Understanding Army s UAS Requirements Through Modelling And Simulation

Transcription

1 Understanding Army s UAS Requirements Through Modelling And Simulation Shane Rogers, Senior Manager; Richard Aplin, Operations Analyst Systems Analysis Laboratory Boeing Defence Australia Ltd 363 Adelaide St, Brisbane Queensland 4001 Shane.rogers@boeing.com Richard.j.aplin@boeing.com Abstract. Unmanned Aerial Systems (UAS) to support tactical manoeuvre are prime candidates for experimentation in order to gain an improved understanding of Army s current and future requirements. Modelling and simulation enables a range of platform types, quantities and payloads to be introduced into realistic tactical scenarios alongside existing and proposed Army capabilities to assist in determining the effectiveness of multi-tiered UAS options. This paper outlines the conduct of the first of a proposed series of experiments to be undertaken by Boeing s Systems Analysis Laboratory (SAL) exploring alternative UAS combinations, with varied payloads, to enhance force effectiveness and tactical decision making. It highlights the creation of an experiment scenario, required software development and the overarching simulation architecture to achieve a realistic virtual, network-centric operational environment for discovery experimentation. In addition, the proposed measures for analysing changes in force effectiveness observed for each UAS option will be discussed, with subsequent descriptions of the visualisations required to appropriately communicate these findings to the customer. The use of the terms Army and Australian Defence Force (ADF) do not imply nor indicate Defence endorsement 1. INTRODUCTION The first of a planned series of experiments, known as NIGHTJAR, was run by the SAL in February The initial experiment was a short UAS discovery experiment designed to explore the employment options of current and future Tactical UAS (TUAS) whilst showcasing the SAL s Virtual Land Experimentation capability to Boeing s business units. The success of the experiment in both the UAS and Network Centric Operations (NCO) arenas has lead to great interest from Boeing s International Modelling and Simulation community and also elements within the ADF to conduct follow-up activities to support programs ranging from Tactical Networking Systems for Command and Control on-the-move through to further, more specific UAS studies. 2. LAND EXPERIMENTATION 2.1 Modelling and Simulation of a TUAS Late in 2009 it was determined that modelling and simulation of proposed UAS to support tactical manoeuvre would be able to demonstrate possible combinations of airframe quantities and associated payloads to meet the perceived longterm needs of Army whilst providing an avenue to elicit and better understand customer requirements. Inspired by technologies developed by Boeing Defence UK (BDUK) in 2008[1], the SAL developed a virtual, net-centric, warfighting, experiment environment that can enable Army personnel to use Human-in-the-Loop (HITL) desktop consoles to participate in a realistic Land scenario either in a mounted or dismounted role. Realistic HITL UAS Ground Control Stations (GCSs) operate virtual UAS with representative payloads that can then be accessed by the tactical commander to enhance situational awareness (SA) and decision making throughout the scenario. Through the careful capture of all the data produced in the experiment analysts are then able to identify where the capabilities provided by the UAS either assisted, hampered or had nil effect on the tactical commander in the conduct of the mission. Rather than focus on a single Tier II UAS, such as Shadow or Hermes, the SAL concentrated on creating an experiment environment that can demonstrate to Army, through modelling and simulation, that a greater quantity of smaller UAS with a variety of interchangeable payloads may provide the desired tactical capability enhancements as an alternative to the procurement of a limited quantity of larger Tier II UAS. The ability to work with Defence to identify alternative solutions to procurement problems has existed in Boeing for a significant period of time. From a Land perspective; however, it has only been recently that International- Analysis, Modelling, Simulation and Experimentation (AMSE) laboratories have been working in collaboration with Defence in the Land domain.

2 2.2 Defence Collaborative Experimentation During the period 2004 to 2007 Boeing Australia Ltd and the Australian Defence Force conducted a campaign of experiments as part of the Joint Collaborative experimentation Boeing (JCEB) program. JCEB highlighted to Boeing that collaborative experimentation with operators from the ADF provides both organisations with insight into how future forces can operate more effectively and the technologies required to support them. Whilst JCEB focused heavily on Air and Maritime warfare[2], Boeing identified a need for a representative HITL experimentation capability in the land domain. An iteration of this Land HITL capability was developed for a series of experiments in the BDUK Portal during 2007 and 2008 [1]. The SAL elicited a great deal of knowledge from this development and Experiment NIGHTJAR thus provided an opportunity not only to inform TUAS payload and platform quantity requirements but also to test this capability in Australia for potential future use by the ADF and other customers. 2.3 A Discovery or Hypothesis Experiment? During the planning phase of NIGHTJAR it was decided that a discovery experiment would best suit the broad task of examining TUAS requirements. An experiment of this type allows for new systems, concepts and technologies to be introduced and analysed for military utility[3]. The intent was to use NIGHTJAR as an opportunity to examine the broad impact of these new technologies, with follow-on experiments testing specific hypotheses in a constructive manner or focusing on key areas of interest using HITL activities. Whilst the SAL personnel were very aware of the limitations of conducting a discovery experiment in a short timeframe, it did provide the best return for the team as it allowed a variety of options to be examined over the course of five days (including training). For increased analytical rigour plans were made to examine the impact of specific sensors and UAS implementation strategies in the SAL s constructive tools allowing for batch runs to be conducted. This process would provide larger sample sizes against which the SAL s analysts could test for significance and examine areas of particular interest in detail under a number of operating conditions. At the time of writing this paper further, in-depth analysis has not been flagged as a requirement; however, tool development to support this has commenced. 2.4 Aims of Experiment NIGHTJAR The primary aim of Experiment NIGHTJAR was to conduct discovery experimentation focusing on the impact of differing UAS quantities and payloads on tactical commander situational awareness and decision making. The focus areas of the experiment through which the key Measures of Effectiveness were drawn were: Situational Awareness (SA) Decision Cycle (Boyd Cycle)[4] UAV interoperability (types and payloads) Network Centric System interoperability Tempo of operations A secondary aim of experiment NIGHTJAR was to develop capabilities for the International-AMSE arm of Boeing s Phantom Works. With an ever expanding array of Laboratories across the globe, an International Analysis Environment is being created to provide a suite of tools and capabilities to support any Defence program worldwide. Experiment NIGHTJAR, whilst principally a Land experiment was also used to develop UAS and Net-centric capabilities applicable to any domain. 3. CAPABILITY DEVELOPMENT 3.1 Development of a Virtual NCO UAS Environment In order for the experiment to be successful a number of capabilities required significant development. Within the Commercial-off-the-shelf (COTS) software realm, there were few products that accurately represented both control of a UAS as well as providing a high degree of fidelity for sensors. In 2008, The Boeing Defence UK (BDUK) Portal developed a Reconfigurable Operator Console for Experimentation (ROCX), to simulate a UK Bowman Battle Management System (BMS) console. This application is an extension of MÄK s VR-Forces[5] and provides Common Operating Picture functionality, Reporting, Calls for Fire, Targeting and remote sensor control. Capitalising on this initial development, the SAL, through assistance from BDUK, was able to reconfigure ROCX to simulate a UAS ground control station console. Through discussion with Boeing s Unmanned Systems and through study of the NATO Standard for UAV interoperability (STANAG 4586) the basic GCS functionality was replicated. This functionality included air speed, altitude, turn rate, loiter patterns, moving map and navigation data. Figure 1: GCS Operator Console with GMTI Radar on

3 The resulting HITL GCS enabled the user to set either a pre-determined mission route for the UAS or be able to control the UAS through allocation of loiter patterns and waypoints whilst in-flight. In order for the UAV operator to impact the Virtual Warfighting Experiment, it was essential the Electro-optical sensor was able to view the same terrain and entities that the participants in the Land component were interacting with. Terrain correlation always poses a great deal of problems in virtual warfighting simulation as the experiment developers seek to federate a number of applications each modelling the terrain in a slightly different fashion.[2] Rather than trying to copy the next-generation terrain representations of Virtual Battlespace 2 (VBS2)[6], it was determined the best way for the UAS operator to view the same scene was for the sensor of the UAV to be hosted in VBS2. Through attaching a VBS2 EO/IR sensor to the ROCX entity it enabled the UAV operator to view all the levels of detail of the VBS2 terrain as well as the 3D models of each entity present in the scenario. Importantly the VBS2 sensor could be controlled the GCS console, minimising operator workload and providing increased realism (see fig 1). One of the other key benefits of using ROCX as the underlying application was that it enabled the GCS to interact with the ROCX BMS used by the experiment participants. This provided the tactical commander with a fully net-enabled SA environment where the UAS operator could designate target locations through his control station and the commander could interact with these designations on his Battle Management System and provide targeting data to Offensive Support assets. It also meant that a tactical whiteboarding capability could be used across the network to accurately convey target indications, updates to orders, changes to routes etc. The final development for the ROCX BMS was the creation of a basic GMTI radar capability. This sensor could be turned on or off depending on the option s UAS payload requirement. The GMTI radar provided the GCS operator with a radar track of any moving vehicle within the sensor field of view and also enabled the operator to slew the EO sensor to the track for rapid Positive Identification (PID) of hostile vehicles. Further capability development took the form of video capture. International AMSE software engineers developed a program to capture the video of VBS2 which could then be streamed over the network and viewed by other personnel participating in the federation. This enabled replication of Rover III/IV UAS feeds as well as net-centric video on demand capabilities. Further iterations of this software will enable communications degradation of video quality and enable the annotation of video and stills as part of the tactical white-boarding toolkit. This video streaming capability also enabled the resolution of VBS2 to be realistically transmitted so that both the GCS operator and those accessing the feed were not receiving the near-perfect video available on a VBS2 desktop console. Other key capability developments were of the SAL s Boeing Analysis for Simulation Environment (BASE) analytical tool.[7] Though the development of VBS2 scripts which capture and transmit key data at regular intervals, engineers were able to capture the VBS sensor field of view intersections with the ground and consequently plot detailed coverage areas of the EO sensor to be use in follow-on analysis. Figure 2: VBS2 UAS sensors represented in BASE BASE is generally a quantitative analysis tool; however, during NIGHTJAR the difficult task was to capture changes to subjective measures such SA and decision making. Often changes to these measures can only be captured by looking at the results of participant journals. In the lead up to the experiment it was identified that the crucial qualitative measures of effectiveness (MOEs) were going to be generated by the commander who was central to the activity and whose reactions to situations needed to be recorded throughout each mission. In order to capture these reactions, engineers within the SAL developed ScribeL, a scribe logging tool that enables a member of the experiment controlling team to input comments either directly through a text box, or via pre-configured buttons in the GUI which output a pre-determined text string based on the metrics that are trying to be captured in the runs. These strings are then sent as a DIS comment PDU across the network, logged by the simulation logging software and then able to be analysed in BASE. To the engineers credit all the new capabilities worked flawlessly thus enabling a very successful experiment conduct. Whilst this paper is presented by only two members from the SAL, the success of the activity was a direct result of the work of the entire SAL team and members of the International AMSE organisation.

4 4. EXPERIMENT CONDUCT 4.1 Experiment Construct One of the principles of warfighting experimentation is to ensure that the scenario in which the experiment takes place is realistic and provides ample opportunity to test the effectiveness of the system that is being evaluated.[8] In conjunction with this and as discussed with the development of ScirbeL, changes in SA can be very subjective and can be extremely hard to draw out of an experiment. Consequently, the experiment was designed to immerse a tactical commander in the scenario and provide him with enough command realism to ensure that he was able to adequately react to changes in the situation based on the information presented to him. platoons HQ which represented an entire company of 14 vehicles. The Officer Commanding (OC) and platoon commanders each had access to a Battle Management System which enabled excellent co-ordination and control of the non-military personnel. The OC was able to use a Combat Net Radio simulator, CNR-Sim by Calytrix[9], to give orders by voice, whilst being able to use the tacticalwhiteboarding functionality of the ROCX BMS to provide real-time visual representations of the changes to routes, locations of interest and target indications. To ensure realistic communications flows, the platoon commanders had to relay orders from the OC to their three virtual section commanders. Although the section commanders were not present in the experiment, by simulating the voice traffic between the platoon commanders and the remainder of the platoon, a more realistic depiction of the time lags faced by the OC in being able to enact new plans was experienced and a more realistic communications picture for a mission could also be recorded and subsequently analysed. Due to the nature of this being discovery experimentation, a great number of runs was not a key facet of this experiment. Instead, it was taken that a couple of runs of a number of different options were going to provide spread of results leading to insights which would drive future experimentation. Figure 3: NIGHTJAR Bushmasters with UAS The scenario was based around a Bushmastermounted infantry company given a time sensitive targeting mission to kill/capture a High Value Target (HVT) in Oruzgan Province in Afghanistan. The commander had varying numbers of UAS in support which were able to stream video into the company headquarters. He was also provided with a Battle Management System to interact with the platoon commanders and had various offensive support assets available. Enemy forces consisted of Taliban insurgents acting as escorts to the HVT, spotters and a utility vehicle with a heavy machine gun mounted in the rear. In order challenge the UAS operator/s as well as adding to the realism of the scenario, clutter in the form of civilian personnel and vehicle traffic were present in vicinity of the target compound. As a result of having a limited number of staff, the experiment had to make best use of the available technologies to ensure that the scenario was realistic without creating a large training overhead for non-military SAL participants. As Artificial Intelligence entities in VBS2 can be difficult to control and; therefore, exhibit realistic behaviours, it was decided to limit the Company to four vehicles; a company HQ and three The options (see figure 4) included a baseline UAS, similar to that provided by the current, in-service ScanEagle. It consisted of an Electro Optical /Infrared (EO/IR) sensor only. The other options consisted of a number of different payloads including a basic Ground Moving Target Indication (GMTI) radar capability a Laser Marker/Designator capability and a weapon system. Options also included different quantities of a generic UAS platform that could host each of the differing payloads. OPTION QTY UAS PAYLOAD TYPE 1 1 EO/IR 2 1 EO/IR, GMTI, LASER 3 2 EO/IR, x 2, LASER 4 2 GMTI x 2, LASER 5 2 EO/IR x 2, GMTI x 1, LASER 6 3 EO/IR x 3,GMTI x 2, WPN x 1, LASER Figure 4: Experiment Options 4.2 Architecture Participants were set up over three rooms in the SAL. The main theatre was used for the OC s Bushmaster, which consisted of a VBS console for the OC s driver, the OC s out the window view, a BMS and a UAS sensor feed. Also situated in the theatre were the white force stations. These included positions for a System Administrator, BASE operator, administration stations for VBS2, ScribeL and the experiment lead.

5 complete multiple runs of each option. This resulted in a total of 12 completed and valid runs by the end of the week. Each run was conducted in the following manner: overarching mission orders OC confirmatory orders federation sequence mission conduct mission debrief. Figure 5: Company Commander s station A second room consisted of the enemy player consoles and a strategic UAS console to be used in option 6. The final room consisted of all the blue force players. There were three platoon commander VBS2 consoles each with access to a BMS and two UAS GCS positions. Each ground control station consisted of a ROCX GCS terminal and a VBS2 sensor console. Figure 6: Platoon Commander s Station All participant consoles, including the experiment co-ordinators, were connected through the SAL s UNCLASSIFIED network with each participant connected to a virtual communications network through CNR-Sim. Battle Management System (BMS) Voice Communications Simulation Analysis Tools Simulation backbone Virtual Crew Stations Crew Station Server Virtual Ground Control Station (GCS) Operational Analysis Communications Effects God s Eye Viewer Server Figure 7: Federation Functional Architecture 4.3 Conduct of the Activity The experiment was conducted internally within the SAL over the course of one week. Manning was broken into three components, a White Cell for co-ordination of the experiment; a Blue Cell controlling all the friendly forces and a Red Cell controlling the enemy forces. Each option was initially run only once; however, with technology achieving an unprecedented high level of stability the SAL team was able to All components of the experiment were recorded either through logging of the network traffic, through journal inputs or through note-taking during mission conduct or during the debrief. These results were then able to be analysed and key insights communicated to potential customers. 4.4 Analysis and Results The initial experiment planning sessions for NIGHTJAR focused on what Measures of Effectiveness (MOEs) would be used, how that data would be captured and ultimately analysed. It was evident that some of the key metrics were going to be of a qualitative nature and this posed some problems from both a capture and analysis perspective. The development of ScribeL and a number of unique data capture capabilities in BASE provided the experiment analysts with a means to extract the required metrics during the conduct of each run. Combined with this development, the SAL s history with conducting experiments of this type provided a set of procedures that catered for capturing the qualitative comments of the operators via journals and scoring realism. Detailed After Action Review (AAR) sessions were held on completion of each run to allow the experiment coordinator to capture input from the wider group of participants and these findings were very useful in determining key insights from the activity. Two of the key MOEs in NIGHTJAR were tempo and total UAS sensor coverage. In the case of NIGHTJAR, tempo was assessed through recording the time it took for the commander to complete a decision cycle in reaction to a standard event during the mission. ScribeL buttons were preconfigured to cater for Observation, Orientation, Decision and Action and they were used to rapidly capture the stage of the decision cycle from the moment the commander was made aware of a particular event to an action based on that event. By plotting the tempo of decision making across runs, the impacts of differing UAS quantities and payloads could be both analysed and displayed. It was also recognised early on that the total area surveilled by the UAS would be a MOE of significant interest. Whilst the capability to capture this MOE was developed by the SAL Technical Team in time for the activity, an area of interest that was not appreciated by the staff was how to capture the time multiple UAS spent with their sensors dwelling on the same area.

6 3. Alberts, D.S & Hayes, R.E.(2002) Code of Best Practice for Experimentation, CCRP Publication Series: United States 4. Osinga, F.P. (2007) Science, Strategy and War: The Strategic Theory of John Boyd (Strategy and History), Routledge: United Kingdom 5. MaK Technologies, VR-Forces web page, 6. Bohemia Interactive, VBS2, Virtual Battle Space web site, Figure 8: BASE 3D visualisation during debrief The ability to display how employment of multiple UAS often lead to duplication of coverage areas could be appreciated by experiment participants using the BASE 3D visualisation capability. Although this method did highlight how the UAS were employed it was deemed a necessity to also display the time each sensor dwelled on the same area at the same time. The degree to which the employment of multiple UAS impacted overall coverage areas was certainly a surprise to the experiment planning staff. 7. Nixon, A. (2004), "A Flexible Software Architecture For Visualising Simulation Data", SimTect Proceedings Kass, R.A. (2006) The Logic of Warfighting Experiments. CCRP Publication Series: United States 9. Calytrix, CNRSim, 5. CONCLUSION Experiment NIGHTJAR was an enormous success for the SAL. It provided an excellent environment to implement a suite of new capabilities for virtual and constructive simulation whilst also showcasing the SAL s ability to conduct HITL experimentation in a net-centric Land domain. With a high degree of interest from internal and external customers to replicate the NIGHTJAR experiment environment for a range of UAS and net-centric studies, capability development in this field will continue for both the SAL and other International AMSE laboratories. REFERENCES 1. Arnott, A., Lines, A., Winskill, J., McMahon, A. (2008) Creation of a Warfighting In Complex Terrain System of Systems Analysis Environment SimTecT Proceedings Brownie, R & Rogers, S. (2007), Supporting the ADF through Virtual Warfighting and Experimentation, SimTecT Proceedings 2007