UDT 2020 Missionized Riptide Unmanned Underwater Vehicles J. E. - - PDF document

UDT 2020 UDT Extended Abstract Unmanned, Remotely Piloted & Autonomous Systems

Not export controlled per ES-FL-021720-0033 Approved for public release; unlimited distribution.

UDT 2020 – Missionized Riptide Unmanned Underwater Vehicles

• J. E. Filiberti1 and R. M. Carvalho Jr.2

1 Senior Principal Scientist, BAE Systems FAST Labs, Arlington, Virginia, USA, julia.filiberti@baesystems.com 2 Technical Director, BAE Systems FAST Labs, Merrimack, New Hampshire, USA, ronald.carvalho@baesystems.com

Abstract — Following the acquisition of Riptide Autonomous Solutions, BAE Systems FAST Labs’ autonomy and unmanned underwater vehicle engineers teamed to missionize a 9-inch diameter Riptide unmanned underwater vehicle (UUV) for The Advanced Naval Technology Exercise (ANTX) demonstration in August 2019, hosted by the Naval Undersea Warfare Center (NUWC) Newport Division. The work presented includes an overview of the design, integration, and demonstration of several modular autonomous sensing and communications payloads. The missionized Riptide software architecture promotes flexibility to plug and play future modular autonomous payloads developed across government, academia, and industry alike. Once deployed, the vehicle is designed to execute a military mission without any human intervention, including graceful recovery operations. Optional user commands are able to redirect the autonomous Riptide vehicle during mission execution; additionally, users can track mission progress through a real-time mission display. The ANTX-19 demonstration proved out the concept that a single tactical UUV could perform an information collection task without any human intervention, leveraging multiple sensor modalities and high-level mission autonomy.

1 Introduction

1.1 Motivation Militaries could strongly benefit from tactical UUVs for the performance of a myriad of naval tasks, acting as force

• multipliers. A small class (<10” diameter) tactical UUV

working singularly, or in a group can characterize environments, survey vessels of interest, collect other information from the waterspace, and communicate that information to its manned counterparts over the horizon. A Design Reference Mission (DRM) for a single tactical UUV was selected as being most representative of the typical tasks faced by a UUV in an information collection

• situation. In this case, UUVs and their payloads serve as

long dwell information collection sources for vessel detection and identification. The need to conduct wide area search, collect information of interest, and communicate over long distances with limited bandwidth place high demands on mission system autonomy. 1.2 Objective The purpose

• f

the Missionized Riptide UUV development effort was to develop a UUV mission suite capable of performing an information collection task and demonstrating it at the Advanced Naval Technology Exercise 2019 (ANTX-19). More specifically, we sought to develop a Small-class Riptide UUV with multi-modal sensing, communications, and autonomy functionality to successfully perform the information collection task with no human intervention. The objective sensor payloads included radio frequency (RF) and imagery sensor

• modalities. The objective autonomy functionality

included payload control to achieve smart energy management and mission control to extend system survivability and govern task progression. A secondary objective of the Missionized Riptide UUV project was to lay the foundation for future mission extensions and interoperability with other payloads. This included the development of a software interface which promoted application-only interaction with a common message bus, baseline model-based systems engineering (MBSE) implementations, rapid prototyping and integration practices, and test infrastructure investment and exercise.

2 Approach

2.1 Wampus Overview The resulting Small UUV System, dubbed Wampus, was

• utfitted with a BAE Systems RF Sensor System (ported

directly from unmanned aerial vehicles), a retractable mast, off-the-shelf panoramic camera, acoustic & radio communications, and a processor outfitted with mission controller software. The system was designed in a MBSE environment, and was designed, integrated and tested within a five- month period. Wampus successfully demonstrated the ability to geolocate a signal of interest, proceed to the vessel of interest, collect close-area imagery of the vessel, and rendezvous with a recovery platform, all while operating fully autonomously. 2.2 Wampus Hardware Development

UDT 2020 UDT Extended Abstract Unmanned, Remotely Piloted & Autonomous Systems

Not export controlled per ES-FL-021720-0033 Approved for public release; unlimited distribution.

2.2.1 Small-class UUV Development BAE Systems’ Small-class vehicle platform (9.375 inch diameter) represented the newest addition to the Riptide family of vehicles when it was redesigned in June

• 2019. The vehicle preserves the same modular, open

architecture and open source philosophy of its Riptide predecessors; the nose contains navigation sensors and the main processor; the tail has fin actuation, propulsion, safety systems and communications; and the mid-body houses the battery system and power distribution. The base vehicle is a dry volume, to which a wet section may be added to house subsystems that require direct contact with the water, as was done for the ANTX-19 payloads. From a mechanical perspective, the most notable update to the Small-class is the joining ring architecture, which supports rapid prototyping activities and their associated sustainment actions (e.g., battery replacement) by removing much of the effort associated with vehicle

• deconstruction. Another significant update was to the

battery system and mid-body housing for the battery assembly, which allow for repeated, precise placement of and improved cable management. Given that the battery represents a significant portion of the vehicle’s ballast and trim weight, exact repeatable placement is

• important. Further improvements have been made to the

fin shaft seals in the tail for smooth operation and robustness. 2.2.2 Mast Development The Wampus mast supported the camera, RF, GPS, and communications antennas. The prototype was based on a “Roll a Tube” product that is currently used by ground

• forces. The advantages of this approach were a rigid,

hydrodynamic mast structure that could be stored in a relatively low volume, and allowed for approximately 5lbs

• f

weight support. A mast structure was required to provide the transition space between the tube in its stored form and its deployment in tubular form. Mounted to this mast structure were the camera and a boom that provided horizontal separation of GPS, RF communications and sensor antennas. 2.2.3 Mission Computer Processing Hardware The mission computer integrated into Wampus was a prototype of BAE System’s SQUID INC (Secure Qualified Undersea Integrated Device for Identify Friend

• r Foe (IFF), Navigation & Communications) mission
• processor. SQUID INC will provide cross-domain (air,

sea, subsea), IFF, navigation, communication and mission- level autonomy capabilities in a commercial open systems

• architecture. Capabilities within the system are envisioned

as being Application (App)-based, with app accesses dependent upon the user’s security level, optimized for interoperability with other software apps and hardware instantiations, and tailored to task requirements. Specific capabilities incorporated into SQUID INC and demonstrated in ANTX-19 include those described in Section 2.4.3. 2.3 Sensors & Communications Payloads Wampus employed and autonomously controlled four sensor and communications payloads during ANTX-19. Communications payloads included a commercial off-the- shelf radio and acoustic modem. The two sensor payloads are described next. 2.3.1 RF Sensor Payload Although the RF sensor payload was initially designed for unmanned aerial vehicle use, its functionality and form adhered to the requirements for the UUV payload. Therefore, the core payload hardware was ported without modification from the UAV implementation. The antenna apertures were modified to provide a waterproof environmental seal and different frequency coverage. The SW baseline was modified to provide detection and identification of multiple maritime waveforms. The UUV and support vessel both hosted an instance of the RF software, enabling cooperative geolocation

• perating over a radio network.

The system occupied just over 25 cubic inches of payload space, weighed less than 1.7 lbs, and its power consumption was 18.5 Watts under its heaviest processor load. 2.3.2 Camera Payload The camera used was a lightweight 360-degree spherical commercial camera. The camera was chosen based on its compact design and high image quality for future programs and because it was an inexpensive option amenable to a proof-of-concept demonstration.

UDT 2020 UDT Extended Abstract Unmanned, Remotely Piloted & Autonomous Systems

Not export controlled per ES-FL-021720-0033 Approved for public release; unlimited distribution.

The RF sensor payload, camera, radio, and acoustic payloads were located in the vehicle consistent with figure X/add in captions and figure references. 2.4 Autonomy 2.4.1 Software Architecture The maritime autonomy and mission controller software architecture was developed by BAE Systems in the spirit

• f the US Navy’s future Unmanned Maritime Autonomy

Architecture (UMAA), as currently defined. Each payload, including sensors, communications, and mast subsystems, interfaced with the mission controller via a common public-subscribe message bus. This software architecture enabled energy efficiencies through autonomous payload control, thereby extending the endurance of the overarching system. Moreover, it enabled autonomous health and system monitoring through published subsystem statuses used for determining whether the system had the capacity to continue the

• mission. Because the subsystems communicated over a

common message bus and employed the same maritime-

• riented message schema definitions, software integration

timelines were reduced from weeks and months to days during the sprint leading up to the ANTX-19 exercise. 2.4.2 Lower-level Autonomy The software system on the base Small-class vehicle is identical to that used across all Riptide UUVs. It is an

• pen source platform that allows customers and end-users

to modify or extend the system, as they require for their

• application. The software system may be conceptualized

in four levels. At the lowest level, Hardware Control provides command/control of the base vehicle hardware including sensors, actuators, safety and power systems. In addition to interfacing to this hardware, the software also abstracts hardware-specific protocols to a common Riptide

• protocol. At the Vehicle Control level, data from the

sensors are fused into a navigation solution providing vehicle position and attitude, which is in turn used to command fins and propulsion motors to achieve stable vehicle flight. Above this is the Behavior Control level, where the behaviors that direct the vehicle to achieve specific maneuvers (e.g., depth, speed, heading, waypoints, etc.) are generated. In the current vehicle software system, this Behavior Level control is provided by the open source MOOS-IvP system, though it can also be provided by a ROS bridging application, or other system such as MOAA/check all acronyms. For ANTX- 19, a fourth level of control, Mission Autonomy, was provided by the SQUID autonomy components interfacing with MOOS-IvP via an Message Data Bus-to-MOOS bridging application. 2.4.3 Higher-level Mission Autonomy Functionally, the entire information collection task was executed without human intervention using the Wampus mission controller software through adaptation to the perceived operating environment in real time. The mission controller was able to perform each step in the information collection task autonomously, including actioning fused data reports, making decisions for when to proceed to the next stage of the task, and determining whether the system was suitable to continue the task under current conditions.

UDT 2020 UDT Extended Abstract Unmanned, Remotely Piloted & Autonomous Systems

Not export controlled per ES-FL-021720-0033 Approved for public release; unlimited distribution.

The Wampus mission controller backbone governed autonomous task progression, management of payload activities, IFF functions, managed energy usage, and monitored health & status, with optional interaction with human operators. To support optional user interaction, autonomy applications interfaced with a beta display for interactive mission monitoring, and leveraged human- machine acoustic and radio communications for IFF and mission tracking functionality. The mission autonomy module is easily ported to low SWaP hardware, is easily extensible due to disciplined

• bject-oriented implementation, and is easily integrated

with other application services via the software architecture approach (Ref. 2.4.1/check all refs). The mission controller software was containerized using Docker for increased portability. 2.5 Model-Based Systems Engineering Approach Design of the baseline system architecture was conducted via a Model Based Systems Engineering (MBSE) approach using a MagicDraw software in a SYSML

• framework. This approach was chosen for its efficiencies

in design and documentation in a shared workspace

• environment. The Design Reference Mission’s (Use Case)

Key Performance Parameters (KPPs) were recorded as upper level requirements and the Use Case was then decomposed into a series of State Diagrams and further into State Activity Diagrams allowing for development of lower level subsystem requirements. A Requirements Verification Matrix was rapidly constructed from the requirements matrix, which formed the basis of the test

• plan. Implementation of the autonomy software modules

was founded on the State Flow and Activity Diagrams developed in the MBSE environment; the MBSE design approach was carried through to both software and hardware implementations as designed. SYSML was used to document the model down to physical connectors, signal flow and power diagrams. 2.6 Integration and Test Approach Bench testing and integration were conducted prior to the construction of the Small-class UUV with communications, sensing, and mission control payloads. Because of the application-only interaction with the message bus, software integration timelines were greatly accelerated and successful. Software was developed in more than four disparate locations across BAE Systems, so efficient bench integration events enabled by the software architecture (Ref 2.4.1) were even more beneficial. After bench integration and the vehicle construction was complete, Wampus was tested in a parking lot. Parking lot tests exercised the autonomy, communications, and sensor payloads, without the added complications of in- water deployment. Prototype launch practices were seamless and enabled by trailer transportation, a support boat, and a two-personnel launch procedure. The vehicle was launched by trailering it into the water, and towing it behind a support vessel out to the test area as a surrogate operational deployment

• initialization. After arriving, personnel removed the tow

rope and released the UUV into the water.

3 ANTX-19 Demonstration Results

During the ANTX-19 demonstration, Wampus was launched into the test area. Once deployed, users on- board the support vessel sent a command message to Wampus from the mission laptop. The message contained all of the external parameters necessary to begin the task. After mission receipt, Wampus began the information collection task.

UDT 2020 UDT Extended Abstract Unmanned, Remotely Piloted & Autonomous Systems

Not export controlled per ES-FL-021720-0033 Approved for public release; unlimited distribution.

As planned, Wampus successfully transited to the patrol

• area. MOOS-IvP waypoint behaviors were executed and

the vehicle’s built-in navigation and control modules exercised appropriate mechanisms to proceed to the desired waypoint. Transit in a partially submerged state was performed with only the mast and antennas exposed. Radio communications were reliable and supported user mission monitoring throughout the demonstration. After arrival at the patrol site, the mission controller commanded the RF sensor payload to perform wide area

• search. Here, the vehicle executed a racetrack patrol

pattern (a MOOS-IvP loiter behavior) until the vessel of interest was detected. Wampus proceeded to the vessel

• f interest location. As Wampus got closer to the vessel,

the RF sensor suite was able to detect the signals of interest (see Lessons Learned Section 4.1). After successful waypoint transit to the vessel of interest, Wampus autonomously captured images using the 360 degree camera payload while circling the vessel of interest. Once the vessel was imaged, the vehicle proceeded to the recovery point, as intended. As Wampus approached the recovery area, it successfully conducted “own asset identification” procedures with the support vessel over acoustic communications, simulating verification that the UUV had not been compromised and that the ingressing UUV was indeed a friendly agent. After IFF verification, the vehicle traveled the remaining distance to the recovery point. Test support personnel attached Wampus to a tow vehicle and returned to the test facility.

4 Lessons Learned

4.1 Hardware A number of lessons learned were captured during our rapid integration and testing initiatives. Overall Wampus was successful in proving out multi-modal sensing on an autonomous Small-class UUV platform, using prototype mast and antenna designs. We are developing the next iteration of mast and antenna designs, leveraging high- fidelity modelling and simulation tools. 4.2 Software Additional lessons learned were captured from our prototype autonomy software implementation, its supporting interface, and autonomous payload control

• approaches. Software containerization posed a number of

limitations during in-water testing events, mainly due to its locks on resources and need to connect to the internet for software rebuilds. As a result, future efforts will allocate more resources to containerization nearer toward production, after initial prototype testing. Additionally, though not an issue during ANTX-19 demonstration, future live tests will have exact duplicate mission processors on-hand should there be any processing hardware issues.

5 Future Work

5.1 Hardware Near term hardware improvements include a redesign of the antenna system and the mast deployment structure. Both are required to increase sensor range performance and decrease burdens to vehicle dynamics and energy

• efficiency. As a result, the vehicle should be able to dive

to 300 meters and RF sensor suite should experience better

UDT 2020 UDT Extended Abstract Unmanned, Remotely Piloted & Autonomous Systems

Not export controlled per ES-FL-021720-0033 Approved for public release; unlimited distribution.

range performance. Moreover, we plan to increase vehicle speed capabilities through propulsion system

• improvements. And lastly, we plan to make better use of

vehicle inserts to promote easier assembly and disassembly practices, thus expediting testing procedures. 5.2 Software Future software work already underway includes integration with the BAE Systems’ mature tracking and data fusion algorithms, machine learning-based classification algorithms, and additional sensing and communications payloads. A ROS/Gazebo modelling and simulation capability is also in progress. Lastly, we plan to implement multi-task and multi-agent mission autonomy for extensibility to more military applications.

6 Conclusions

The ANTX-19 results support the proof-of-concept that a low-cost single tactical diameter UUV can perform an information collection task, acting as a force multiplier for military forces in impermissible maritime situations. Moreover, the design, implementation, integration and demonstration of such a capability for the first time in less than just five months proves that platforms can be missionized for highly specialized tasks under very short

• timelines. The software architecture design and

implementation is effective for rapid integration and testing with components developed over various locations. The multi-int sensor configuration showed utility for in- water applications. The communications approach proved sufficient for in-water applications. The test procedures were amenable to frequent development cycles. And the mission autonomy approach proved sufficient for complex tasks and real-time situation understanding. In summary, the synthesis of high-level mission autonomy, multiple communications and sensing payloads, and a prototype platform can amass significant capabilities to perform a range of complex tasks, concluding that deployment of tactical UUVs can be worthwhile to military operators.

7 Acknowledgements

The authors express their sincere gratitude to the other hardworking team members who made missionized Riptide UUVs a reality, including, but not limited to: Dani Goldberg, Langdon Tarbell, Dean Schifilliti, James Plummer, Kayla Magee, Peter Schibly, Lily Liu, Jason Weiss, Giovanni Gensale, Christopher Van Valkenburgh, Jay Lustig, Jeffrey Smith, and Geoffrey Edelson.

8 References

FINAL REPORT UNDERSEA WARFARE CONFERENCE MOOS ROS/GAZEBO References should be cited in the text by placing sequential numbers in brackets (for example, [1], [2, 5, 7], [8-10]). They should be numbered in the order in which they are

• cited. A complete reference should provide enough

information to locate the article. References to printed journal articles should typically contain:

• The authors, in the form: initials (only the first letter

capitalized with full stops after the initials) followed by family name;

• The journal title (abbreviated).
• The volume number in bold type;
• The article number or the page numbers,
• The year of publication (in brackets);

Authors should use the forms shown in Table 3 in the final reference list. Here are some examples: [1] A. Mecke, I. Lee, J.R. Baker jr., M.M. Banaszak Holl, B.G. Orr, Eur. Phys. J. E 14, 7 (2004) [2] M. Ben Rabha, M.F. Boujmil, M. Saadoun, B. Bessaïs, Eur. Phys. J. Appl. Phys. (to be published) [3] F. De Lillo, F. Cecconi, G. Lacorata, A. Vulpiani, EPL, 84 (2008) [4] L. T. De Luca, Propulsion physics (EDP Sciences, Les Ulis, 2009)

Table 3. Font styles for a reference to a journal article. Element Style Authors Normal Initials followed by family name Journal title Normal Abbreviated Book title, Proceedings title Italic Volume number Bold Page number Normal Year Normal In brackets

9 Author/Speaker Biographies

• Ms. Julia E. Filiberti is a Senior

Principal Scientist at BAE Systems FAST Labs and served as autonomy lead for the Missionized Riptide effort

• presented. She is currently a

principal investigator for DARPA and ONR programs, and supports a range of maritime autonomy initiatives.

UDT 2020 UDT Extended Abstract Unmanned, Remotely Piloted & Autonomous Systems

Not export controlled per ES-FL-021720-0033 Approved for public release; unlimited distribution.

• Mr. Ronald M. Carvalho Jr. is

a Technical Director at BAE Systems FAST Labs and served as Principal Investigator for the Missionized Riptide effort. A retired naval officer with a background in Model-Based Systems Engineering, he supports a large portfolio of Navy capture and execution efforts across FAST Labs maritime systems programs.