Ding et al.
Ocean Engineering 235 (2021) 109355
9
C. Cheng et al.
Fig. 8.
An illustration of force analysis for the artificial potential field algorithm.
Source:
Modified from
Cao et al.
(
2006
).
situations that AUVs need to confront in the marine environment, AUV
can move along the boundary of the obstacle and start the dynamic
recognition range to prevent forgetting obstacles.
Saravanakumar and
Asokan
(
2013
) put forward a multi-point potential field method, which
discretizes the surface of the obstacle facing the AUV head into multiple
points. By analyzing the force acting on multiple points and obstacles,
the underwater vehicle is guided to the point where the total potential
energy is minimum. The proposed method is tested in a 3D simulation
environment and experimental results show it can effectively avoid
dynamic and static obstacles. In addition, they emphasize that the
proposed multi-point potential field method is more suitable for the
6-DOF AUVs working in 3D space and is not meant for autonomous
ground vehicles (AGVs) operating in 3D terrain.
Cheng et al.
(
2015
)
proposed an artificial potential field method in consideration of both
the current and dynamic obstacles. This method combines the current
and the speed of the AUV to resist the adverse effect of the current
on the AUV navigation. Then the direction of AUV’s velocity can be
accurately determined with the relationship between repulsion force
and gravity. Their results show it can overcome influence of the cur-
rent and avoid collision with dynamic obstacles.
Villar et al.
(
2016
)
proposed an artificial potential field method for obstacle avoidance of
AUV with a mechanical scanning sonar. Specifically, different attraction
and repulsion constants are used in the potential function to create
an optimal path, and virtual obstacles are also introduced to solve
the local minimum problem. In a target search task,
Ge et al.
(
2018
)
proposed an improved potential field method for multi-AUV obstacle
avoidance. In their method, the dispersion degree, the homodromous
degree and district-difference degree are considered in the potential
field function, which can allow AUVs to complete the search task
accurately and avoid falling into local traps.
Noguchi and Maki
(
2019
)
proposed a path planning method based on artificial potential field
algorithm and reinforcement learning (
Rummery and Niranjan
,
1994
).
Because the environmental information obtained by sensors is limited
and fuzzy, binary Bayes filter (
Thrun
,
2002
) is used to estimate the oc-
cupancy probability of obstacles in the workspace of AUV. Experiments
show that this method can generate a safe path without contact with
obstacles.
Khalaji and Tourajizadeh
(
2020
) extracted the kinematics
and dynamics model of AUV in Section
2
. Then, based on Lyapunov
theory (
Lefeber et al.
,
2003
), a new nonlinear controller is proposed,
which uses the potential field algorithm to avoid local stationary or
moving obstacles. In addition, an adjustable avoidance gain is added to
the potential function, and the simulation results show that AUV can
bypass the obstacle with the minimum deviation from the reference
trajectory.
Fan et al.
(
2020
) proposed an improved artificial potential
field method for real-time path planning of AUV. In their method, a
distance correction factor is added to the exclusion function to solve the
local minimum problem, and the relative velocity method is introduced
to detect and avoid dynamic obstacles. Considering the potential obsta-
cles and forbidden areas in the monitoring system of the underwater
Fig. 9.
Illustration of the basic principle of fuzzy logic algorithm for obstacle avoidance
of AUV.
Source:
Modified from
Cordón et al.
(
1996
).
internet of things,
Lin et al.
(
2020
) proposed a joint control model for
multiple AUVs based on the artificial potential field algorithm. In their
method, the Software-Defined Networking (
Lin et al.
,
2018
) paradigm
is used to improve the flexibility and controllability of group based
AUV. In addition, each AUV in the system not only pays attention to
the influence of the surrounding obstacles, but also takes dynamics of
all AUVs into account. The simulated experimental results show that
the proposed method can effectively manage the path planning for all
AUVs.
The artificial potential field method has great advantages in avoid-
ing unknown obstacles. However, the local minimum problem should
be considered to avoid the instability of AUV caused by limited force,
which might bring more loss to the equipment or environment.
Do'stlaringiz bilan baham: 
