajectory for MWOMR,
which does not consider any constraints that might arise from robot dynamics and environment
limitations. (5) has taken an effort in formulating model predictive control for a holonomic mobile
robot including MWOMR but focuses only on setpoint stabilization which does not follow
varying references.
Implementation of a control method to a real robot has been getting huge attention
in the literature. The most important problem arising in real hardware implementation
is the robot localization which is crucial in order to define the robot position.
(6) shows the results of combining the camera and encoder data to measure the position
of the robot and apply an explicit model predictive control to drive the robot to
the desired trajectory.
This paper proposes a control framework to minimize the position error of a mecanum-wheeled
omnidirectional mobile robot to its desired path in real hardware. A modified nonlinear
model predictive path-following control is proposed to ensure that the robot moves
in the desired path (7). To implement the NMPC in a real MWOMR system, the combination of lidar, encoder,
and IMU is used to obtain the robot position. The tracking performance of NMPC and
average execution time per iteration are evaluated to see whether it is applicable
to a real robot or not.
The rest of the paper is organized as follows. In section 2, a brief overview of the
dynamics of a mecanum-wheeled omnidirectional mobile robot and path-following problem
definition are described. In the third section is presented the MPC for path-following
of a mecanum-wheeled omnidirectional mobile robot and the implementation in hardware
using ROS. In section 4, the proposed NMPC framework is validated in simulation and
experiment. In section 5, the results of MPC for path-following performance in the
mecanum-wheeled mobile robot are presented. Finally, some conclusions are drawn in
the final section.
Fig. 1. Mecanum-wheeled mobile robot. The figure is reprinted from (8).
2. Problem Definition
In this section, the dynamic model of a mecanum-wheeled omnidirectional mobile robot
and the constraint describing the physical limit of the robot speed are introduced.
Furthermore, the path-following problem definition and goals are presented.
2.1 Kinematics of MWOMR
The dynamics of MWOMR can be described by
where the states $x=\left[p_{x}, p_{y,}, \theta\right]^{T} \in \mathbb{R}^{n}$ consist
of the position of the mobile robot in cartesian coordinate $[p_{x,\:}p_{y}]^{T}$
(m,m) and the orientation angle $\theta$ (rad), and the inputs $u=\left[v_{x}, v_{y},
\omega\right]^{T} \in \mathbb{R}^{m}$ are the linear speed $v_{x}$ (m/s), the lateral
speed $v_{y}$ (m/s), and the angular speed $\omega$ (rad/s) of the robot. Rotation
matrix $R(\theta_{t})$ in eq(1) is denoted as
The configuration of all wheels in MWOMR is essential as it determines the movement
of MWOMR. Fig 1 illustrates the proper orientation of each wheel. In MWOMR, a motor drives each wheel
separately so it can move in any direction freely. The relation of the motor’s angular
speed and the input vector $u$ is given by
where
$H =\begin{pmatrix}1&& -1&& -a-b \\ 1&&1 &&a+b \\ 1&&-1 &&a+b \\ 1&& 1&&-a-b\end{pmatrix}.$
$a$ (m) and $b$ (m) are the half of robot’s wheel base and half of robot’s wheel track,
respectively. $w_{m} \in \mathbb{R}^{4}$ (rad/s) is a vector which describes the angular
speed of each motor matching the setup in Fig 1 and $r$ (m) is the wheels’ radius. The full derivation of the mecanum-wheeled omnidirectional
mobile robot is presented in (8).
From the relation of each motors’ speed and robot’s speed in eq(3), the robot velocity constraint is constructed as introduced in (5),
where
$\|H u\|_{\infty}:=\left|v_{x}\right|+\left|v_{y}\right|+(a+b)|\omega| .$
This paper assumes that the states are bounded by a compact set
2.2 Path-Following Problem
There are particularly three types of control problems in order to steer the system
to the desired point: setpoint stabilization, trajectory tracking, and path-following.
The setpoint stabilization is a problem to drive the system states to a constant reference,
whereas trajectory tracking and path-following steer the states to follow varying
references.
The objective of the trajectory tracking is to follow a certain reference in a specified
time. This may limit the accuracy because it does not incorporate the hardware limit
that is already stated in eq(4) and eq(5). In contrast, the goal of path-following is to reach and follow a geometric path
that does not depend on the time variable. Path-following does not require the system
to be at one point in a specific time.
In the path-following, the references to be tracked is denoted as path $P$ which can
be defined as a set of points $r$ parameterized by a dimensionless $s$
where the $s$ is called a path parameter. The map $p: \mathbb{R} \mapsto \mathbb{R}^{n}$
is assumed to be continuously differentiable. Note that the time evolution of $s(t)\in[s_{0},\:s_{1}]$
is not known a priori, but it is another degree of freedom to choose in this problem.
Here, a virtual input $\phi_{t}$ is chosen to parameterize the evolution of $s_{t}$
as
In the path-following, the controller is designed to achieve these conditions (1):
i) Path convergence
$\lim _{t \rightarrow \infty}\left\|x_{t}-p\left(s_{t}\right)\right\|=0$
ii) Forward motion
$\dot{s}_{t} \geq 0$ and $\lim _{t \rightarrow \infty}\left\|s_{t}-s_{1}\right\|=0$
iii) Constraint satisfaction
$x_{t} \in X, u_{t} \in U \quad \forall t \in\left[t_{0}, \infty\right)$
Condition i) ensures that the system moves along the desired path. Condition ii) makes
sure that the robot is moving forward along the path. The last condition guarantees
that the system follows the hardware limitations.
3. Model Predictive Path-Following Control
In this paper, a discrete-time system of MWOMR dynamics eq(1) and path parameter evolution eq(7) are considered to be used in the model predictive control. A simple Euler’s formula
is used to discretize the system and path parameter evolution as
where $t_{s}$ is the sampling time.
3.1 MPC Formulation for Path-Following
In order to calculate the optimal input for path-following problem, the following
cost function is considered
where $P,\: Q,\:$ and $R$ are positive definite and $(·)_{k+j | k}$ means the predicted
value at time step $k+j$ based on the value measured at time step $k$. The first term
of eq(9) ensures that the system tracks the specified path, and the third term forces the
robot to move forward along the specified path.
Suppose that $bold{u}=\left\{u_{k|k},\: u_{k+1|k},\:\ldots ,\: u_{k+N-1|k}\right\}$
and $bold\Phi =\left\{\phi_{k|k},\:\phi_{k+1|k,\:}\ldots ,\:\phi_{k+N-1|k}\right\}$
denote the sequences of all predicted inputs and virtual inputs, respectively. Assume
that $x_{0}=\left[p_{x},\: p_{y},\:\theta\right]^{T}$ and $s_{0}$ are given, the MPC
control law is obtained by solving the following optimization problem
subject to
The first element of sequence $bold{u}$, $u_{k|k}$, is applied to the robot and $s_{k+1|k}$
becomes the next initial value $s_{0}$ of problem eq(10).
In the case where no initial condition for parameter $s$ is given when $t=0$, $s_{t_{0}}$
can be obtained by solving the following optimization problem:
3.2 MPC Implementation in ROS
Since MPC needs the state of the robot to solve the optimization problem, knowing
the robot position and orientation angle becomes a critical issue. To measure the
state of the robot, this work uses three kinds of sensor to get the robot position
and rotation angle: encoder, IMU, and lidar. The data from each sensor is fused using
EKF localization inside a robot localization package in ROS (9). The results of the localization are the robot position and orientation angle and
can be accessed through ROS localization nodes.
After collecting the required information for the optimization problem, the MPC problem
can be solved directly from the robot. In ROS, the MPC problem is defined using CasADi
as a tool to model nonlinear optimizations (10). CasADi creates an optimization object only once before the robot starts. Then, the
optimal control input can be computed by applying the IPOPT algorithm to solve the
optimization problem eq(10) (11).
The optimal control input, which is the robot speed commands, is passed through the
robot and converted to wheels’ angular speed by eq(3). The MPC implementation in MWOMR using ROS is illustrated in Fig 2.
Fig. 2. MPC implementation using ROS
4. Simulation and Experiment Setup
In this section, the setup for validating the MPC formulation in simulation and hardware
experiment is presented. Two kinds of paths are used to examine the MPC performance
for the path- following problem in MWOMR. Both simulation and hardware experiment
deal with circle path and 8-shaped path as described in eq(12) and eq(13), respectively,
where $p_{x_{e}}$ and $p_{y_{e}}$ are the differences between the reference and current
position and $s\in[0 ,\:1]$.
Since MWOMR can move in any direction without changing the orientation angle, the
performance comparison between the robot moving with fixed and varying orientation
angle is evaluated. Varying the orientation angle is considered in the path formulation
eq(12) and eq(13). In order to reach a path with a fixed orientation angle, $\theta_{path}$ can be
assigned to a constant $c$ (rad) as the desired angle of the robot.
Fig. 3. (a) Mobile robot model in Gazebo, (b) Experimental mobile robot
4.1 Simulation Setup
The simulation is performed in the GAZEBO simulator using Nexus Robot 4WD mecanum
wheel mobile robot model as illustrated in Fig 3a. The simulation is carried out in Ubuntu 18.04 LTS in a computer with AMD Ryzen 3950X.
Table 1 summarizes all parameters and weights used in NMPC formulation for simulation.
Table 1. Robot parameters and weights for simulation
Parameters
|
Value
|
Wheel track
|
0.3 m
|
Wheel base
|
0.3 m
|
Wheel's radius
|
0.05 m
|
Maximum wheel's speed
|
12 rad/s
|
Q weights
|
diag([20,20,10])
|
R weights
|
diag([5,5,5])
|
P weights
|
15
|
MPC prediction horizon
|
15
|
4.2 Experiment Setup
Jetson Nano 4GB is installed on the robot to solve the MPC optimization problem. The
picture of the developed MWOMR is given in Fig 3b. The robot’s parameters and weights are given in Table 2.
Table 2. Robot parameters and weights for experiment
Parameters
|
Value
|
Wheel track
|
0.215 m
|
Wheel base
|
0.165 m
|
Wheel's radius
|
0.039 m
|
Maximum wheel's speed
|
33 rad/s
|
Q weights
|
diag([15,15,10])
|
R weights
|
diag([5,5,5])
|
P weights
|
10
|
MPC prediction horizon
|
15
|
5. Results
In this section, the performance of the proposed method with a circle and 8-shaped
path is presented. First, the proposed method is validated if it can follow a circle
and 8-shaped trajectories in simulation. The results of tracking performance of both
paths are shown in Fig 4 and Fig 5, respectively. From the figures, it is verified that MPC path-following can steer
the robot to move along the specified path while satisfying the robot dynamics and
constraints.
Fig. 4. Tracking performance for a circle path
Fig. 5. Tracking performance for a 8-shaped path
In order to look into the robustness of the proposed design against load variations,
the NMPC based tracking controller is applied to the circle shape reference trajectory
with three different loads: no load, 20 kg load, and 40 kg load. Fig 6 demonstrates that the proposed NMPC results in acceptable tracking performance with
different loads.
Fig. 6. Tracking performance for load variations
Next, the experiment results are presented when the method is applied to the real
hardware using two kinds of paths with fixed and varying orientation angle. The results
of a circle and 8-shaped paths are shown in Fig 7 and Fig 8, respectively. From the figures, the robot could follow the path sufficiently well
with both fixed and varying orientation angle.
In addition to tracking performance, the average time execution per iteration is given
in order to show that the computation time, both in simulation and real hardware,
is less than the robot sampling time. The sampling time is set to 0.1 s which adopts
the sampling time in the real robot lidar. The average execution time taken for completing
one iteration is summarized in Table 3.
Table 3. Average execution time
Case
|
Average time in simulation
(Ryzen 3950X)
|
Average time in real robot
(Jetson Nano)
|
Circle path with fixed orientation
|
3.476 ms
|
31.579 ms
|
Circle path with varying orientation
|
3.010 ms
|
36.184 ms
|
8-shaped path with fixed orientation
|
3.153 ms
|
31.074 ms
|
8-shaped path with varying orientation
|
3.355 ms
|
36.072 ms
|
Fig. 7. (a) Tracking circle path with fixed orientation; (b) Tracking circle path
with varying orientation. Arrows indicate the orientation angle of the robot.
Fig. 8. (a) Tracking 8-shaped path with fixed orientation; (b) Tracking 8-shaped path
with varying orientation. Arrows indicate the orientation angle of the robot.
6. 결 론
In this paper, a control framework for mecanum-wheeled mobile robot using model predictive
control is presented. To this end, a nonlinear model predictive path-following control
is designed for the mobile robot first. Then, real hardware setup is developed in
which the designed MPC is implementable with several sensors. Simulation and experiment
results show that the mobile robot tracks given trajectories successfully while satisfying
all the physical constraints. Moreover, the optimization problem for MPC can be solved
during two consecutive system sampling times in real embedded systems.
Future work along this direction includes considering a speed reference as well in
the middle of path following of the mecanum-wheeled mobile robot.
Acknowledgements
This work was supported by the Industrial Fundamental Technology Development Program
(No. 20014786, Development of AI-based Concrete Slab Finishing Automation System)
funded by the Ministry of Trade, Industry & Energy (MOTIE) of Korea
References
J. Matschek, T. Bäthge, T. Faulwasser, R. Findeisen, 2019, Nonlinear Predictive Control
for Trajectory Tracking and Path Following: An Introduction and Perspective, in Handbook
of Model Predictive Control S. V Raković and W. S. Levine Eds. Cham: Springer International
Publishing, pp. 169-198
A. P. Aguiar, J. P. Hespanha, P. V Kokotovic, 2005, Path- following for nonminimum
phase systems removes performance limitations, IEEE Trans. Automat. Contr., Vol. 50,
No. 2, pp. 234-239
Z. Sun, S. Hu, D. He, W. Zhu, H. Xie, J. Zheng, 2021, Trajectory-tracking control
of Mecanum-wheeled omnidirectional mobile robots using adaptive integral terminal
sliding mode, Comput. Electr. Eng., Vol. 96, pp. 107500
Z. Sun, H. Xie, J. Zheng, Z. Man, D. He, 2021, Path- following control of Mecanum-wheels
omnidirectional mobile robots using nonsingular terminal sliding mode, Mech. Syst.
Signal Process., Vol. 147, pp. 107128
M. W. Mehrez, K. Worthmann, J. P. V Cenerini, M. Osman, W. W. Melek, S. Jeon, 2020,
Model Predictive Control without terminal constraints or costs for holonomic mobile
robots, Rob. Auton. Syst., Vol. 127
Y. Jin, S. Han, S. Lee, May 2019, Implementation of tracking control system for autonomous
mobile robot based on state estimation with multi-sensor, Trans. KIEE, Vol. 68, No.
5, pp. 678-684
T. Faulwasser, R. Findeisen, 2016, Nonlinear Model Predictive Control for Constrained
Output Path Following, IEEE Trans. Automat. Contr., Vol. 61, No. 4, pp. 1026-1039
K. M. Lynch, F. C. Park, 2017, Modern Robotics, Cambridge University Press
T. Moore, D. Stouch, 2016, A Generalized Extended Kalman Filter Implementation for
the Robot Operating System, in Intelligent Autonomous Systems 13, pp. 335-348
J. A. E. Andersson, J. Gillis, G. Horn, J. B. Rawlings, M. Diehl, 2019, CasADi: a
software framework for nonlinear optimization and optimal control, Math. Program.
Comput., Vol. 11, No. 1, pp. 1-36
A. Wächter, L. T. Biegler, 2006, On the implementation of an interior-point filter
line-search algorithm for large-scale nonlinear programming, Math. Program., Vol.
106, No. 1, pp. 25-57
저자소개
압둘 아리스 우마(Abdul Aris Umar)
He received his B.E. degree in Electrical Engineering from Universitas Indonesia,
Depok, Indonesia in 2018.
He is a Ph.D. candidate at the Department of Electrical and Information Engineering,
Seoul National University of Science and Technology (SeoulTech).
His research interests include optimal control, learning-based control, and its application
to robotics.
He received B.S,, M.S. and Ph.D. degrees in electrical engineering from Korea University,
Seoul, South Korea, in 1998, 2000, and 2005, respectively.
He was a postdoctoral researcher at Seoul National University in South Korea, the
University of Stuttgart in Germany and the University of Leicester in the UK from
March 2005 to February 2006, from March 2006 to December 2007, from January 2008 to
January 2009 respectively.
He is currently a full professor in the Dept. of Electrical and Engineering Information
at Seoul National University of Science and Technology, Seoul, South Korea since 2009.
His research interests include model predictive control, distributed control, application
of deep learning to energy systems and robotics.