To ensure that ROVs can complete depth tracking in the presence of external disturbances and system uncertainties, this paper proposes a double-loop sliding mode controller with a disturbance observer. The double loop strategy which concludes an inner-loop control system and an outer-loop control system is applied to design the controller. A disturbance observer is also used to estimate and compensate the whole disturbances and uncertainties. Simulation results illustrate that the proposed controller can improve the performance of sliding mode controller. Robustness and accuracy of the controller are maintained at a high level while the thruster chattering is reduced by 34.1%-57.8%.


According to their operation methods, underwater robots can be classified into autonomous underwater vehicles (AUVs) and remotely operated vehicles (ROVs). AUVs are autonomous, however, it cannot communicate with the land in real time. So it is hardly to be used for complex underwater operations. Work-class ROVs are generally equipped with manipulators, which can perform underwater heavy-duty operations. Moreover, ROV can achieve real-time communication with mother ship through the umbilical cables. Therefore, ROVs have been widely used in underwater engineering, such as deep-sea exploration, pipeline maintenance, and deep-sea mining efforts. ROVs are manually operated that two operators are required to cooperatively control one ROV. One person is responsible for controlling the body movement, and the other person controls the manipulators. The operator's experience determines the control accuracy and efficiency, which also results in high operating costs. Therefore, improving the automatic control capability of ROVs is currently a hot issue(Schjølberg & Utne, 2015).

ROVs are often disturbed by the undercurrent due to the complexity of the submarine environment. Moreover, ROV is always operated at a large depth and its umbilical cable will significantly interfere with the body movement. At the same time, ROV dynamics are highly coupled and nonlinear. ROVs' irregular shape makes it difficult to accurately describe its hydrodynamic characteristics and create model uncertainties. The above problems all lead to considerable difficulties in ROV automatic control. To solve these problems, many advanced control methods have been applied to ROV control, such as adaptive control, sliding mode control (SMC), fuzzy logic control, model predictive control, and neural network control (Huo, Ge, & Wang, 2018; Maalouf, Chemori, & Creuze, 2015; Shen, Shi, & Buckham, 2018; Soylu, Buckham, & Podhorodeski, 2008; Wang, Zhang, Wilson, & Liu, 2015). According to ROVs' special character, sliding mode control has been proven to be a very attractive approach to cope with the problem due to its simple form, strong robustness, tolerance of model uncertainty and transient stability (Soylu et al., 2008). However, sliding mode control has the problem of chattering phenomena. The sliding surface function contains a discontinuous switching term, which is generally served by a signum function. This character is the source of the robustness of the SMC law, but it also acts on the system in a bang-bang manner thus creating chatter in the actuators (Slotine & Li, 2004). Many advanced methods have been proposed to solve the chattering problem of sliding mode control. As mentioned above, the discontinuous switching term is the main reason for chattering, thus Slotine (2004) and Kim (2015; proposed replacing the discontinuous signum function by a continuous saturation function or hyperbolic tangent function. Soylu(2008) proposed a chattering-free method for ROV systems that uses an adaptive term in place of the conventional, discontinuous switching term. Fuzzy sliding mode logic controllers are designed to assure performance and stability (Chen, 1999; Guo, Chiu, & Huang, 2003). Meanwhile, the fuzzy partition of the switch term minimized the chattering.

This content is only available via PDF.
You can access this article if you purchase or spend a download.