Wargaming is an indispensable and versatile instrument for the Swiss Armed Forces, which combines several advantages. Like every method however, wargaming also has certain limitations. War games are thus driven by the decisions of their players. Each game is unique and not reproducible. In addition, the number of games that are played by human players is severely limited. Statistical statements on a particular game outcome are thus not possible. Last but not least, the outcome of a wargame, like in a game of chess, for example, is heavily dependent on the ability of the players. As part of this project, we develop and test algorithms of artificial intelligence in order to overcome existing limitations. Using reinforcement learning, we analyse digital wargames completely automatically with the objective of supporting decision-makers and planners, developing new tactics and operating procedures and find out which combinations of military skills are advantageous and which requirements are critical for systems to be procured.

The latest developments in the field of artificial intelligence and in particular reinforcement learning offer new possibilities in the education and training of soldiers and commanders. As part of this project and in cooperation with the Dalle Molle Institute for Artificial Intelligence Research (IDSIA), we examine the use of reinforcement learning to increase efficiency in the tactical training of fighter pilots. Here, we develop and test intelligent algorithms which can control a wide range of enemy aircrafts in flight simulations in order to optimally challenge the trainees (Red Teaming). The focus in this project is on the collaborative cooperation of several systems as well as the explainability of successful tactics.

In cooperation with the armasuisse S+T Test Center, we are developing vulnerability models of various systems for the Swiss Army and Air Force. These models enable us to simulate the interaction between various types of effectors (missiles, projectiles, etc) and targets (planes and ground vehicles, etc.) and thus to analyse it precisely. Using these models eliminates to a certain extent time-consuming, expensive tests, while on the other hand, its integration in simulators can enable more realistic and effective training. 

In preparation for the procurement of the new local ground-based air defence, we support the Swiss Air Force by means of simulations in the development of new tactical operation concepts. Here, we simulate various different attack and defence strategies in high-resolution Swiss terrain in order to be able to assess them in terms of their efficiency. Simulations of sensor coverages in the terrain, critical flight paths of low-flying threats and location optimisations of radar systems also enable us to derive initial critical requirements and to minimise risks for the subsequent procurement at an early stage.

SimCo-19 is an application for decision support to fight Covid-19 pandemic and is based on a synthetic demographic model of Switzerland, also known as a digital twin. The Swiss population is replicated by composition, in other words, age, gender, place of residence, etc., without referring to specific people. Statistically speaking, no differences can be identified between the real data and this synthetic data. The advantage of this method lies in the fact that human behaviour of a large group of people can be simulated precisely and at low cost. SimCo-19 thus helps crisis units to recognise medical bottlenecks early on and to test the efficiency of various countermeasures before they are implemented.

OPTIMA is a constructive multi-agent simulation platform for supporting the Swiss Armed Forces in the field of capability development and mission support. OPTIMA enables complex concepts and scenarios to be explored at  strategic, operative and tactical level under a variety of conditions. Due to the high performance, OPTIMA enables statistical analyses and robust statements based on a very high number of simulation runs (Data Farming). Results and the knowledge gained are subsequently used by the Swiss Armed Forces to develop new or optimize existing capabilities. OPTIMA is based completely on open source software and has been developed together with research partners. It will be continuously developed according to the needs of the Swiss Armed Forces.

In cooperation with the Military Aviation Authority (MAA), we are developing a software solution which will enable the Swiss airspace to be statistically analysed with regard to density, type and behaviour of the air transport as well as to identify risks of near collisions. The developed software supports the MAA and the Federal Office of Civil Aviation (FOCA) in certifying and integrating the reconnaissance drone system ADS15 into the Swiss airspace. In addition, it enables the development of an algorithm for avoiding collisions, which is indispensable for the complete integration of the ADS15. In future, SAoSat aims to enable risk assessments of relevant flight routes and operations for any civil and military aircraft.

Today’s simulation landscape within the Swiss Armed Forces consists of a variety of highly-specific systems. These proprietary applications can neither be networked nor can cost- and time-saving development synergies be used. On behalf of Training and Education Command, in this project we are exploring novel simulation applications which will enable increasing networking and a high degree of flexibility. All this enables the simulation of joint and multi-domain scenarios, which can be executed via a network independent of location. We test these novel simulation environments with regard to their suitability in terms of capability development , mission support as well as education and training. The knowledge gained will flow directly into the simulation concept of the Armed Forces and enable an assessment of these technologies as early as the conceptual phase.