1 Towards a Global Partial Operational Based on Qualitative Spatial Reasoning Zakaria Maamar Driss Kettani zakaria.maamar@drev.dnd.ca driss.kettani@drev.dnd.ca Defence Research Establishment Valcartier 2459 Pie-XI Blvd North, Val-Bélair QC G3J 1X5, Canada & Research Center in Geomatics and Computer Science Department Laval University, Ste-Foy QC G1K 7P4, Canada 1. Overview In the Global Partial Operational (GPOP) project, we aim at providing a framework in which the Canadian Forces (CFs), namely Land (L), Air (A), and Maritime (M), could collaborate despite their individual differences, such as mission types. To reach this objective, a facility that supports the interoperability of these forces Command & Control Information Systems (CCISs) is set up. A CCIS plays a crucial role in the military field. In a battlefield, for example, a commander takes decisions concerning his troops' deployments and operations using the information that is provided by the CCIS. In order to achieve CCISs interoperability, a set of intelligent components, called Software Agents (SAs) [1], are introduced. For instance, SAs could act on behalf of the military forces, by satisfying their needs and monitoring the events that could interest them. The main motivation behind the development of an interoperable environment for systems, in general, and CCISs, in particular, is to exchange services without dealing with distribution and heterogeneity constraints. An example of services could be to request the weather from a partner in order to update its operational picture. In the GPOP project, it occurs that LFs require information from AFs about the enemy as well as allied positions, before these LFs commit themselves to a risky operation. Currently, each military force, i.e. L, A, and M, has its own picture of the situation in which it is involved. In the GPOP picture, this picture is called partial. However, in order to have a complete assessment of the situation the partial picture has to be enhanced and enriched with external information that could be obtained from other partners. Therefore, the partial picture evolves into a global picture. In this paper, we focus on the issue of how a military user, could rely on new advanced qualitative spatial reasoning techniques, in order to use a global partial operational picture. Better than quantitative reasoning [2], qualitative techniques comply with the human perception of space and address efficiently the issues of incertitude and inaccuracy. The remainder of the paper is organized as follows. Section 2 presents the GPOP project. Section 3 describes a qualitative reasoning approach that is currently under investigation in the GPOP project. Finally, Section 4 summarizes the paper. 2. Presentation of the GPOP project 1
2 2.1 What is a CCIS? Nowadays, information technologies are an inherent part of the commanders' decisionmaking process. Particularly, CCISs help commanders to obtain a view of the situation in which they are involved. A CCIS consists of a structure, tasks, and functions. The CCIS structure presents an assembly of facilities that are arranged to meet the CCIS's objectives. To reach these objectives, the CCIS's functions are initiated in order to carry out the needed tasks. Tasks require structure facilities, in terms of personal, technical equipment, etc. Figure 1 presents a simplified architecture of a CCIS. Several types of functions are offered to users, ranging from planning and weather forecast to logistic. These functions are built on top of a support structure, in terms of hardware (e.g. PC stations) and software (e.g. DataBase Management System) resources. Furthermore, certain functions of a CCIS receive formatted messages from the external environment, e.g. sensors, through a communication module that parses these messages. For example, the air monitoring function receives messages from radars as well as from patrol planes. Next, this function uses these messages to automatically update the appropriate databases. CCIS Users Communication Planning Logistic Resources (software,hardware) Functions Structure External environment Figure 1 CCIS Simplified Architecture 2.1 CCIS Agentification Agentification aims at allowing a system to behave like an agent. Therefore, this system could take advantages of the agent's main characteristics, such as autonomy and sociability. Furthermore, this system could join a multiagent environment. The approach that is proposed to agentification is to build an agent on top of a system. In the GPOP project, to agentify a CCIS we introduced an agent, called CCIS-Agent (Figure 2). Similar to a Wrapper-Agent, a CCIS-Agent is the front-end of the CCIS to communication networks, acts on its behalf, and maintains its autonomy. In addition to carrying out these different activities, the CCIS-Agent advertises through its services the different functions the CCIS performs. A service could be initiating the CCIS's weatherforecast function. As stated previously, a CCIS offers different types of functions to users. Generally, these functions are very complex, for instance a planning function could be a distributed-objects client/server application that runs on top of an Object Request Broker middleware. Because of this complexity and the fault-tolerance and efficiency criteria that a CCIS should meet, new types of SAs, called Function-Agents are introduced at the CCIS-Agent level. Therefore, each Function-Agent corresponds to a CCIS's function. As a result, the CCIS-Agent manages and monitors a group of Function-Agents (Figure 2). For instance, a 2
3 request to the planning function of a CCIS is initially sent to the CCIS-Agent that forwards this request to the appropriate Function-Agent. External Environment CCIS-Agent Requests/Answers forwards Function-Agent 1 Function 1 CCIS Function-Agent i Function i Figure 2 CCIS Agentification 2.3 GPOP environment Figure 3 illustrates the GPOP environment. It consists of three parts that are related through a communication middleware, such as the Common Object Request Broker Architecture (CORBA). Each part corresponds to a specific military force, namely L, A, and M. Furthermore, each military force relies on its CCIS in order to develop its own operational picture of the situation in which it participates. For instance, LFs use LFCS, MFs use MCOIN, and finally, AFs use AFCCIS. Achieving these CCISs interoperability is a complex operations. Indeed, each CCIS has its own structural and functional characteristics. In order to obtain a global operational picture, the different CCISs of Figure 1 need to collaborate. Therefore, they have to exchange services without being aware of their individual characteristics, such as localizations and querying languages. To this end, we suggest to entrust the realization of this collaboration to SAs. Each SA will be in charge of different types of operations, such as interacting with users and requesting, on their behalf, the CCISs that are required in the design of a global picture. In Figure 1, SAs are set up on top of CCISs. L-Global L-Partial Middleware (CORBA) M-Global M-Partial Middleware (CORBA) A-Global A-Partial L-Software Agents M-Software Agents A-Software Agents Land Forces (LFCS) Maritime Forces (MCOIN) Air Forces (AFCCIS) Inputs for the L-GP Inputs for the M-GP Inputs for the A-GP Figure 3 GPOP Environment In the GPOP project, a global picture integrates parts from different partial pictures. In addition, these pictures are from different origins. Figure 4 provides an example of a global picture for LFs. This picture is based on the while LFs partial picture and on appropriate 3
4 parts from the AFs and MFs partial picture. These parts enrich the LFs partial picture. SAs carry out this enrichment process (this matter is not discussed in this paper). M-Partial A-Partial M-PP A-PP L-Partial L-Global Partial Operational Figure 4 Example of a Global Operational Partial Once a global picture is obtained, military users use this picture in their daily activities. In what follows, we explain how qualitative reasoning could provide support and assistance to these users activities. 3. Qualitative reasoning in GPOP It is well know that spatial localization is becoming a major issue in the military field. Being aware where is the enemy and where are our forces as well as allied forces is primordial. However, it occurs that certain types of queries, such as how many tanks are there near the river in the left side, require a qualitative reasoning in order to be satisfied. In the GPOP project, we believe that people build an Influence Area (IA) around objects [2]. IA s notion is an abstraction of the way these objects influence people s perception. As an illustration, two aircraft carriers separated by 10 nodes are close, comparing to the distance that separates them. However, two frigates separated by the same distance are far given that they are relatively small comparing to the distance that separates them. Then, the reasoning has been influenced by the relative importance of objects and their associated influence areas. 3.1 What is an Influence Area? Given an object O of any shape, the IA of O is a portion of space surrounding O such that (Figure 1.A): IA has two borders (an interior border and an exterior border); IA s borders have the same shape as O s border; if from any point Oi located on O s border BO we draw a perpendicular line, this line crosses IA s interior border at point IAIBi and IA s exterior border at point IAEBi such that ( Oi BO) (dist(oi,iaibi) = c1 and dist(oi,iaebi) = c2 and c1>c2). The distance dist (ISIBi,IAEBi) is called the width of the influence area. Thanks to the IA, the qualitative definition of neighborhood can be formulated as follows: Object O 2 is a neighbor of object O 1 IFF (O 2 IA(O 1 )) Ø. In order to handle the subtle way that people qualify distances between objects such as very close, close, and relatively close, multiple influence areas around each object, have to 4
5 be designed. Each IA would represent a certain degree of proximity. Hence, the qualitative definition of distance is now formulated as follows: Object O 2 is at a certain degree of proximity (dp) of Object O 1 IFF (O 2 IA dp (O 1 )) Ø where IA dp (O 1 ) denotes the influence area characterizing the qualitative distance dp to Object O 1. 3.2 Search & Rescue, as an application In land search and rescue operations, spatial reasoning is omnipresent and almost all the controllers have to deal with spatial and temporal constraints. A typical example of this reasoning consists of identifying the search area. Generally, a controller delimits the extent of the area according to predefined rules. In fact, this controller has a partial picture of the current situation. Thereafter, he studies the topology of the terrain and the nature of the spatial objects it contains in order to extend the search area. Moreover, before the controller commits the available resources to the current rescue operation, this controller s operational picture is enriched with information provided by other partners, such as Air Forces and their patrol planes. Therefore, this controller s operational picture evolves into a global operational picture. Next, the controller uses the QSM in order to satisfy his needs, such as identifying the last known position of the object in distress. It occurs that witnesses, who do not know the exact localization of this object, have reported this position. 4. Conclusion In this paper, we presented an overview of the GPOP project. In this project, we aim at designing in which distributed and heterogeneous CCISs could interoperate. Advanced technologies are used, such as software agents. In addition, a new qualitative approach supports the spatial reasoning in this framework. This approach uses the concept of influence areas that people build around objects that they perceive in the environment in order to contextually reason about space, evaluate metric measures, qualify positions and distances, etc. References [1] S. Green, L. Hurst, B. Nangle, P. Cunningham, F. Sommers, and R. Evans. Software Agents: a Review. Trinity College Dublin and Broadcom Éireann Research Ltd. May 1997. [2] D. Kettani, Conception et implantation d un modèle spatial qualitatif qui s inspire du raisonnement spatial de l'être humain, Ph.D. Thesis, Computer Science Department, Laval University 1999. 5