Since its foundation in 1976, LIG Nex1 has led developments in the Korean high-tech defense sector. LIG Nex1 has contributed to modernizing and boosting Korea’s self-reliant national defense capability through the development of original core technology and advanced weapon systems for virtually all areas of military operation, including C4I, ISR, PGM, EW, and avionics. In 2014, LIG Nex1 started developing the LIG Nex1 AUV as a new business area.
According to a survey conducted by The Statistics Portal, 9,856 people died in marine accidents over the past 14 years. One recent ocean shipping-related accident was the sinking of the MV Sewol on April 16, 2014. The South Korean ferry capsized on its way from Incheon to Jeju City carrying 476 passengers and 24 crew members. Despite heroic efforts by rescue teams, it became a catastrophic event due to strong winds, rain, fog, and low ocean temperature, leading to 290 causalities. Parts of the ocean environment, such as high-speed currents, disturbed the rescue operation. We think that autonomous unmanned vehicles (AUVs) could help rescue people or eliminate threats in oceanic environments. We want to build a prototype AUV for an algorithm test bed as soon as possible with minimal cost.
Unlike traditional unmanned air vehicles (UAVs) or unmanned ground vehicles (UGVs), AUVs need to satisfy additional design parameters specific to the marine environment, which causes issues for developers. First, an AUV requires a high accuracy inertial navigation system to determine motion control because it is not easy to use GPS under water. Second, an AUV requires a watertight body to protect electric devices from salty water. Third, since AUVs maneuver with fluid resistance, they consume more energy than UAVs or UGVs. This means that an AUV has limited operating time and motion coverage. Finally, the cost to test an AUV is higher because it requires transportation to site, launching, and recovery from water. Therefore, developing a real AUV hardware system and testing in site plane is not an easy job to execute.
Choosing CompactRIO and LabVIEW to Develop AUV
AUVs comprise guidance navigation and control (GNC), propulsion, communication, and sensing systems. Among these systems, GNC communicates with all other components and controls the component behavior to accomplish a given mission. Therefore, GNC requires a high processing and multitasking capability and supports various communication methods. By using CompactRIO and the LabVIEW FPGA Module, we could easily implement controllers with algorithms based on The MathWorks, Inc. MATLAB® software, which helped us reduce development time from 18 months to a mere three months.
AUV GNC Prototyping
Our GNC device consists of a CompactRIO controller and insertable interface modules. The GNC controller should interface with more than 11 devices at a maximum 10 ms loop time period. The CompactRIO controller could reconfigure the interface type by changing the interface module (for example serial communication, analog input or output). Therefore, we can easily accommodate interface changes caused by changing hardware components. The GNC controller does not change regardless of interface change. The structure of the interfacing software also needs only a partial update.
GNC Software Based on LabVIEW
We designed the proposed GNC software to operate in a real-time OS (RTOS) with the LabVIEW Real-Time Module. LabVIEW has been applied to system development in various fields with proven stability. GNC should support an RTOS since a control system for altitude and position control requires a highly deterministic loop time.
Functions of LabVIEW are represented as VI files. Figure 2 describes the VI structure of the proposed GNC software. Control.vi and Naviation.vi are the subroutines for the control and navigation algorithms. HCL.vi implements the HCL algorithm. OPlogic.vi and Data Recording.vi manage the operation state and record raw data. SSS and additional equipment are handled in Equipment.vi. All subroutine VIs are controlled in Main.vi. The majority of data is exchanged between tasks by using process shared variable.
To guarantee a deterministic loop time, we needed to use a timed-loop structure and self-callback multithread with a precise loop time for basic loop configuration.
The GNC software should be compatible with MATLAB simulation code to reduce the software migration job and generate simulation and test. When researchers use MATLAB as a simulation language, they need to convert code to C/C++. It seems to be a trivial problem. However, an additional code conversion process might cause unexpected errors in the algorithm since navigation and control algorithms need accurate calculations. The combination of MATLAB and LabVIEW can deliver a fast prototyping method.
Reducing Test Time and Cost Using an NI HIL Test System
AUV testing is costly because the AUV requires transportation to the site, launching, and recovery from the water. Therefore, developing a real AUV hardware system and test in site plane is not easy. We developed an AUV motion test HIL platform using a PXI Express system equipped with an NI RTOS and successfully completed unmanned submarine testing within one week, which typically takes longer than one month.
In general, we apply HIL test when developing a system that requires higher reliability in a short period of time. Figure 6 shows the HIL test configuration with flight motion simulator (FMS), IMU, and LIG Nex1 AUV. We applied the same test scenario to check the performance of the integrated system. The depth changed from 0 to 5 m and heading changed from 0° to 180°.
Since 2013, LIG Nex1 has adopted the NI PXI Express system as FMS real-time controller. We developed HIL test software based on LabVIEW Real-Time and have tested various underwater vehicles for three years successfully. In our AUV project, we could reduce the time to set up the AUV HIL system by combining conventional real-time software and a code synthesis approach based on MathScript. This means that the key model of the M&S SW tool is transplanted into the HIL system. From the viewpoint of field engineers, it is an evolutionary change in the way of algorithm development. By using this solution, we could develop the navigation algorithm within a week—from algorithm design to HIL test.
We successfully completed the water tank test on our first try by conducting thorough verification with HIL test before we actually started. In the past, we had to go through multiple test iterations to correct issues found during water tank tests. These numerous test iterations were the main reasons for extended test time and high test cost. On the contrary, sufficient verification through the HIL system in the AUV project helped us succeed in the first water tank test and dramatically reduced development time.
Developing solutions like AUV is one of the challenges that today’s engineers face to prevent marine accidents from happening and to respond to them more efficiently. We successfully developed the AUV platform for algorithm test and HIL test system for the AUV platform using CompactRIO, LabVIEW, and PXI Express. We plan to further expand this solution to an open architecture-based solution and provide it to various academic and industrial institutions so that they can use it to develop applications related to AUV. We expect our solution to lay a foundation for many organizations in the field to come up with various solutions for conducting maritime tasks.
Moon Hwan Kim
Maritime R&D Lab, LIG Nex1 Co., Ltd