Content area
Current uncrewed system (UxS) solutions tend to operate with tightly coupled command and control systems, making it difficult to contribute to operating as an integrated force. The case presented in this article is used to reason at the conceptual level about the different requirements and approaches for a future Norwegian UxS Integrated C2 system in order to inform the national development of an UxS Integrated C2 Reference Architecture. This is one in a series of papers that develops a mission engineering approach and represents functional analysis needed for future acquisition of Norwegian UxS. Based on this work and the development of the situated cognitive engineering (sCE)-method eliciting knowledge, and knowledge acquisition information, we make key findings for outlining a strategic guide for an initial Norwegian UxS reference system and set-up (manning, organization, and technical know-how). UxS solutions must be available to support ISR services for a variety of tasks and units on all military branches and levels. An UxS reference system must be adapted to the operational area and be available to operate within a harsh environment at the Northern Flank of NATO supporting those who need the information from sensors and/or decider and effector capability. Modern UxS solutions are based on human control and management, which entails human autonomy teaming which can be labor-intensive, with the potential for cognitive overload as well as bottlenecks in information processing (Frey et al., 2018; Hamstra et al., 2019). In the article, we present a framework that support future acquisition of Norwegian UxS that suggests that autonomy must be distributed to reduce vulnerability and be scalable to handle emergency adapted to the Northern Flank of NATO environment, e.g., an autonomous system that interacts with its surroundings demonstrating a cooperative design approach with new opportunities (e.g., with and without artificial intelligence (AI) support Endsley (2023)). We claim that a common future acquisition framework of Norwegian UxS applications (with AI) can reduce the burden on the operator based on results from our functional analysis (sCE-method) and empirical studies.
Introduction
In the article, we draw on research by the Norwegian Defence Research Establishment (FFI) (Midtgaard and Nakjem 2016; Mathiassen et al. 2022; Nummedal 2021). Norway has chosen a development strategy that involves different parties from industry, defense, and research under a common UxS-program. We name this the triangular collaboration model between industry, defense, and research demonstrating a fast method to adapt military UxS solution systems to its primary users. Small nations such as Norway need to conserve their development resources. One of the strategies is to allow a few demonstrator systems to be explored with almost identical framework of autonomy with minor adjustments to different UxS applications, e.g., a joint commitment to autonomy and several application projects under a common framework, which ensure synergies (Kaber 2018a, b). Exploration of limitations and possibilities with the “same” autonomy demonstrate a lean approach to separately handle UxS dilemmas and simultaneously integrate UxS in a command and control (C2) Reference Architecture to inform future national UxS investments (hybrid autonomy layer (HAL); Krogstad et al. 2020). In this particular paper, we expand on how human-autonomy teaming mechanisms could be related to the currently available frameworks for autonomous systems.
There has been considerable work conducted towards a common C2 architecture. The US Navy and US Air Force have followed services-oriented architecture (SOA) approaches with the Unmanned Systems Control Segment (UCS) architecture and Open Mission Systems (OMS) Universal Command and Control Interface (UCI) respectively. The US Army has adopted human portable “common controllers” with android smartphone-like technology and the development of the Robotics and Autonomous Systems—Ground (RAS-G) Interoperability Profile (IOP) AEP 4818. In parallel, the NATO community have restarted development of STANAG 4817 aiming for a multi-domain control system standard. Commercially, there are a number of industry offerings for multi-domain common control systems (O’Neill et al. 2023 [p. 2]; Perkins 2017; Priebe and Douglas 2020).
On behalf of Norwegian MoD (Norwegian Defence Staff), FFI explore several demonstrator systems at the same time. This ensures synergies with FFI’s partners and makes it uncomplicated to have a joint investment in autonomy in several application projects (UGV’s, AUV’s, and UAV’s). Common for all these applications are as follows: (1) Autonomous systems interact with its surroundings and cause high operational demands. (2) Modern UxS operations are labor-intensive and are dependent on human factor knowledge. (3) Autonomous systems cause high pressure on drone pilots during operation and need uncomplicated interactions design (Mathiassen et al. 2022).
In this paper, we build on our prior discussion (Stensrud et al., 2024) of how some human-autonomy teaming (HAT) design approaches (mechanisms for coordination), specifically levels of automation (LOA) (Frame et al. 2020; Minos-Stensrud et al. 2021; Stensrud et al. 2021a, b), mixed-initiative (MI) (Chanel et al. 2020), and coactive design (COAD) (Johnson et al. 2011, 2014& 2018), could be combined. The different mechanisms are, among other things, distinguished by their different degree of specificity of the task done by both the human and machine. While LOA is traditionally requiring great specificity, this is less the case in MI and COAD where what the machine may do and how it interacts with the human is more emergent. MI may use standardized protocols for emergent collaboration while COAD is even more open in that the exact interdependencies are continuously worked out.
We have now added more detail on the various system evaluation templates and schemes that are part of a USV and UxS (AUV) system may use and its influence on autonomy level and teaming. We also expand on this by suggesting how the different collaboration modes (LOA, MI, and COAD) may be used to leverage the strength of the two principal interaction schemes that we currently consider and develop. As indicated in Stensrud et al. (2024), much of the research to date have discussed these interaction approaches separately. In Stensrud et al. (2024), we discussed how humans and artificial cognitive systems can be orchestrated to enable the handling of complexity and dynamics of an environment (Kirlik et al. 1993), e.g., handling military threats, and how different designs are affecting mission solutions. Specifically, we suggested that there are trade-offs between the HAT designs so that LOA and to some degree MI provide better coordination in low complexity and low dynamics environment (Grote et al. 2018), while COAD could support coordination in high complexity and high dynamics. LOA and MI could be relatively less costly in low complexity and dynamics while the opposite holds for COAD. Ways of using these HAT designs in a complementary way were suggested to support coordination through both pre-scribed route planning and feedback, such as by integrating external and internal feedback in prediction of future action. We illustrated our suggestions previously made through a conceptual use case of autonomous unmanned drone collaborating with fighter jets, which provide additional nuance to our theoretical discussion (Stensrud et al., 2024).
We claimed in Stensrud et al. (2024) that “A problem facing HAT is how to be able to adjust own actions, both human and machine, considering environmental characteristics, in such a way that the HAT is continued and able to perform the joint tasks despite various environmental conditions.” Evaluation of coordination through HAT designs under different environmental characteristics are more broadly discussed in Stensrud et al. (2024) related to a simplified air-to-ground military task. In the current article, we go into more detail regarding the specific characteristics of the sub-surface environment where UxS could be used and build an understanding of the functional characteristics of such a UxS and how the functions can be supported by human-autonomy teaming. We take as our principal example how the MI approach can be distinguished according to how much tasks are not only delegated but also coordinated among the USV and UxS. Before doing so, we introduce the combination of swarm technology and HAT-designs when using UAVs. As we will see, the degree to which interactive collaboration can emerge very much depend on how advanced the reference architecture is. This nuances the knowledge of how task environment influences a particular interaction mechanism, and this knowledge can provide the ground for further discussions of the specific man-unmanned interaction mechanisms, e.g., LOA, MI, and COAD, useful for directing and controlling UxS under different environmental conditions in underwater mine-hunting.
Human-autonomy teaming handling challenging environments
In Stensrud et al. (2024), we conceptualized the external environment according to two main dimensions: “(…)complexity as the number of elements and number of relations among elements in an environment (Schneider et al. 2017) and dynamics as the rate of change in elements in the environment (Dess and Beard 1984).” We furthermore suggested that there are trade-offs between different interaction mechanisms so that there are trade-offs between the HAT designs so that LOA and, to some degree, MI provide better coordination in low complexity and low dynamics environment, while COAD could support coordination in high complexity and high dynamics. LOA and MI could be relatively less costly in low complexity and dynamics while the opposite holds for COAD. The LOA requires that we specify a concrete definition of the tasks which the machine can do and so lend itself to environments that have low complexity (i.e., is easy to understand) and are less dynamic (i.e., tasks can be repeated). MI accommodate that the human and machine can take initiative according to the situational demands, and so require a more advanced machine as well as man operation but with more possibilities in complex and dynamic environments, although with more coordination needed between the entities. Lastly, COAD can be useful when there are unclear and changing task requirements because interdependencies between man and machine are not predefined. This comes to an uncertain and relatively high coordination cost due to the multiple possible interdependencies. Summarized, we claimed in Stensrud et al. (2024) that “We broaden this outlook and suggest that MI and COAD are particularly relevant as ways of conceptualizing HAT considering more dynamic interactions between man and machine than the LOA perspective. By comparing these approaches to HAT we aim to contribute to the need for developing more unified theories of HAT as suggested by O’Neill Th. et al. (2022, p. 927).” “Using the different interaction-mechanisms in different phases, with different levels of complexity and dynamism, could be one solution to overcome the trade-offs between the interaction-mechanisms. In the following we exemplify how the UxS and human operator may be complementary in face of different degree of autonomy for the UxS..” Stensrud et al. (2024).
LOA, MI, and COAD can be used in numerous ways to specify the collaboration between USV (unmanned surface vessel) and sub-surface UxS. We now outline (in Table 1) some ways that the collaboration can be instantiated by the mixed-initiative (MI) method of collaboration. We use MI as our exemplary interaction-mechanism as it encompasses both the structured LOA approach as well as point to a more emergent interdependencies in COAD.
Table 1. Mixed-initiative within a dialogue framework (Allen et al. 1999 [p. 15])
Mixed-initiative levels | Capabilities | Remote human coordinator | USV system | Sub-surface UxS |
|---|---|---|---|---|
Unsolicited reporting on sensing | Agent (USV) may notify others (update sub-surface UxS) of critical information as it arises | Initiate the USV system to do situation reporting based on clear criteria interpretable by the USV system within a limited complex and dynamic situation | Are notified if the situation or plan changes | Are notified if the situation or plan changes |
Subdialogue initiation on identification of mines | Agent (USV) may initiate subdialogues to clarify, correct, and so on | Granting the USV a role as interpreter of the situation. The USV has a broader complex and dynamic range | Coordinate the subsequent interaction by subdialogue initiation to clarify status | Answer the subsequent interaction by subdialogue initiation feedback on clarifying status |
Fixed subtask initiative on detection of mine (optionally launch UxS) | Agent (USV) takes initiative to solve predefined subtasks (launch sub-surface UxS into water) | Granting the USV additional subtasks based on their capability of situation awareness and tasking competence | Initiates clearance or request for that might take several fixed subtask interactions | Initiates or request for authorizations (pre-clarification) that might take several fixed subtask interactions |
Negotiated-mixed initiative on evaluation | Agents (USV) coordinate and negotiate with other agents (USVs or sub-surface UxS) based on evaluation. | Delegating responsibility for the USV to “run the mission” of the UxS in complex and dynamic situations | At the final negotiated mixed-initiative level, there is no fixed assignments of responsibility or initiative | At the final negotiated mixed-initiative level (e.g. post-assessments)), there is no fixed assignments of responsibility or initiative |
With respect to MI, we make an example of what the USV (unmanned surface vessel) and sub-surface UxS can contribute to in a future naval/maritime mine countermeasure (NMCM/MMCM) capability that includes unmanned, autonomous vessels. In Table 1, we are specifying how parts of a naval mine countermeasure (NMCM) operation may be designed. We suggest, given the challenging environments, that we need to explore autonomy functions, decision support mechanisms, and human autonomy teaming methods to clear the surface and sub-surface (waterbed) for mines. Reading the table from top to bottom, we see a progression from the machine doing simple tasks to progressively more advanced tasks. This also corresponds to a shift from human-in-the-loop to human-on-the-loop, as well as machines taking on advanced coordination of other machines. We are entering a situation where the sub-surface UxS is better placed to do a specific part of the task (e.g., identification of mine) and at the same time the UxS is best at do general sensor coverage because it can be closer to the waterbed, versus USV. USV will do the collaboration part of the MCM process due to its better sensor-fusion and due to legal requirements that a human in- or on-the-loop should do final clearance (reach back). Based on Allen et al. (1999), we propose that joint activity is about interaction and negotiation, i.e., mixed-initiative within a dialogue framework between the USV (unmanned surface vessel) and sub-surface UxS (Table 1) and becomes more adaptable to the situation than the prescribed LOA framework.
In future developments, we may see that the USV and UxS may suggest task priorities or even develop, through experience and through collaborating with machine and man, novel integration of tasks. In particular when the complexity and dynamism is outside of the range of what was predefined in the task definition, this may become a reality. The realization of collaboration in such situations may initially require COAD type of man–machine and machine-machine interaction mechanisms. In general, the COAD extend the notion of human–machine teaming to a more systemic view than what LOA and MI does, as it emphasizes the information sharing (on OPD (Stensrud et al. 2020b)) between entities, something that is more restricted in the LOA and MI. COAD proposes three essential interdependence relations: observability, predictability, and directability (OPD) (Johnson et al 2014) by defining roles and individual requirements (Johnson and Bradshaw, 2021). Johnsons, Bradshar, and Feltovich (2014) claim that it is an invert relationship between automation and interaction. One of many advantages of COAD is that the focus on core interdependence relations can provide a formative tool for designers called interdependency analysis (IA). Interdependency analysis (IA) supports what could be automated, and as a fundamental principle of COAD, interdependence must shape automation (Stensrud et al., 2024). Over time, as interdependencies are specified and standardized, MI and LOA type of interaction mechanisms may be used. Thus, the reference architecture and HAT mechanisms may co-evolve.
Returning to our example, a mixed-initiative systems offer applications (e.g., an interactive planning application) where both the user (e.g., a remote coordinator) and agent (e.g., a (pre-configured) USV (unmanned surface vessel) and sub-surface UxS system) are notified if the situation or plan changes. At this basic level, the mixed-initiative system does not then coordinate the subsequent interaction on this first step allowing unsolicited reporting. The next level (table) involves subdialogoue initiation; the mixed-initiative system asks for authorizations (pre-clarification and post-assessments) that might take several fixed subtask interactions between USV system (pilot) and UxS (agent). Mixed-initiative system is responsible for choosing routes, communication (emission control), and re-charging for each action along the pre-planned timeline of sub-tasks of a MMCM mission. At the final negotiated mixed-initiative level, there is no fixed assignments of responsibility or initiative. The mixed-initiative system supporting the USV system (including a mother-ship with more operators) and UxS, will according to Allen et al. (1999) help to monitor the current task solution and thereby evaluate whether it should take the initiative or not. It bases this decision on many factors: demands and work load of the agent (s) (UxS) and risk of losing the USV system (including a mothership). An assumption in the MI design is that the one entity best able to carry out a subtask should take the initiative to do so, agnostic of whether it is man and machine. In this way, it, in a sense, incorporates in its mechanism some idea of whether an entity is fit to do the particular subtask, thus requiring at least a basic understanding of the current status of an entity in relation to the task and the environment. Later on, Jiang and Arkin (2015) have developed this definition to encompass robots. Jiang and Arkin (2015) suggest that feedback from an external environment or inferred state of environment can trigger initiative in a reactive or deliberate way; however, uncertainties of the environment can make initiative reasoning challenging (Kirlik et al. 1993).
A strategy to support Norwegian UxS systems design
In this sub-chapter, we present on-going research proposing an approach for development of UxS systems in Norwegian Defence. First, we discuss interaction design patterns to be considered; second, we introduce a fundamental functional analysis method; and third, future tools for evaluating the design and modeling of prototyped UxS applications. And, fourth, an UxS Common C2 System Reference Architecture is in use to ensure that the common controller implements the same set of interfaces as specified in future applications (Fig. 1) (Park et al. 2020).
Fig. 1 [Images not available. See PDF.]
Domains of human systems integration: An autonomous system interacts with its surroundings. The shift of single to multi-UxS operations affects operators of these systems because their attention is limited, so UxS systems have to be built wisely (adapted from Nummedal 2021)
The method and intention of the framework are presented outlining a strategic guide for an initial Norwegian UxS reference system and set-up (evaluation of alternative UxS (classes), manning, organization, and technical know-how).
It has become clear that treating the system as separate from the users results in poor performance and potential failure in the operational setting. Continued growth in technology has not delivered desired results. Systems engineers and others are beginning to understand the role humans play in technology systems. The core challenge is to balance successful hardware and software solutions with human friendly implementations. The Foundation of Technological Principles and Human Factors Knowledge is therefore needed to take the operational demands seriously when prototyping UxS systems (Fig. 2).
Fig. 2 [Images not available. See PDF.]
Domains of human systems integration: Autonomy is a prerequisite for enabling a multi-UAV system (adapted from Nummedal 2021)
In such cases, we have previously suggested (Stensrud et al. 2024) that there are often requirements for interaction design patterns based on foundations in design theory, interaction between humans and UxS systems (Lyons et al. 2021, p. 2).
We focus on human operators that collaborate with UxSs to solve missions (Rebensky et al. 2022a, b; Chanel et al. 2020). There is a need for an easy interaction mechanism and standard operation procedures for HAT, informed by more general HAT design types, to ensure overall authorization of coordination among UxS-vehicle operators and -mission operators when in various contexts. The next step is to embed these tactical man–machine systems into larger organization (e.g., multi domain operations) to gain insights and experience to develop well-structured and clearly defined roles, tasks, and processes that are to support communication and collaboration between involved actors. This indicates foundations for co-evolution of architecture and HAT design.
The systems engineering team relies on specialists and operator to assist in analyzing customer requirements (see Fig. 4). Research has shown that aspects and components remained, until today, with no established methodologies or evaluation tools to link various human aspects to systems engineering models due to two reasons (Meilich, 2008): lack of relevant taxonomy linkage to system engineering (SE) needs (Stensrud et al. 2023a) and poor domain languages.
Most of the requirements for human systems integration are derived from requirements and specification for interaction design that shapes functions needed to provide use case that brings about effect providing the objectives for performance, efficiency, environmental, operational, maintenance, and training (see Table 1 and Fig. 3). One of the obstacles to realizing the substantial potential of proper interaction design patterns is the lack of clear articulation of foundation: technological principles, human factors knowledge requirements (Neerincx and Lindenberg, 2008). A clear articulation should be outlined in some sort of statement of work (SOW) or other authorizing documentation (specification) in collaboration with customer, to avoid that needed functionality or lack of a reference software architecture could cause to missing fail-safe behavior or untested and improper sub-task items imposing risk (i.e., standardization also make it easy to track requirements changes).
Fig. 3 [Images not available. See PDF.]
Domains of human systems integration: Autonomy must be distributed to reduce vulnerability and be scalable (adapted from Nummedal 2021)
An important component of the human systems integration plan should be a verification and validation process that provides a clear way to evaluate the success of human systems integration. The human systems integration team should develop a test plan (Stensrud et al. 2023a) that can easily be incorporated into the systems engineering test plan. The effectiveness and performance of the human in the system needs to be validated as part of the overall system. “It may seem more attractive to have stand-alone testing for human systems integration to show how the user interacts with controls or displays, how the user performs on a specific task.” (Ahram and Karwowski 2009, p. 1849). This methodology can address the performance of the human operator or maintainer with respect to the overall system and the situated use case (i.e., operational demands). The most important thing is to develop a close relationship between foundation and specification (Neerincx and Lindenberg, 2008) (Fig. 4) when evaluating the UxS system engineering process. To guide the functional analysis, we suggest following the method in Fig. 4. In a capability requirement context illustrated in Fig. 6 (i.e., the main boxes labeled Foundation, Specifications, and Evaluation), we use a situated cognitive engineering (sCE) method (Neerincx & Lindenberg, 2008 cited in Vught et al., 2020) (Stensrud et al., 2024).
Fig. 4 [Images not available. See PDF.]
Functional analysis—evaluation method is based on the situated cognitive engineering (sCE) method (Neerincx and Lindenberg, 2008 cited in Vught et al., 2020)
Functional analysis
There are many different definitions of the term functional analysis. Functional analysis involves the use of some form of procedure, that is, a formal procedure, for collecting and organizing data about an empirical phenomenon modeled into an appropriate model with a known format. This is illustrated by a functional analysis example in Fig. 4 and Table 1 (Stensrud et al. 2021a, p. 18).
The purpose of the extended functional analysis assessment model extending a situated cognitive engineering (sCE) method is to create the prerequisites for carrying out a structured assessment of the operational effect an UxS concept has within various areas of military application.
In general, one can say that such an assessment must be seen in the light of how far the concept development has come and thus how mature an UxS concept is both in terms of tasks, architecture, and HAT design. Early in concept development, an assessment model as presented (Fig. 6) can give indications as to whether the concept is sufficiently promising that it should be developed further. In further concept development, the model can also be used to identify hypotheses that must or should be verified. Examples of more sophisticated methods that can be used to verify relevant hypotheses are experimentation with vehicles in synthetical environments and experimental prototype vehicles (Hopkins and Schwanen 2018).
The assessment model is structured by defining one or more contexts that describe the context in which the object of analysis, in this case a configuration of UxS, is to be assessed. The individual configuration is assessed within the relevant context against a set of assessment criteria which are used as factors to estimate the operational effect it has. The result of the assessment will be knowledge about which configuration of UxS is estimated to give the appropriate operational effect in a given context. At the same time, the model will also help, if the assessment is done in several contexts, to be able to tell users the extent to which a configuration provides an operational effect in several contexts. The latter can be particularly interesting in relation to cost efficiency (Fitts and Jones 1947; Di Pasquale and Savill 2022) in connection with an assessment of which configurations of an UxS give the most promising operational effect in an overall perspective.
Use case – Valkyrie a Norwegian unmanned aircraft system
Valkyrie is a research system for distributed autonomy developed at FFI. It integrates autonomy software components from FFI, such as the perception system Warpath and the decision autonomy system hybrid autonomy layer (HAL) (Krogstad et al. 2020) on multiple autonomous UAV. The Valkyrie architecture is modular with respect to platform, payload, and communication infrastructure, and is built around a companion computer. The UAV that we use are a prototype platform built at FFI called the Flamingo which is a small quadcopter drone with an onboard companion computer.
Valkyrie is an unmanned aircraft system (UAS) in which associated unmanned aerial vehicle (UAV) units fall into class 1 A according to EASA rules and regulations. The Valkyrie UAS is a generic unmanned swarm system consisting of several cooperating UAV. The system integrates sensors, platforms, and control logic with a user interface that enables an operator to control several UAV simultaneously. An operational framework describes operator training for Valkyrie specialists, and technicians how to test and trial with different operations, e.g., Valkyrie UAS in an ISR configuration with up to 6 UAV units with a range of up to 5 km that can be operated by 1–2 people.
The purpose of the acquisition of the system is testing of an operational framework adjusted for UAV. In connection with this, a professional out-check and plan for qualifying operators for type-checking is prepared for operators on the Valkyrie system. The preparation of this plan is a collaboration between the Norwegian Armed Forces and Norwegian Defence Research Establishment (FFI), the Norwegian Armed Forces School (HVS), and the Norwegian Air Force’s Air Operational Training and Certification Center (LFTS). Type inspection and training will be held by FFI, for personnel deemed suitable. Type check Valkyrie T&P requires technical and administrative approval for test and trial operation (TFG-TP) from Norwegian Defense Materiel (FMA) Air Capacities.
The Norwegian Chief Air Force Weapons School is per instruction given the authority to issue regulations, informative and directive documents, and subject plans within the assigned area of authority applicable to the Norwegian Armed Forces, of which the area of authority UAS is included in the same instruction.
UAV class 1 A requires UAS introduction course and type check provided by LFTS. Approval of Type Check Valkyrie T&P includes dispensation for FFI to be responsible for type check in consultation with LFTS.
The purpose of the UxS system Valkyrie is to support side-ordered and superior military units with updated sensor information (red dot) to increase the supported department’s situational awareness (SA).
Execution and performance of the experimental swarm system will depend on the support that the certified operators get from the system’s interface, and which will depend on the visualization of ongoing task solving, linked to possible dynamic changes, possibly based on emerging (red dots) from the ongoing sensor data acquisition and/or by single- or multiple-source processing of this information, and/or in the case of direct or indirect assignment changes such as, e.g., change of operation area.
Valkyrie consists of Flamingo UxV units as described in previously ungraded FFI reports (Nummedal 2021), hybrid autonomy layer (HAL) (Krogstad et al. 2020) for decision autonomy, and with Warpath for scene understanding and a ground segment (ground station for control and information handling) from the swarm by UxVs. Each UxV communicates directly with each other over a mesh network and can act as relays for each other to the ground segment.
HAL uses Battle Management Language (BML) to both give and receive orders, and as such, BML is implemented in the LandX-UI (user interface), and can be used to give orders to vehicles controlled by HAL. The HAL framework reasons about such complex high-level commands by breaking them down into a sequence of simpler, low-level sub-tasks that are more suitable for a machine to execute. All tasks, both high level and low level, can start and stop behaviors which are processes that produces the appropriate outputs or internal state changes that solves the task at hand. The most obvious example is the waypoint guidance behavior that actually sends steering commands to the autopilot system in order to move the UAV to the waypoints computed by various tasks or other behaviors. HAL receives data from all the subsystems, and how it reasons internally about the sequence of steps that must be executed. In parallel with all of this, HAL runs multiple other behaviors, which each are responsible for handling specific situations or subproblems. For collision avoidance, HAL implements multiple algorithms, but uses an artificial potential field method as default.
One or more UxVs can be used to support guard and security missions (perimeter security of areas, monitoring of axis) to contribute to force protection during the movement of military units. During implementation, it will be possible to use single UxVs for sub-missions such as following mobile units.
UxV swarm behavior will be flexible and will be able to easily scale to perform several parallel monitoring missions such as axis reconnaissance, in parallel with reconnaissance and guarding and securing.
Cooperation with operational militaries: in experimental use and phasing in, will provide input for experience-based upgrading of user experience (UX) and user interface (UI) and procedures for using the system.
Use case—underwater mine hunting
In this sub-chapter, we present an evaluation model and explains how it can be used. Our UxS system capacities will be codified to be surface vessels, underwater vessels, and unmanned aircraft. Undersea, we have identified three contexts that may be interesting to investigate further. These are intelligence surveillance and reconnaissance (ISR), anti-submarine warfare (ASW), and mine counter measures (MCM).
To guide the development of man-unmanned concepts, we outline a set of tenets according to the following intentions: (1) Secure that the need of UxS development activities is identified; (2) Secure that relevant UxS ideas are developed further to solve future missions and tasks; (3) To ensure the integrity of the UxS framework, secure that the good ideas generated on lower level in the defense organization is deeply top-down rooted; (4) Provide for sufficient resources to be allocated to the development of actual UxS concepts; and (5) Secure concept development (R&D activity) is improving the military decision-making process concerning UxS prototyping. To be able to satisfy this, there is a need for a systematic approach (Stensrud et al. 2021a, p.14 i.e., purpose of evaluation; Stensrud et al. 2021b).
“In overt operations, an unmanned surface vehicle (USV) is used for high-speed transport and deployment of survey AUVs. The AUVs perform the search-classify-map phase over the full operation area using synthetic aperture sonar (SAS) and automated target recognition (ATR) processing. The USV constitutes a communication relay, with acoustical links to the AUV and radio link to the home base. Optional radio nodes on unmanned air vehicles (UAV) can increase the communication ranges. This network allows status and result samples to be sent from the AUV to the home base and operator commands the opposite way. The vehicle continuously evaluates its system performances in the local environment using a through-the-sensor approach and adapts its behavior according to the given mission aims. After the initial SAS survey, a new mission plan is automatically generated for recording identification images of the ATR contacts with the AUV’s optical camera. The confirmed mines are finally neutralized with a one-shot mine disposal weapon (OSMDW) launched from the USV.” (Midtgaard and Nakjem 2016, p. 1).
“Naval MCM technology has evolved formidably in recent years and Norway has played a leading part in this development. On-going development of new platforms and systems in Europe and the USA follows three main lines:
Modular systems on-board vessels that can perform both mine hunting and mine sweeping (as opposed to two separate vessel classes).
Increased use of unmanned vehicles (at surface and underwater) enabling the mother ship (with crew) to operate from a safe distance outside the mine field.
The MCM systems (equipment) are built as replaceable modules making it possible to refit the vessel for each operation. The sum of these changes is greater flexibility, increased safety for crew and reduced costs relative a one-to-one replacement of current vessels.” (Midtgaard and Nakjem 2016, p.2)
AUV swarm behavior will be flexible and will be able to easily scale to perform several parallel monitoring missions such as axis reconnaissance, in parallel with reconnaissance and guarding and securing supported by hybrid autonomy layer (HAL) (Krogstad et al. 2020) for decision autonomy. Equivalent HAL uses battle management language (BML) to both give and receive orders, and as such, BML is implemented in the SeaX-UI (user interface), and can also be used to give orders to unmanned maritime vehicles controlled by HAL in the same way as unmanned ground vehicles.
Evaluation of a use case
The assessment of a possible broad introduction of UxS in the Norwegian Armed Forces can be regarded as concept development where the hypothesis is that the UxS will provide a relevant operational effect that can be valuable in a larger campaign context.
This article presents a framework and a method for how operational impact assessments can be carried out supporting a strategic UxS plan for Norwegian Defence. The chapter is structured by arguing why this type of framework is important. The following sub-chapters give an overall principled presentation of the method and model that has been developed, and present and discuss the various parts of the model. The context can be an operation type. Anti-submarine warfare (ASW) and mine counter measures (MCM) are examples of what we can call types of operations. As for the functions, the operation types will have to be carried out within several scenarios. However, there may be reason to assume that the number of scenarios can be reduced in relation to more generic functions which should initially be able to be carried out within all types of scenarios. Readers interested in more details are referred to how such a model can be used in Stensrud et al. (2021a).
Attempts have been made to illustrate the connection between scenarios (use case), functions, and types of operations (operational demands) in Fig. 4 and Table 2 (Stensrud et al., 2008; Stensrud et al. 2021a, pp. 18–19). Our principled approach to context is done primarily through scenarios, for example, ISR/MCM, i.e., underwater mine hunting; this often means that the analysis object’s contribution to the exercise of the function must be considered within several scenarios. This focuses the analysis group on their assessment of the UxS system, i.e., sub-surface UxS system operational effect in various contexts.
Table 2. Functional analysis (Neerincx & Lindenberg, 2008 cited in Vught et al., 2020)
Specification | Description |
|---|---|
Function | Function description (command, control, communicate, collaborate, sense, inform) |
Use case (operation type) | Use case description (ISR/REA/SAR/MCM/ASW) |
Interaction design pattern | HAT design approach (LOA, mixed-initiative, coactive design) |
Objectives | Optimizing route plan, optimum track |
Effect | Emission control, counter measure |
Use case: subsurface UxS system
In the sub-chapter, we present a use case: subsurface UxS system as an analysis object. Subsurface UxS’s can be classified based on several criteria. The future Norwegian naval mine countermeasure (NMCM) capability includes unmanned, autonomous vessels. The Royal Norwegian Navy uses the Hugin autonomous underwater vehicle (AUV) in mine hunting operations. To facilitate transportation and extension of Hugin’s operational range in NMCM operations, the Norwegian Defence Research Establishment (FFI) is developing technology for automatic launch and recovery of sub-surface UxS (i.e., AUVs) using unmanned surface vehicles (USVs). Bitar et al. (2022) describes the Stinger launch and recovery system (LARS) mounted onto the USV Frigg, a method for automatic recovery of the Hugin AUV using this system, and a system that can estimate relative position and velocity of the AUV using a lidar sensor mounted on the USV. At time, NMCM operation is done with the help of a pilot USV and UxS system demonstrator. In the sub-chapter, we are looking five to ten years ahead in time, and analysis what system components to build in a future implementation of USV and sub-surface UxS for future NMCM operations.
Subsurface UxS payload/equipment and functions are presented in Fig. 5, framing on autonomy of a subsurface UxS as a future challenge.
Fig. 5 [Images not available. See PDF.]
Autonomous control of underwater platforms is a main feature for future success in maritime operations (adapted from Kalloniatis et al. 2020; Bitar et al. 2022)
Subsurface UxS’s can be classified based on a number of criteria.
In this context, analysis object means the object whose operational effect is to be assessed within a defined context.
Two overarching criteria have so far been defined. One is related to size and weight (hereafter referred to as weight class). The second is related to which payload/equipment the object carries (hereafter referred to as equipment class).
Size/weight provides strong guidance for many other criteria—performance scales in many areas with size. In the project, a simple categorization according to dry weight has been made as follows:
Small – under 100 kg
Medium – 100 to 1000 kg
Large – over 1000 kg
For the time being, four equipment classes have been defined as follows:
Sensor carrier
Weapon bearer
Communication/navigation carrier
Load carrier (personnel/material)
Within each equipment class, subclasses can be defined, for example, with different types and combinations of sensors.
In principle, each configuration, i.e., combination of size and equipment class, will be a separate object of analysis (Stensrud et al. 2021a, p. 17).
In the assessment context, it will obviously be appropriate to limit the number of variants. A first step in such a variant limitation process will be to sort out combinations which, for technical and practical reasons, will obviously not be relevant. The next step will be to carry out a closer assessment to identify which combinations are believed to be feasible within realistic technological and financial frameworks.
The description of the objects of analysis shall form the basis for the assessment of the operative effect within various contexts. The description must therefore give a practically applicable picture of the characteristics and capacities of the object of analysis which is sufficient to make sense in such a context. This means, among other things, that the technical specifications must be translated into practical terms. For example, a sensor range x and sensor resolution y mean that a sensor can detect an object of size z at a distance u, classify it at a distance v, and identify it at a distance w (Stensrud et al. 2021a, p. 18).
Some of the properties and capacities, such as speed, range, endurance, and load capacity, will typically be linked to the platform and be generic for the individual weight class. Others, such as sensor range and resolution and effect, will be linked to the equipment.
We have left most of the detailed analysis out of the article (Hellesnes and Bjørnsgaard 2007; Synnes and Hansen 2020; Stensrud et al. 2021a, pp. 16–19; Bitar et al. (2022)). Anyway, a systems engineering team relies on functional analysis to assist in analyzing customer requirements (see Fig. 4).
The principal structure of the functional analysis, evaluation, and compilation results (e.g., a semi-structured soft OR evaluation method) is presented in Fig. 6. However, showing where the analysis object with the maximum suitability in each context might not be the object with the highest overall score. There may be instances where the unmanned platform may for each context have a lower than maximum score, but where the sum over all the contexts might be the highest of all the analyzed objects/configurations, as shown in Figs. 7 and 8.
Fig. 6 [Images not available. See PDF.]
A slightly extended functional analysis (Neerincx & Lindenberg, 2008 cited in Vught et al., 2020)
Fig. 7 [Images not available. See PDF.]
Possible evaluation criteria template: scenarios—functions—types of operations when assessing a new capability within the strategic planning process (adapted from Stensrud et al., 2008)
Fig. 8 [Images not available. See PDF.]
A possible evaluation: scenarios—functions—types of operations when assessing a new capability within the strategic planning process (adapted from Stensrud et al., 2008)
The analysis object can be seen as the specific configuration of man-unmanned systems. Different analysis objects can be enumerated based on whether they conform to design principles. For example, the first analysis object could utilize the traditional LOA-approach (i.e., autonomy function in Fig. 5), while a second analysis object would use the mixed-initiative and the third would use coactive design. Other functions to be assessed are proprietary communication/navigation and sensor payload.
These different configurations, analysis objects, will then be evaluated according to analysis criterions of how well they solve the tasks in different environments. In the analysis criterion 1, they would for example solve a surveillance mission in low complexity and low dynamism, typically, in well-known coastal regions in the country of origin. High complexity and dynamism, a criterion 2, could be off-coast missions or missions in foreign coastal regions, e.g., where there is less preparedness in terms of mapping of the subsurface environment (both the geographical and the exact pattern of movements in that environment). To ensure scientific rigor, one would like to “tease” out the specific effects of say complexity and dynamism and the more precise analysis would treat these both separately and interactively.
Evaluation of the human decision-making and task load (i.e., human factors (Gombolay et al. 2017)) under these different criterions is critical to enable the full analysis of the viability of the different configurations, analysis objects. Such analysis should include an evaluation of the link between the architecture and HAT designs available: e.g., how well do the HAT design ensure the proper use of the machine, to what extent is the human able to coordinate the various entities.
In the assessment model (Fig. 7), the score is set quantitatively on the basis of a qualitative assessment which reflects both the importance of the assessment criteria in the context and the impact of the object of analysis on the existing structure. A score of + 5 thus means both that the assessment criterion has great importance in the context and that the object of analysis has a very positive influence; 0 means that it has no influence and a score of − 5 implies both that the assessment criterion has great importance in the context and that the object of analysis will have a very negative impact. The total score for an analysis object is the sum of the scores on the assessment criteria, as shown in Fig. 8.
The quantitative score gives an indication of which operational effect it is assumed that the object of analysis will be able to assess against an assessment criterion in each context. However, it says nothing about which properties of the object of analysis are assumed to produce this effect. It is therefore conceivable that two objects of analysis receive the same score on an assessment criterion in the same context but based on different characteristics. However, these properties can turn out to be completely different when assessed against other criteria in the same context or the same criterion in other contexts, and thereby lead to an object of analysis overall being considered suitable than the other within the framework of one context or across of multiple contexts. It is therefore essential to document the qualitative assessments that form the basis for determining the score (Stensrud et al. 2021a, pp.17–18).
The assessment is repeated for all analysis objects that are assumed to be relevant within the context.
In the specific case of sub-surface UxS, various sub-surface objects were assessed within various contexts according to Figs. 7, 8, and 9. The analysis object 1 evaluated in context #1: a maritime operational environment—underwater—with one or more analysis objects focused on detecting, identifying, and neutralizing mines (MCM). AUV class 1 was assessed to be most appropriate. Analysis object, AUV class 1 (small), is equipped with relevant sensors and is evaluated against AUV class 2 and 3. Aim of the assessment was to find out what operational effect the object of analysis will have in the various contexts. Analysis object #n was a future option sub-surface UxS.
Fig. 9 [Images not available. See PDF.]
A slightly extended functional analysis (Neerincx and Lindenberg, 2008 cited in Vught et al., 2020; Stensrud et al., 2008, Stensrud et al. 2021a)
Today’s small AUVs (e.g., Remus100) are less capable than medium and large AUVs (e.g., Hugin1000), regarding key performance parameters such as sensor range, sensor data quality, mission endurance, and depth rating. The Royal Norwegian Navy operates both Remus100 and Hugin1000 AUVs, but the latter is the preferred solution for mine hunting and other seabed surveys, except in very shallow waters (Midtgaard and Nakjem 2016).
Conclusion
The pilot UxS systems presented demonstrate the challenges in the current UxS solutions and provide solutions by leveraging on actual use cases to enhance on that, by leveraging on actual use cases of autonomous underwater drone and subsurface UxS system. The paper provides a framework and method for conducting operational impact assessments to improvise on the UxS solutions under different kinds of challenging environments. Through the development of UxS solutions through numerous experiments and demonstrations in Norway (Midtgaard and Nakjem 2016; Mathiassen et al. 2022; Nummedal 2021), co-evolution of the architecture and HAT designs was exemplified. Although the autonomy function and design approaches to human autonomy teaming (HAT) considerations weigh heaviest and are central to our conceptual exploration of analysis objects’ performance, much of the work is still in its infancy regarding sensors and scene analysis. HAT interaction types (LOA, MI, COAD) are evaluated over several contexts and/or operation types, iteratively for each object and assessed qualitatively on functions (controlling, sensing, etc.) where possibly sensor fusion assessment criteria are the most critical for autonomy functions: adjusted automation (Endsley 2023) and resource utilization by transition from studying human–automation interaction to human–autonomy teaming (Lyons et al. 2021).
We have presented research of Norwegian Defence Research Establishment (FFI) that explored limitations and possibilities with a common experimental approach independent of application with a common framework of autonomy adjusted for different UxS application (UAV, USVs, and AUVs). This enables FFI to explore several demonstrator systems at the same time. This ensures synergies with FFI’s partners and makes it uncomplicated to have a joint investment in autonomy in several application projects (UGV’s, AUV’s, and UAV’s).
Generally, a multi-UxS system increases capacity without increasing the number of UxS-operators. Autonomy makes it possible to control several vessels at the same time. Autonomy will redefine UxS operations. A multi-UxS system allows control of different types of vehicles. Due to the use case of underwater mine hunting, two overall criteria have so far been defined. One is related to size and weight (hereinafter referred to as weight class). The second is related to which payload/equipment the object carries (hereinafter referred to as equipment class). The hypotheses is that a combination of surface and sub-surface vehicles USV (unmanned surface vessel) and sub-surface UxS (AUV) can contribute to a future Norwegian naval mine countermeasure (NMCM) capability that includes unmanned, autonomous vessels. To gain this overall objective, we need to work further on autonomy functionality. On this background, the purpose of this article is to elucidate the following research question: How do different HAT designs (i.e., interaction design patterns according to Neerincx and Lindenberg, 2008) contribute to support the coordination of task under various environmental characteristics? While we have discussed this problem generically in prior work Stensrud et al., 2024; Human-autonomy teaming in high and low environmental complexity and dynamism, Stensrud et al. 2024; where we focused on a particular type of control: “bump-less” time-shift of authority during emergency response where it is not desirable to interrupt task resolution (Dess and Beard 1984), e.g., due to criticality of sustained performance), this study is diving into a practical naval domain problem of detecting and identifying mines. The functional analysis, so far, we have been focusing on autonomy (and different design approaches of human autonomy teaming interaction types) of a subsurface UxS as a main feature for future success in maritime operations. It is emphasized based on the extended functional analysis that one should rather say that it is best to choose an AUV model (subsurface UxS) that is just large enough to (1) carry suitable sensors (and any other payload) with the desired range and resolution; (2) provide the desired endurance with this sensor equipment, specified speed, and processing power; (3) be sufficiently stable and maneuverable to produce sensor data of good quality; and (4) have the necessary depth rating. Taking full advantage of the functionality in a larger campaign requires co-evolution of the technical architecture framework and HAT designs. Future experimentations with different HAT designs is thus recommended.
Acknowledgements
The authors would like to acknowledge colleagues at Norwegian Defence Research Establishment (FFI) participating in the Autonomy program at the Institute contributing to this research (e.g., Midtgaard & Nakjem, 2016; Mathiassen et al., 2022; Nummedal, 2021).
Declarations
Conflict of interest
The authors declare no competing interests.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
Abbink, DA; Carlson, T; Mulder, M; De Winter, JCF; Aminravan, F; Gibo, TL; Boer, ER. A topology of shared control systems—finding common ground in diversity. IEEE Trans Hum-Mach Syst; 2018; 48,
Ahram, T; Karwowski, W. Human systems integration modeling using systems modeling language. Proc Hum Factors and Ergon Soc Annu Meet; 2009; 53,
Allen, JE; Guinn, CI; Horvitz, E. Mixed-initiative interaction. IEEE Intell Syst; 1999; 14,
Bekier, M. Automation acceptance in air traffic management; 2013; UNSW Sydney: [DOI: https://dx.doi.org/10.26190/UNSWORKS/16253]
Bitar, G; Rundhovde, M; Ruud, E-LM; Sandrib, J; Selvåg, J; Smestad, R. Automatic underway AUV recovery using a USV (FFI report no. 22/01519; p 48); 2022; Norwegian Defence Research Establishment: https://ffi-publikasjoner.archive.knowledgearc.net/bitstream/handle/20.500.12242/3223/22-01519.pdf
Brehmer, B. Dynamic decision making: human control of complex systems. Acta Psychol; 1992; 81,
Brehmer, B. Command and control as design. 15th international C2 conference, track 1: concepts, theory, and policy; 2010;
Cabrall, CDD; Sheridan, TB; Prevot, T; De Winter, JCF; Happee, R. Karwowski, W; Ahram, T. The 4D LINT model of function allocation: spatial-temporal arrangement and levels of automation. Intelligent human systems integration; 2018; Springer International Publishing: pp. 29-34. [DOI: https://dx.doi.org/10.1007/978-3-319-73888-8_6]
Chanel, CPC; Roy, RN; Dehais, F; Drougard, N. Towards mixed-initiative human–robot interaction: assessment of discriminative physiological and behavioral features for performance prediction. Sensors; 2020; 20,
Chen, JYC. Fortino, G; Kaber, D; Nürnberger, A; Mendonça, D. Transparent communications for human–machine teaming. Handbook of human-machine systems; 2023; 1 Wiley: pp. 105-114. [DOI: https://dx.doi.org/10.1002/9781119863663.ch10]
Demir, M; Cooke, NJ; Amazeen, PG. A conceptual model of team dynamical behaviors and performance in human-autonomy teaming. Cogn Syst Res; 2018; 52, pp. 497-507. [DOI: https://dx.doi.org/10.1016/j.cogsys.2018.07.029]
Dess, GG; Beard, DW. Dimensions of organizational task environments. Adm Sci Q; 1984; 29,
Di Pasquale, D; Savill, M. The importance of coupling aerodynamic and cost analysis in aircraft design. CEAS Aeronaut J; 2022; 13,
Endsley, MR. Supporting human-AI teams:transparency, explainability, and situation awareness. Comput Hum Behav; 2023; 140, [DOI: https://dx.doi.org/10.1016/j.chb.2022.107574]
Fitts, PM; Jones, RE. Analysis of factors contributing to 460 “pilot error” experiences in operating aircraft controls; 1947; Dayton, OH, Aero medical laboratory, air material command, Wright-Patterson air force base:
Frame, ME; Boydstun, AS; Lopez, JS. Development of an autonomous manager for dynamic human-machine task allocation in operational surveillance. 2020 IEEE international conference on human-machine systems (ICHMS); 2020; pp. 1-4. [DOI: https://dx.doi.org/10.1109/ICHMS49158.2020.9209414]
Frazier, SN. Measuring cognitive workload in automated knowledge work environments (p 141) [A dissertation submitted to the Faculty of Purdue University in Partial Fulfillment of the requirements for the degree of doctor of philosophy]; 2022; Faculty of Purdue University:
Frey TL, Aguilar C, Engebretson K, Faulk D, Lenning LG (2018) F-35 information fusion. In: 2018 aviation technology, integration, and operations conference. Atlanta. https://doi.org/10.2514/6.2018-3520
Gombolay M, Bair A, Huang C, Shah J (2017) Computational design of mixed-initiative human–robot teaming that considers human factors: situational awareness, workload, and workflow preferences. Int J Robot Rese 36(5–7):597–617. https://doi.org/10.1177/0278364916688255
Goodrich, MA; Adams, JA; Scheutz, M. Arai, K. Autonomy reconsidered: towards developing multi-agent systems. Intelligent systems and applications; 2022; Springer International Publishing: pp. 573-592. [DOI: https://dx.doi.org/10.1007/978-3-030-82193-7_38]
Grote, G; Kolbe, M; Waller, MJ. The dual nature of adaptive coordination in teams: balancing demands for flexibility and stability. Organ Psychol Rev; 2018; 8,
Grover, J; Mohanty, N; Liu, C; Luo, W; Sycara, K. Noncooperative herding with control barrier functions: theory and experiments. 2022 IEEE. 61st conference on decision and control (CDC); 2022; pp. 80-86. [DOI: https://dx.doi.org/10.1109/CDC51059.2022.9992986]
Hamstra JW (ed) (2019) The F-35 lightning II: from concept to cockpit. American Institute of Aeronautics and Astronautics, Inc. https://doi.org/10.2514/4.105678
Hellesnes, J-I; Bjørnsgaard, T. A model for assessing the operational effect of different configurations of AUV (Publication Excluding public distribution) (External Report No. 2007_00622; p 23); 2007; NDRE (FFI):
Hopkins, D; Schwanen, T. Experimentation with vehicle automation. Transitions in energy efficiency and demand; 2018; Routledge:
Jiang, S; Arkin, RC. Mixed-initiative human-robot interaction: definition, taxonomy, and survey. 2015 IEEE international conference on systems, man, and cybernetics; 2015; pp. 954-961. [DOI: https://dx.doi.org/10.1109/SMC.2015.174]
Johnson, M; Bradshaw, JM. Lawless, WF; Llinas, J; Sofge, DA; Mittu, R. How interdependence explains the world of teamwork. Engineering artificially intelligent systems: A systems engineering approach to realizing synergistic capabilities; 2021; Springer International Publishing: pp. 122-146. [DOI: https://dx.doi.org/10.1007/978-3-030-89385-9_8]
Johnson, M; Bradshaw, JM; Feltovich, PJ. Tomorrow’s human–machine design tools: from levels of automation to interdependencies. J Cogn Eng Decis Making; 2018; 12,
Johnson, M; Bradshaw, JM; Feltovich, PJ; Hoffman, RR; Jonker, C; Riemsdijk, BV; Sierhuis, M. Beyond cooperative robotics: the central role of interdependence in coactive design. IEEE Intell Syst; 2011; 26,
Johnson, M; Bradshaw, JM; Feltovich, PJ; Jonker, CM; Van Riemsdijk, MB; Sierhuis, M. Coactive design: designing support for interdependence in joint activity. J Hum-Robot Interact; 2014; 3,
Kaber, DB. A conceptual framework of autonomous and automated agents. Theor Issues Ergon Sci; 2018; 19,
Kaber, DB. Issues in human–automation interaction modeling: presumptive aspects of frameworks of types and levels of automation. J Cogn Eng Decis Mak; 2018; 12,
Kalloniatis A, Kwok H, Oxenham M, Unewisse M (2020) One ring to rule them all, and through the headquarters bind them: AI, emergence, and the planning-execution-evaluation continuum in a fifth generation headquarters. 25thICCRTS, pp 1–22
Kirlik, A; Miller, RA; Jagacinski, RJ. Supervisory control in a dynamic and uncertain environment: a process model of skilled human-environment interaction. IEEE Trans Syst Man Cybern; 1993; 23,
Kouchaki, M; Okhuysen, GA; Waller, MJ; Tajeddin, G. The treatment of the relationship between groups and their environments: A review and critical examination of common assumptions in research. Group Organ Manag; 2012; 37,
Krogstad TR, Mathiassen K, Ruud ELM, Seehuus RA, Simonsen AS, Wiig MS (2020) HAL-A decisional autonomy module for unmanned systems. (Scientific no. 20/00896). FFI, NDRE - Norwegian Defence research establishment
Lindner S, Mund D, Schulte A (2022) How human-autonomy teams change the role of future fighter pilots: an experimental assessment. AIAA SCITECH 2022 Forum. AIAA SCITECH 2022 Forum, San Diego, CA & Virtual. https://doi.org/10.2514/6.2022-2551
Luciano, MM; Nahrgang, JD; Shropshire, C. Strategic leadership systems: viewing top management teams and boards of directors from a multiteam systems perspective. Acad Manag Rev; 2020; 45,
Lundberg, J; Johansson, BJE. A framework for describing interaction between human operators and autonomous, automated, and manual control systems. Cogn Tech Work; 2021; 23,
Lyons, JB; Sycara, K; Lewis, M; Capiola, A. Human–autonomy teaming: definitions, debates, and directions. Front Psychol; 2021; 12, [DOI: https://dx.doi.org/10.3389/fpsyg.2021.589585] 589585.
Mathiassen, K; Baksaas, M; Seehuus, R. LandX20 experiment report – experiment for future land warfare capabilities with focus on increased situational awareness and unmanned systems (Scientific no. 22/00274; pp 1–46); 2022; NDRE (FFI): https://www.ffi.no/publikasjoner/arkiv/landx20-experiment-report-experiment-for-future-land-warfare-capabilities-with-focus-on-increased-situational-awareness-and-unmanned-systems
McNeese, NJ; Demir, M; Cooke, NJ; Myers, C. Teaming with a synthetic teammate: insights into human-autonomy teaming. Human Factors: The Journal of the Human Factors and Ergonomics Society; 2018; 60,
Meilich A (2008) Integration of MBSE and HSI 35. https://ndia.dtic.mil/wp-content/uploads/2008/systems/6997meilich.pdf
Mercado, JE; Rupp, MA; Chen, JYC; Barnes, MJ; Barber, D; Procci, K. Intelligent agent transparency in human–agent teaming for multi-UxV Management. Human Factors: The Journal of the Human Factors and Ergonomics Society; 2016; 58,
Midtgaard Ø, Nakjem M (2016) Unmanned systems for stand-off underwater mine hunting. OMi, 7. At: Lillestrøm, Norway
Miller, CA; Parasuraman, R. Designing for flexible interaction between humans and automation: delegation interfaces for supervisory control. Human Factors: The Journal of the Human Factors and Ergonomics Society; 2007; 49,
Minos-Stensrud, M; Moen, HJF; Dyre Bjerknes, J. Information sharing in multi-agent search and task allocation problems. IEEE Symposium Series on Computational Intelligence (SSCI); 2021; 2021, pp. 1-7. [DOI: https://dx.doi.org/10.1109/SSCI50451.2021.9660121]
Mouloua M, Hancock PA, Ferraro J (2020) Human performance in automated and autonomous systems. In: Current theory and methods. CRC Press/Taylor & Francis Group
Myers, C; Ball, J; Cooke, N; Freiman, M; Caisse, M; Rodgers, S; Demir, M; McNeese, N. Autonomous intelligent agents for team training. IEEE Intell Syst; 2019; 34,
National Academies of Sciences, Engineering, and Medicine. Human-AI teaming: state-of-the-art and research needs; 2022; Washington, DC, The National Academies Press: [DOI: https://dx.doi.org/10.17226/26355]
Neerincx M, Lindenberg J (2008) Chapter 18: situated cognitive engineering for complex task environments. https://www.researchgate.net/profile/Mark-Neerincx/publication/287918596_Situated_cognitive_engineering_for_complex_task_environments/links/56a2394a08ae2afab886816c/Situated-cognitive-engineering-for-complex-task-environments.pdf
Nummedal OR (2021) Flamingo - a UAV for autonomy research (Experimental No. 21/00318). FFI-Norwegian Defence Research Establishment. http://hdl.handle.net/20.500.12242/2839
O’Neill J, McDonald J, Heaton B, Thotahewa K, Augur C, Brown K (2023) Applying a Mission-engineering thread approach using sense-decide-effect to the problem of a common C2 system for UxS. In: 29thICCRTS, Baltimore, US
O’Neill, T; McNeese, N; Barron, A; Schelble, B. Human–autonomy teaming: A review and analysis of the empirical literature. Human Factors: The Journal of the Human Factors and Ergonomics Society; 2022; 64,
Parasuraman, R; Sheridan, TB; Wickens, CD. A model for types and levels of human interaction with automation. IEEE Trans Syst Man Cybern Syst Hum; 2000; 30,
Park, S; Grosser, TJ; Roebuck, AA; Mathieu, JE. Understanding work teams from a network perspective: a review and future research directions. J Manag; 2020; 46,
Perkins, DG. Multi-domain battle: the advent of twenty-first century war. Mil Rev; 2017; 97,
Petousakis G, Chiou M, Nikolaou G, Stolkin R (2021) Human operator cognitive availability aware mixed-initiative control. https://doi.org/10.48550/ARXIV.2108.11885
Priebe M, Douglas C (2020) Multiple dilemmas: challenges and options for all-domain command and control. RAND Corporation. https://doi.org/10.7249/RRA381-1
Rebensky, S; Carmody, K; Ficke, C; Carroll, M; Bennett, W. Teammates instead of tools: the impacts of level of autonomy on Mission performance and human–agent teaming dynamics in multi-agent distributed teams. Front Robot AI; 2022; 9, [DOI: https://dx.doi.org/10.3389/frobt.2022.782134] 782134.
Rebensky, S; Carroll, M; Bennett, W; Hu, X. Impact of heads-up displays on small unmanned aircraft system operator situation awareness and performance: A simulated study. Int J Hum–Comput Int; 2022; 38,
Rico, R; Hinsz, VB; Davison, RB; Salas, E. Structural influences upon coordination and performance in multiteam systems. Hum Resour Manag Rev; 2018; 28,
Schmitt, F; Roth, G; Barber, D; Chen, J; Schulte, A. Karwowski, W; Ahram, T. Experimental validation of pilot situation awareness enhancement through transparency design of a scalable mixed-initiative mission planner. Intelligent human systems integration; 2018; Springer International Publishing: pp. 209-215. [DOI: https://dx.doi.org/10.1007/978-3-319-73888-8_33]
Schneider, A; Wickert, C; Marti, E. Reducing complexity by creating complexity: A systems theory perspective on how organizations respond to their environments. J Manag Stud; 2017; 54,
Schraagen, JM; Barnhoorn, JS; Van Schendel, J; Van Vught, W. Supporting teamwork in hybrid multi-team systems. Theor Issues Ergon Sci; 2022; 23,
Seeber, I; Bittner, E; Briggs, RO; De Vreede, T; De Vreede, G-J; Elkins, A; Maier, R; Merz, AB; Oeste-Reiß, S; Randrup, N; Schwabe, G; Söllner, M. Machines as teammates: A research agenda on AI in team collaboration. Inf Manag; 2020; 57,
Sheridan, TB; Ferrell, WR. Man-machine systems: information, control, and decision models of human performance; 1974;
Sheridan T, Verplank W, Brooks T (1978) Human and computer control of undersea Teleoperators. https://ntrs.nasa.gov/citations/19790007441
Simon, HA. Models of man, social and rational; 1957; Oxford, England, Wiley:
Stensrud R, Rutledal F, Bjørnsgaard T, Danjord F (2008) A proposal for a Norwegian Defence conceptual framework. 25th ISMOR, Hampshire, UK, 25–29 (August 2008)
Stensrud, R; Valaker, S. Methods to meet changes in the security environment a proposal of qualitative and quantitative assessment attributes for coordination performance. 20th ISCRAM conference; 2023; pp. 676-691. https://idl.iscram.org/files/stensrud/2023/2556_Stensrud+Valaker2023.pdf
Stensrud, R; Mikkelsen, B; Valaker, S. 5th generation fighter aircraft and future command and control: A note on some possible changes. 25th international command and control research and technology symposium; 2020;
Stensrud, R; Valaker, S; Haugen, T. Interdependence as an element of the design of a federated work process. 2020 IEEE international conference on human-machine systems (ICHMS); 2020; pp. 1-6. [DOI: https://dx.doi.org/10.1109/ICHMS49158.2020.9209423]
Stensrud, R; Mikkelsen, B; Betten, S; Valaker, S. A proposal for a simple evaluation method in support of the initial concept phase assessing a future unmanned loyal wingman for Royal Norwegian air Force (RNoAF); 2021;
Stensrud, R; Valaker, S; Haugen, R; Mikkelsen, B. Possible consequences for command and control of shifting coordination form: navigating the murky waters of adverse situations and complex missions. International command and control research and technology symposium; 2021;
Stensrud, R; Valaker, S; Haugen, T. Unified vision-series of military exercise - the methodological struggles in conducting evaluation of human factors and systems interaction. 14th international conference on applied human factors and ergonomics (AHFE 2023); 2023; [DOI: https://dx.doi.org/10.54941/ahfe1003605]
Stensrud R, Mikkelsen B, Valaker S (2024) Exploring human-autonomy teaming methods in challenging environments: the case of fighter pilots and loyal wingmen. Hum-Intell Syst Integr. https://doi.org/10.1007/s42454-024-00050-y
Sycara, K; Hughes, D; Li, H; Lewis, M; Lauharatanahirun, N. Adaptation in human-autonomy teamwork. IEEE international conference on human-machine systems (ICHMS); 2020; 2020, pp. 1-4. [DOI: https://dx.doi.org/10.1109/ICHMS49158.2020.9209410]
Synnes SAV, Hansen RE (2020) Norwegian Defence research establishment (NDRE) development of SAS micronavigation - innovations, public knowledge and potential patent infringements (publication excluding public distribution) (External note No. 20_02545; p 28). FFI (NDRE)
Vagia, M; Transeth, AA; Fjerdingen, SA. A literature review on the levels of automation during the years. What are the different taxonomies that have been proposed?. Appl Ergon; 2016; 53, pp. 190-202. [DOI: https://dx.doi.org/10.1016/j.apergo.2015.09.013]
Valaker S, Stensrud R (2020) The influence of harmonization levels on multi-domain operations effectiveness: the moderating roles of organizational environment and command and control variety. International Command Control and Technology Symposium. https://www.academia.edu/45184131/The_influence_of_harmonization_levels_on_multi_domain_operations_effectiveness_The_moderating_roles_of_organizational_environment_and_command_and_control_variety
Ven, AHVD; Delbecq, AL; Koenig, R. Determinants of coordination modes within organizations. Am Sociol Rev; 1976; 41,
Vught. Future hybrid human-robot teams: Functional requirements for Non-human Intelligent Collaborators (NICs) in an operational context; 2020;
Zabrodskyi, M; Watling, J; Danylyuk, O-V; Reynolds, N. Preliminary lessons in conventional warfighting from Russia’s invasion of Ukraine: February–July 2022, 30 November 2022, RUSI (special resources, p 69); 2022; RUSI: https://www.rusi.org/explore-our-research/publications/special-resources/preliminary-lessons-conventional-warfighting-russias-invasion-ukraine-february-july-2022
Copyright Springer Nature B.V. Dec 2024