A Network Centric Info-Structure for the Swedish Armed Forces

A Network Centric Info-Structure for the Swedish Armed Forces Klas Wallenius, CelsiusTech Systems AB klwa@celsiustech.se Summary The commonly accepted concept of Situational Awareness is discussed, where the knowledge of own and enemy resources and their relations to the environment is essential to make good decisions. In addition, the concept of Information Awareness is suggested, to measure the information and its value to the users. A network centric info-structure is proposed, with a Common Battle-Space Model that is supporting situational awareness for all decision-makers in the organisation. The common model includes information regarding the situation, the environment and the decisions. The difficulties in designing such a model include most of all the need for information control, for which Information Awareness is essential. The Saab solutions to design essential parts of the Common Battles-Space Model include the Wide Area Situation Picture, WASP, and the Multi- Sensor Tracker, MST. Yet a substantial research effort remains to complete the battle-space model. The development of a common planning system to maintain and evaluate decision alternatives is particularly important. 1. Introduction A Revolution in Military Affairs, RMA, is the label often used to indicate the vast changes that are foreseen in modern armed forces around the world. [5] The revolution will enforce completely new doctrines and organisations for warfare. New ways of running business in the commercial sector, exploiting the fast development of information technology, will influence this revolution. The concepts of warfare leveraged by the new technology are sometimes described as Network Centric Warfare [1]. The network centric view, as opposed to the traditional platform centric view, will imply that information obtained somewhere in the organisation can be shared by anyone else that is connected to the network. This enables decentralised decision making and a reduced number of levels in the management hierarchies. This will, in turn, give much faster reactions to events in the battle-space. The Swedish Armed Forces has extensive intentions to develop an RMA concept [5]. One effort in this direction is the research on a Mobile Joint Command and Control Function, with the Swedish acronym ROLF [10]. The IQ Ledning programme aims at proposing an info-structure that is supporting such a ROLF unit. It is a joint study programme between Saab AB, The National Defence College of Sweden (FHS), the Defence Research Establishment (FOA), the Swedish Defence Materiel Administration (FMV), and the Royal Institute of Technology (KTH). Another study programme is performed in cooperation between SaabTech Systems AB and FMV, to find methods and concepts for a Wide Area Situation Picture, WASP. The views presented in this article are mainly based on experiences from the IQ Ledning and the WASP programmes. However, these are only two of the activities in the RMA arena that SaabTech Systems AB takes part in. 2. Situational Awareness and Information Awareness The concept of Situational Awareness usually includes the states of own and enemy forces, the environment, and the relationship between the forces and the environment [4]. There is a decision scheme for officers in the Swedish Armed Forces [6]. This scheme requires that the purpose and goals of the task and knowledge of own and enemy forces are considered to maintain the decision alternatives. The knowledge of the forces includes a comparison of the capacity under present terrain and weather conditions. By considering these aspects, operators will make their decisions based on a well-founded awareness of how to optimally exploit their own However, not only weapons and forces should be included among the resources that the operators have to be aware of. As the achievement of dominant battle-space awareness will take a more vital part of the operations, decisions concerning how to best use the information resources (such as sensors, intelligence, data links and even the 220

capacity of the decision-makers) will gain in importance. Everything cannot be known at a given moment since there is a cost for using the resources needed to obtain the information. At the same time the difference between knowing something, and not knowing something, may be essential for the outcome of the operation. To maximise the total benefit of the information resources, a means of measuring the usefulness of the information is required. The decision-makers should always be aware of to what extent they can trust the information, and what information they have, compared to what information they need in their current assignments. They should also (somehow) be aware of how they could benefit by using more of the information Thus we state that Information Awareness must be included to give full situational awareness. In older systems, Information Awareness came with no or little computer-based support. The data presented usually originated from one source (e.g. radar) or a few similar sources only. The operators then learnt how to trust the data by intuition and experience, and also by some rather primitive quality numbers. In future systems, the origin of the data will be so complex that intuition will be of little or no help anymore. Computer-based support to achieve better Information Awareness will thus grow in importance. Such support in its simplest manner is performed independently of the intended usage of the information. Full support for Information Awareness, though, will also require a model of the decisions to be considered, and will thus be much harder to achieve [8, 9]. In [2], three measures are defined that, if presented together with the information, will give a larger degree of Information Awareness to the users: 1. Precision to denote measures of the "correctness" of data, 2. Quality to denote its fitness for purpose, 3. Utility to denote the expected benefit for the use. Of these three measures, only 'quality' and 'utility' depend upon the purpose to which the information is applied. 3. A Network Centric Info-Structure the Common Battle-Space Model Traditional systems of systems for military organisations have been platform centric. The communication bandwidth between the units (command and control centres, aircraft, tanks, etc.) has been very limited compared to the communication within each unit. The information flows between different units have been explicitly defined and thus very inflexible. The information exchange in such a platform centric info-structure is naturally message based, i.e. certain message formats are predefined for different kinds of reports and orders. The different units maintain local models of the assumed situation in the battle-space. As a result of the difficulty that they have had sharing information, different units have historically used their own sensors and data processing capabilities to maintain their own local models of the battle-space. There has been little or no means to keep these models consistent between the platforms. By use of modern network technology, communication bandwidth now has been much more achievable. At the same time, standardised communication protocols have made it considerably cheaper to offer services across all units connected to a network. Thus, the view can be changed from defining information flows between the platforms to, instead, defining the common information model available to all units. Force Coordination Other Organisations Sensors Force Control The Common Battle-Space Model: Decisions Environment Situation Intelligence Weapons Control Other infosources Data Fusion Figure 1. The suggested info-structure a common model of the battle-space made available for the different levels of decisionmakers. We sugges t such a common model, the Common Battle-Space Model (see Figure 1), which will include all aspects of situational awareness for all users. In other words: the model should include pieces of information relevant to all decisionmakers on all different levels in the military organisation. The implementation of this model would imply a tremendous improvement of decision performance. Better and faster situational awareness together with a larger flexibility would lead to a much better utilisation of available Referring to the definition of Situational Awareness, we see three major categories of 221

information to be shared by the decision-makers in the Common Battle-Space Model: 1. Situation. Estimated states of objects in the battle-space. The states include identity, type, kinematics, and logistic status of the objects. 2. Environment. Information on phenomena that cannot be affected by own or enemy actions, e.g. weather, terrain, and doctrines. 3. Decisions. Purpose and goals for the missions, together with both rejected and selected decision alternatives ( plans ). The decision alternatives describe how available resources can be exploited to accomplish the purpose of the missions. For each alternative an estimate of the expected outcome is required in terms of utility for the organisation compared to the costs for resource exploitation. One commander s decision to exploit resources, to achieve his goals, will of course define new goals for the commanders of those The mission definitions and the decision alternatives are thus included in the same category of information, which will require a common planning service, available to all the users. 4. New difficulties Clearly, the network centric approach presents new difficulties for the designers of the info-structure since there are very different requirements on the information for different decision-makers. At the lowest control levels, the decision-makers have very specialised assignments. They need detailed information, with good timeliness and accuracy. At the same time they are very often mobile, leading to limited communication capacity. At higher levels there is a tendency for generality and a demand for overview. The decisions may be seen more as a matter of resource allocation rather than controlling different weapons and platforms. At these higher levels, the focus may be on other, non real-time, information, e.g. information on political, international law and treaty, and economical issues. A key difficulty to overcome in designing a network centric info-structure will be how to keep the battle-space model relevant to each of the different classes of users, given their substantially various information requirements. We propose the usage of hierarchical aggregation as a partial solution. By arranging objects in groups (and in groups of groups), the users could easily select the level of abstraction that suits their current assignments. The design must also accommodate many contributors of information to the common battlespace model, any of whom may contribute information which is inconsistent with other information, whilst maintaining model consistency. Maintaining information security will be another key concern, especially when there are information exchanges with other organisations: media, civil organisations, and allied forces (and maybe even with the enemy to apply pressure during negotiations). Thus it must be possible to make decisions on the access to the information for the different users. The competing demands of precision versus limited bandwidth need to be considered. The data streams must be controlled, which will limit the precision available for some users. This will be of particular importance to users connected to the network via low bandwidth data links. There are also competing demands of real time versus non-real time data due to the nature of the data types - non real-time data tend to come in larger chunks, congesting the network in its narrow parts. Real-time data streams must then be given higher priority to meet their timeliness requirements. Furthermore, the info-structure designed for warfare also must anticipate that C2 centres and sensors may be captured or destroyed and that the network occasionally may be cut off due to attacks and jamming performed by the enemy. It should still be possible to make full use of the remaining resources to assemble a battle-space model of as high quality as possible. Emphasis on co-ordinating the collection and dissemination of information will increase, since sensors and intelligence services will be limited resources available to the competing needs of decision-makers in the organisation. Further issues include the limited capacity of the decision-makers themselves. Methods to compile and present the information to the decision-makers, suitable to their current tasks and work-pressure, constitute an essential part of the research on ROLF [10]. Altogether there must be a control of the information collection and the information flows in the system, to make optimal use of available resources, and to meet security requirements. The emphasis on information control will be much larger than before. Since the information by default is shared in a network centric info-structure, it is the exceptions from sharing information that must be managed. In platform centric systems, the information control is built in, by the fact that all information flows are defined explicitly in hardware and software. Information control will require methods for data reduction, bandwidth regulation, data filtering, managing of public keys and resource management. The usage of these methods should be kept simple. Metaphors have to be developed for bandwidth consumption and 222

information classification to make the problems understandable and solvable for decision-makers. Common to all problems of information control is, however, the need for information awareness, as defined in Section 2. There has to be an understanding of the usefulness of information and the possibilities to achieve better information. 5. WASP the Wide Area Situation Picture SaabTech Systems has developed the methods needed to meet the requirements for a common description of the situation in the battle-space. WASP the Wide Area Situation Picture is truly network centric, and is designed to Give information on moving objects for different users, with common target numbers. Be robust to attacks, jamming and technical disturbances, Fully utilise currently available information resources: sensors, C2 centres, data links, and data fusion capabilities (all from many different suppliers). Always indicate the resulting data precision due to the currently available information Be fully scalable, allowing for thousands of both users and contributors. Be easy to integrate, even with existing systems and data links. The achievement of the WASP is performed on two levels. Data from sensors and human reporters are entered into the system on the sensor data level. The active and passive sensors may be of very different kind, giving measurements of one up to three or even more dimensions. A measurement could be a one-dimensional bearing to the target or a complex data record including bearing, range, elevation and Doppler information. A measurement may also give indications that can be used to estimate the object s type or identity. Data fusion is performed to combine these sensor data to estimate what objects there are, and where they are in the battle space. SaabTech Systems AB has the solution for this - the Multi-Sensor Tracker (MST), with the future add-on of automatic type estimation ( MST+). The MST is capable to utilise many different sensors from any supplier as long as there is output on the sensor data level. Theoretically, data fusion is best performed when data from as many sensors as possible are being used. In some cases it is even necessary to feed data from more than one sensor to the tracker, for instance when only passive sensors are used. Thus it would seem appropriate to have only one data fusion node in the network, using data distributed from all available sensors. There are however several reasons why data fusion on the sensor data level will be performed at more than one place and why there has to be means to combine data also on the track level. First of all, the usage of one single MST in the entire system would imply a very vulnerable solution. Redundancy will be needed to meet the requirements for robustness. Secondly, from some sensors there can only be track data distributed, because of the construction of the sensor or because of the available bandwidth. Thirdly, the connection of centres and systems belonging to other organisations will give tracks in most cases. The WASP Correlator Unit (WCU), is used to perform the required track correlation. The participating track sources (C2 units, data fusion nodes and networks belonging to other organisations) are connected to the network, and to each other, via the WCUs. All software necessary to establish the WASP is encapsulated in these WCUs. Thus, the adaptation effort for each type of track source is kept at a minimum, see Figure 2. Figure 2. The WASP network. The local track data is correlated to the global track data, received from other WCUs (via the network), to determine which tracks that correspond to the same real objects. This correlation process is fully automatic, although it will be possible to perform manual interaction. Special care is taken in the WCUs to estimate the precision of the tracks. Local data that is either unique or of better precision than the global data, is reported to the other WCUs. In this way the network will distribute only the best track data that is available for each object. Finally, the actual Wide Area Situation Picture is assembled for presentation to the local operators. 223

To reduce bandwidth consumption, the total surveillance area is divided into smaller subscription areas. For each such area there are several bandwidth levels that could be subscribed to by using multicast services in the network. On the lowest bandwidth level, information on all objects in the subscription area is reported, although at a rather low update rate. Higher accuracy is achieved by subscribing to higher levels, thus requiring more bandwidth in the network. The WASP concept will be fully scalable, due to the selected correlation algorithm and the use of multicast. Thousands of C2 units can be connected and there will be no limit on the total surveillance area. The survivability of the WASP will be outstanding. Although sensors, data links and C2 units may come and go, each operator will still achieve the best quality and consistency from the WASP. 6. Conclusions The Common Battles-Space Model will increase the decision performance dramatically. The combination of MST+ and WASP has the capacity to take a substantial part of the implementation of the model. These concepts can make flexible use of very different kinds of data sources and they already exist on the prototype level. By the bandwidth management, the subscription for data, and the careful estimation of data precision, we already are on the way to support Information Awareness. [3] P-O Fjällström, G. Neider, M. Persson, T. Risch, and P. Svensson. Architecture principels for information superiority in future command and control systems (in Swedish). Technical Report FOA-R-00-01435-505-DE, ISSN 1104-9154, Defence Research Establishment, FOA (Sweden), 2000. [4] V. Gawron, Situational Awareness (Lecture notes), H. Silver and Associates, UK, 1997. [5] HKV. RMA A new foundation for defense forces development (in Swedis h). Technical Report HKV 09 100:63046, Swedish National Defence, 1999. [6] HKV. StabsR 1 Fu (in Swedish), Swedish National Defence, 1996. [7] L. Jönsson, G. Neider, J. Schubert, P.Svensson. Information Fusion for the Tactical Intelligence Process. An Informal Introduction to the Key Technology for Dominant Battlespace Awareness (in Swedish). Technical Report FOA-R 98-00902-505 SE, ISSN 1104-9154, Defence Research Establishment, FOA (Sweden), 2000. [8] H. Raiffa. Decision Analysis. Reading: Addison-Wesley, 1970. [9] L. Seligman, P. Lehner, K. Smith, C. Elsaesser, and D. Mattox, Decision-centric information monitoring, Journal of Intelligent Information Systems, 14(1), March, 2000. [10] C. Sundin, and H. Friman. Rolf 2010 a mobile joint command and control concept. Technical Report ISBN 91-87136-33-3. ISSN 1403-2120, Swedish Defence College, 1998. Yet there is a large amount of research remaining to make the model work in its full extent. Especially the service required for maintaining decision alternatives and mission definitions need certain care. Such a planning tool is essential for effective resource exploitation. Acknowledgments The views here presented are based on discussions with the people in the IQ Ledning and WASP programmes, and with Daniel Wängelin, SaabTech Systems AB. Thank you all for these stimulating moments. References [1] D.S. Alberts, J.J. Gartska, and F.P. Stein. Network Centric Warfare. U.S. D.o.D., 1999. [2] S. Arnborg, H. Artman, K. Wallenius, Information Awareness in Command and Control: Precision, Quality, Utility. 3 rd International Conference on Information Fusion, Paris 2000. Klas Wallenius, M.Sc. E.E., has been working with the data fusion group at SaabTech Systems AB since 1993. He is currently the technical manager for the WASP development program. 224