## 1. Introduction

In the offshore industry, dynamic positioning systems (DPSs) are widely applied; this can be seen in pipe laying, offshore wind farms, and drilling rigs, for example. As a result of the restrictions on the use of anchor in deep water, a vessel has equipped by rudders, propellers, and thrusters; these are used to automatically keep its position and heading stable and safe from environmental disturbances such as waves winds and sea currents [

1]. Historically speaking, the DPS was first used in the 1960s with the aim of controlling the motion of vessels in three horizontal degrees of freedom, such as sway, surge, and yaw. The single-input-single-output (SISO) and the proportional-integral-derivative (PID) with a low pass filter control systems were used. Unfortunately, the PID control system caused a phase change that affected the stability of the system [

2]. Balchen and his colleagues used the Kalman filtering techniques, stochastic optimal control theory, and more advanced control algorithms to improve the DPS application on the vessels. This was based on the assumption of kinematic equations [

3,

4,

5,

6,

7]. However, the results showed that the equations of motion were linear in a series of predetermined fixed yaw angles, and the stability of the DPS could not be guaranteed. The study in [

8] introduced the nonlinear control rules of universal identical asymptotical permanency based on the backstepping method where environmental disturbances are not considered. However, the ecological disruption caused by sea conditions is dynamic and could not be unnoticed. The passive nonlinear observer (PNO) that consists of the bias formal approximation of low-frequency position was proposed to measure the velocity of the vessel during the movement, and wave filtering to decrease the quantity of adjusting parameters [

9]. The performance of their methodology used to design the proportional-derivative (PD) control rule, for the output feedback evaluations of the dynamic position system, was investigated in [

10]. In this experiment with a vessel model, the observer filters eliminate the noises from the measurements of Ship position velocity by designing a PD controller that is gradually changing due to environmental disturbances.

New control techniques have been Constituting an advanced model of intelligent behavior, and computational methods have been developed to support them. An adaptive nonlinear PID controller is considered to decrease the deviation of vessels from the wanted position while environmental disturbances cause unexpected sudden changes of positions [

11,

12]. The research presented in [

13] proposed an adaptive observer for dynamic positioning on the output of the feedback controller to approximate the remotely operated underwater vehicle (ROV) speed and uncertainties of parameters. Moreover, it has proposed a linear Kalman Filter (LKF), an Extended Kalman Filter (EKF), an adaptive Kalman Filter, and a passive nonlinear observer-based mathematical model on the ROV Minerva. Besides, an adaptive controller has been developed to estimate the nonlinear DPS parameters by adaptive fuzzy logic theory in [

14,

15,

16,

17,

18,

19]. The study in [

20,

21,

22,

23] is presented a neuro-fuzzy algorithm, which includes a fuzzy control-based neural network algorithm (NNA) so that the basis for the fuzzy rules and membership function can be created during the network learning process. By applying the NNA and it’s setting, the self-regulation of the membership functions is desirable. In addition, in fuzzy and neural network algorithms, the mathematical model does not consider the deriving controller due to timesaving consideration.

In [

24], the DP technology, DP vessel mathematic model, DP controllers and supervisors, thrusters allocation, as well as the hybrid control of DP vessels experiment results were reviewed. To emphasize the requirement of further investigation in this filed, the research presented in [

25,

26,

27,

28,

29] was focused on hybrid control techniques and operation in DP vessels from calm to extreme sea conditions. It analyzed a hybrid control method for switching between linear or nonlinear controls in maximum operating conditions. In this control method, the nonlinear controller uses an independent scale for switching control to ensure overall system stability and prevent snoozing.

Intending to further enhance control performance, researchers in [

30,

31,

32,

33,

34,

35] proposed the model predictive control (MPC). These types of literature are compared to the advantages and disadvantages of the MPC for typical non-linear control applications in DP control problems with other conventional algorithms. Especially, in advanced vessels such as cable laying and shuttle tankers, more sophisticated energy management systems are applied to predict the power demand, route scheduling, and thrust allocation to keep the position of the vessel by using the DP control technique. On the other hand, the power management system (PMS) which was applied in [

36,

37,

38,

39,

40,

41,

42,

43,

44,

45,

46,

47,

48] for controlling of power generators, blackout prevention, power limitation, load sharing, and load shedding. Without a doubt, the DP control is issued according to the information about available power from the PMS to set the desired pitch/rpm load demand to allocate power for thrusters systems.

This paper presents a review of the advantages and disadvantages of the DP control strategies, which have occurred over three decades of investigation and improvement on marine vessels. This literature is divided into five sections.

Section 2 presents the DP system models and components in marine vessels. Moreover, throws the light on the DP control classifications are and provides a summary of the vessel models.

Section 3 discusses different control strategies, including Kalman filter, model predictive control, fuzzy logic control, neural network, and an adaptive sliding mode with finite-time observer-based control in DP vessels. Finally, the challenges facing the traditional DP controls and further study are discussed as a conclusion in

Section 4.