Full duplex hybrid acoustic/RF communication for underwater networked control systems

Official URL: https://risc01.sabanciuniv.edu/record=b3144400_(Table of contents)

Underwater mission planning, monitoring, and coordination of heterogeneous autonomous underwater vehicle (AUV)s require a considerable amount of time and financial resources. This has led to the requirement of establishing reliable communication networks among unmanned underwater vehicle (UUV)s as well as a simulation environment to realistically model the system’s dynamics before actual testing in sea trials. Even though existing solutions can model the dynamics of underwater vehicles, due to complexity, the integration of real-time communication networks has not been considered in the works. To address this issue, this thesis presents an innovative design and realistic co-simulation for a networked control systems (NCS) to achieve navigation of UUVs through communication and control, which is a critical component of real-world marine applications. Traditionally, underwater communication has been based on acoustic communications, characterized by limited data rate and considerably large propagation delay. Taking this issue into consideration, in this thesis, a hybrid acoustic and radio frequency (RF) communication framework is proposed for the underwater NCS where an acoustic link is used for long distance communication and control, and an RF link is employed in the short range. Additionally, to maximize spectrum efficiency, adopting full duplex (FD) communication is proposed for both underwater acoustic and RF links. FD communication enables the feedback signal of the NCS to be iii transmitted rapidly to several AUVs through simultaneous transmission and reception. For the proposed underwater hybrid NCS, a docking scenario is considered, where AUVs perform maneuvers towards a docking station fixed at the seabed. In this scenario, the docking station determines the position of the nearby AUVs, and acoustic or RF communication links carry the position and navigation information from the docking station to AUVs via different medium access control (MAC) protocols. With the help of FD communication, it can be ensured that the underwater hybrid NCS system operates at maximum efficiency, providing the required feedback signal more frequently than NCS with half-duplex communication, resulting in faster and more accurate docking. The AUVs are equipped with two types of controllers for pursuing and actuating docking maneuver: Proportional Integral Derivative (PID) and Linear Quadratic Regulator (LQR) controller, whose gains and sampling times are determined according to the operation of the underwater hybrid NCS. Depending on the communication protocol used in the NCS, protocol delays may be different which forces a change in the sampling times. The different delays of the control loop require further changes in the controller gains to avoid instability. For this purpose, in this thesis, optimization of the controller gains is proposed for the underwater hybrid NCS by applying Sequential Model Algorithm Configuration (SMAC) method for the PID controller and for the LQR controller, by mathematically modeling the hydrodynamics of the AUV to provide better control over disturbances and nonlinearity. By considering the full dynamics of the entire system for controlling the AUVs, the real-time behavior of the underwater networked control system is evaluated realistically using the proposed integrated co-simulation environment, which includes different simulators working together. The performance results indicate that under calm water conditions, our proposed FD underwater hybrid NCS using LQR achieves the shortest docking time of approximately 62 seconds, while the corresponding SMAC optimized approach in FD mode takes around 97 seconds. Furthermore, using FD mode on the acoustic link with the LQR controller reduces the docking time by about 78 seconds. In contrast, for the PID-based method, the docking time is almost doubled to 148 seconds. The underwater hybrid NCS is also evaluated under realistic fluctuating water currents, using two controllers, different MAC protocols, and FD and HD communication modes. Our experiments indicate that with LQR, the proposed FD underwater hybrid NCS’s docking time, when exposed to such currents, is 90 seconds, while the SMAC optimized PID takes approximately 175 seconds. In contrast, the conventional acoustic-based HD mode using LQR for realistic currents has a docking time of around 120 seconds, while the SMAC optimized PID takes about 245 seconds. The penalty to achieve improved performance using FD hybrid is spending 70% more motive energy than the acoustic only system. It is worth noting that communication modes using SMAC optimized PID cannot complete docking maneuvers if the current speed exceeds 0.3m/s, while LQR based methods can handle current speeds up to 0.7m/s. At this velocity, conventional acoustic-based systems take about 140% longer to complete docking than our proposed FD hybrid system. These results demonstrate the feasibility and advantages of using the proposed FD hybrid communication approach for AUV control.