We have seen, in previous article, what a simulation model is and how, over time, the practice of modeling and simulation has gained increasing importance in various contexts. Let us now focus on the current use of modeling & simulations in the defense world.
Why are simulations used in the defense sector
We begin by considering the purposes of the simulation. In a nutshell we can say that today we essentially simulate to achieve one of these 4 goals:
A) The training of military personnel
B) Research and development of winning strategies/tactics in known scenarios
C) The definition of the characteristics of new systems
D) The validation of new systems
Let's see below ("figure 1") what the 4 dimensions of use listed above consist of. And to do this, let's consider a generic simulation, that of an air defense system like the one in the diagram (see following image). It is a very simplistic schematization that provides a model of a command and control (the first on the left, in the figure), the model of a missile launcher, a radar, an interceptor missile and a possible threat (the last blue square in the figure, on the far right). These models of real objects are connected, in the simulation, by a communication infrastructure (the orange block in the figure) capable of making them communicate with each other in order to build a single coherent simulation of an interaction of the defense system with the air threat.
To understand how a simulation is used to achieve goal A, thestaff training1, we can imagine the need to have an operator of the weapon system modeled in the example exercise (see following image "figure 2"). Let's imagine it interacting directly with the command and control model, let's imagine that the simulator is fed with the typical simulation scenarios of an engagement of a hostile threat and let's assume that the system allows to record the operator's interactions and the outcomes of the engagement mission for a subsequent post-analysis of the operator's behavior (how many hostile planes did he manage to engage? With what result? Which operator was better for a given set of scenarios? Why?). Internationally this goal is known as Education, Training, Exercise and Evaluation (ETEE).
To understand how a simulation is used to achieve objective B, research and development of winning strategies/tactics in known scenarios, we have to imagine that we have intelligence data that allows us to model a new threat model. It is legitimate to ask how, other considerations being equal, it is appropriate to configure the system to respond better, that is, to be more effective, for this type of threat: how to arrange the launchers? How to arrange the radars? Which fire doctrine is more effective? What is the most correct behavior of the operators? A similar situation is schematized in the following figure ("figure 3").
To understand why a simulation is used to achieve objective C, i.e. la definition of the characteristics of new systems, we can again refer to a scenario like the one in the previous image ("figure 3"). Suppose we become aware, through intelligence data, of a new threat from a potential enemy. We carry out simulations with the models of the weapon systems at our disposal and we discover that the system at our disposal is not suitable for assuring us a satisfactory degree of protection from the new menace. Being able to procure ourselves a new weapon system, or being able to evolve the existing one, in which direction do we move? Can a more performing radar be enough, capable of seeing the threat earlier and better, or a more maneuverable interceptor missile, or capable of reaching increased heights/distances compared to the existing one? To answer these questions, and then elaborate a request for an offer for an advanced and optimized system to respond to the new needs of the armed force, simulation is a very useful and often irreplaceable tool.
And finally, let's try to imagine how a simulation system can be useful for achieving objective D, i.e. la validation of new systems. Continuing the previous example, we can try to imagine that we have actually received an improved system (or subsystem) in order to counter a new type of threat. Assuming that the system we have decided to acquire is a new radar, we can imagine that, in addition to the radar itself, a model of the radar can also be provided, a model suitable for being part of our simulation2. With this model it is possible to re-run the simulation to verify the performance of the final improved system and then validate the system itself in order to ascertain the effective compliance with the need to face the evolved threat (see following image "figure 4").
Another case that falls under objective D is that of the integration of a complex system which requires development from scratch. Usually, when designing a completely new system, we start by designing a global simulation of the system itself aimed at fine-tuning the detailed requirements and developing the algorithms of each subsystem (exactly the situation of the first scheme of the article). Once, completed the development, the various subsystems will have actually been created with the real HW (at least in prototype form), these prototypes are replaced, first one at a time (see following image "figure 5"), then gradually all, to the subsystems originally conceived in simulation. In other words, the simulation is transformed into a real test, validation and integration tool useful for verifying how far the created subsystem deviates from the one originally conceived (concept of HW in the loop).
Simulation types
There are many ways to differentiate the types of simulations used in Defense. A classic differentiation is between Constructive, Virtual and Live simulations:
- Constructive: a simulation in which both systems and the behavior of unit operators (operators whose behavior is represented by models) are simulated3. The simulation system like the one in "figure 3" can respond to this type of definition, if one imagines that human intervention limits itself to modifying what has been identified as "configuration data" (for example by deciding the deployment of the elements on the ground or to decide the appearance at a certain moment during the simulation of a threat that has a pre-established behaviour). In a constructive simulation everything is actually simulated (usually by computer), and no real element interacts with the simulation.
- Virtual: In this case we have simulated systems controlled by real people (Human in the Loop). What is described in "figure 2" can be a good example.
- Live: In this case, in addition to the Virtual case, we can also have real systems (as already mentioned, the concept is of HW in the loop). Consider, for example, the case of "figure 6" (following image) in which instead of a simulated radar, we have a real radar capable of sending data to the simulated command and control, and enabled to receive both real and simulate (Sim over live).
The classification just exposed has a historical value (it appears on the first documents of the DOD USA from the early 90s) but today it appears, in many ways, outdated. However, some considerations on the various configurations can be made.
A constructive simulation certainly ensures more repeatable results, is relatively cheaper than the others (there are no real objects, everything is simulated on computers) and is more suitable for studying the theoretical behavior of the modeled system (for example: definition of the requirements, definition of strategies, assessment of weaknesses). Fidelity depends on the accuracy of the models used in the simulation itself.
But as non-simulated elements are introduced (ie real operators, as in the case of virtual simulation; or real systems in live simulation) the repeatability of the simulation is more difficult, if not outright impossible to achieve. But representativeness, on the other hand, improves and makes the simulation more suitable for different objectives, such as those of validating a new system or training operators.
We also briefly mention the technology of digital twins, literally "digital twins", which consists in creating virtual and digital copies of real objects, copies capable of suitably modeling the static and dynamic aspects. Such digital copies are capable of influencing the simulation in which they act, of course, but also of transferring some effects/results of the simulation onto the real object of which they are copies. It is a sort of alternative to the situation in "figure 6": instead of making the simulation interact with a real object leaving it in the real world, the real object is transposed into the simulated world by replacing it with its "digital twin".
Finally, it is necessary to reflect how, with the advance of the latest technologies and with the metaverse at the gates, the representation of "figure 7" (following image) is fitting, which shows, alongside the traditional classification (lower part of the figure), the overlapping of concepts of Mixed Reality which we have to get used to more and more (for a discussion on the perspectives of the metaverse it is useful to read the study Ref. 2) .
Figure 1 : Comparison between the concepts of Mixed reality (section a) and simulations (section b). Source: Ref. 3
Also of interest is a classification that differentiates simulations based on the hierarchical category of the simulated systems. "Figure 8" (following image, taken from Ref. 3) synthetically represents a hierarchy of the various types of simulation, highlighting the levels of aggregation, the resolution/fidelity, the area of simulated action (and show, for example, some known simulations in the US defense field).
Our example of "figure 1" would be placed at the penultimate level of aggregation, the "Mission" one. At the lowest level (engagement) one can imagine the model of a single interceptor, or of a radar. At an even lower level (engineering) there are the subsystem models (for example: the model of a seeker4 of a missile).
The levels positioned further down, in "figure 8", are more suited to achieving objectives C and D (definition of requirements and validation) and for this reason they are called level “Engineering”.
The "figure 8" introduces us, in a natural way, to address the issue of interoperability. Interoperability is a broad concept, and expresses the ability of a system to interact correctly with other systems. Interoperability has long become a requirement that weapon systems must meet, and it is a prerequisite for network enabled capabilities and multi-domain operations. If on the one hand the weapon systems must interact with each other, it is equally clear, even simply by looking at "figure 8", that the simulations (or rather: the models on which the simulations are based) must also do so, both for the need for aggregation of different models in a simulation that considers different simulated systems that interact, both for the need for model reuse5.
An example that is easy to understand is that of the modeling of the objects that animate the battle space: if a system sees a certain threat, a system deployed nearby must see the same threat (unless the different characteristics of the sensors).
From this consideration we understand that the need arises to define a) a language for the exchange of simulated data and b) a SW architecture which allows the various simulations to interoperate. To solve this type of problems, such as those of federating objects of a different hierarchical level, we can now consider both the DIS (Distributed Interactive Simulation, the IEEE standard for conducting simulations across platforms) that the HLA (High Level Architecture, which is the architectural standard for distributed simulation design).
(Continued)
Read the first part "What are simulation models: origin and evolution"
Read the third part "What are simulation models: simulation centers in Italy and around the world"
References
1 INTRODUCTION TO MODELING AND SIMULATION, Anu Maria
2 Metaverse and National Security, Italian Institute of Strategic Studies, Fabio Vanorio
3 Open Challenges In Building Combat Simulation Systems To Support Test, Analysis And Training, 2018 Winter Simulation Conference (WSC), Andreas Tolk - Raymond R. Hill - Douglas D. Hodson - Jeremy R. Millar
Footnotes
1 A similar concept is that defined in English by the term "Mission Rehearsal", i.e. proof of the mission
2 Different versions of the model are often commissioned from the supplier, intended to reflect the progress of the subsystem design. In this way it is possible to have intermediate returns, before the final project is released, so as to be able to verify the validity of the requested evolutions and, if appropriate, to be able to correct them in the design phase of the new product, when the cost of the modifications is still reasonable.
3 The behavior of the commanders assigned to exercise the command and control function of the dependent units is not simulated
4 The target tracking subsystem
5 It is however necessary to be cautious about the reuse of simulation models, and warn the reader that it is not always possible to reuse models created for two simulations of different hierarchical levels, or even of the same hierarchical level. There are fidelity level issues, execution requirements, and more, which often make reuse impossible. This is solved either by simplifying the model, or by redesigning it, or even simply by using the data from the more accurate simulation in the context of a higher-level simulation.
Photo: US Air Force / author
