A GIS Assisted Knowledge-Based Approach for Military Operations

Maj. R. Baijal*, M. K. Arora and S. K. Ghosh**

Geomatics Engineering Section, Department of Civil Engineering
Indian Institute of Technology Roorkee, Roorkee 247 667
*rabaijal@yahoo.com, **scnagosh@datainfosys.net

Abstract
Military history is full of incidents wherein a smaller army having a good knowledge of the terrain
has defeated a much larger well-equipped and organised army. Nearly, all military activities are
terrain sensitive and need careful planning and reconnaissance to ensure success. However,
planning of military operations is a complex process and is guided by the experience and
capability of the commander and his staff who provide the necessary inputs to him. This decision
making process can be made intelligent by developing Knowledge Based (KB) expert systems. In
this paper, a knowledge-based approach has been used to produce a number of thematic maps
useful for various military operations. The approach developed is capable of taking inputs in the
form of data layers that may be generated from satellite images, aerial photographs,
topographical maps or other ancillary data. Some common military operations such as selection
of sites for bridges, ferries and helipads, identification of tactically important roads and vehicle
mobility movement are considered. The development of such knowledge-based approached shall
tremendously assist the military commander to provide efficient and real-time information in an
organized way for any military task.

Introduction
With the present cold war situation between India and its adjoining neighbour, the defence forces
have to be on alert at all times. Any emergent situations means that the army has to move
towards the border at a very short notice. The modern battlefield is highly mechanized with heavy
arms and ammunitions to shift around. The mobility of any armoured column depends upon the
terrain conditions over which it has to move. Ground conditions have always played an important
role in all conflicts over the ages. The parameters like topography, soil type and land use land
cover have a direct bearing to key activities like mobility of both men and machines, methods of
crossing obstacles, selection of tactically important areas etc. Logistics also play an equally
important role as weapons in a war. Replenishment of ammunition, fuel and other supplies are
required to reach the fighting troops in time. These require careful planning in terms of routes to
be taken and movement of various types of vehicles to ensure success. Thus, in today’s modern
battlefield, speed of planning and execution of operations is of prime importance.

Fortunately, we are living in an Information Technology (IT) era where the dissemination of
information from one place to another has virtually become real time. The IT tools can be
sufficiently exploited for any challenging task such as planning of wars. Remote sensing,
Geographical Information System (GIS) and artificial intelligence technologies are sitting on the
top of these IT tools that can together be effectively utilised to develop intelligent systems for war
planning. Command, Control, Communication, Coordination and Information (C4I) is one such
system where these technologies can be effectively used. For example, satellite remote sensing
data can be used to generate a wide range of products such as land use land cover maps,
obstacle maps, slope maps, road mobility maps, line of sight plots etc. A GIS can receive,
process, create, store, retrieve, update, manipulate and compress digital terrain data to generate
a number of products. Knowledge of experts is a key input for any C4I system. Knowledge Based
(KB) systems are being developed for war planning that can process inputs from remotely sensed
and GIS derived products and use the knowledge gained to aid the decision making process,
thereby allowing the military commanders make better battle plans. GIS embedded C4I systems
aim to give this KB to field commanders and their staff who despite having little knowledge of
GIS, can work on such systems. Currently only a few C4I systems are in use with embedded GIS,
but their numbers are likely to rise substantially soon as more and more systems are developed
around the world.

This paper presents a study on the use of a GIS assisted knowledge-based approach for some
military operations such as selection of sites for bridges and helipads, identification of tactically
important roads and preparation of vehicle mobility maps.

Some common military operations

The commanders have to carry out careful planning of a range of activities required during any
military operation in war. Some of them are:

Selection of Sites for Launching Bridges

To provide mobility to ground forces across water bodies, two types of bridges are generally
employed by the military. Wet bridges are built across the rivers and large water bodies where
these can float. For small water bodies such as canals and drain, dry bridges clear of the water
surface are provided. However, these bridges have fixed specifications of span, launching slope
and bank conditions. Therefore, a suitable site has to be selected to meet these requirements
with some site preparations. Selection of Ferry Sites

For crossing the canals and rivers, suitable ferry sites are needed till bridges are constructed over
them. The terrain requirement is somewhat similar to bridges except that some form of road or
track on the banks of the water bodies to provide access to the ferry site is required.

Selection of Sites for Helipads

During the movement of the troops and equipment, sufficient air cover essential in today’s
warfare, as these are easy targets from air by the enemy. Keeping the capabilities of a helicopter
in mind, it is being increasingly used in combat role and other tasks like reconnaissance,
evacuation of casualties etc. The dimensions of the helipad required for the landing of a
helicopter varies from place to place but the ground conditions may nearly be the same as for
bridge and ferry sites. For example, the location of a helipad depends upon the tree cover, soil
conditions and slope of the ground.

Identification of Tactically Important Roads

In order to provide fast and safe movement of troops and equipment, identification of tactically
important roads is essential. Roads and tracks that lead up to the likely bridge or ferry site are
tactically important and need to be identified and suitably constructed so that these may be used
as the axis of maintenance. Ideally these roads should not pass through any obstacles like the
minefields.

Preparation of Vehicle Mobility Maps

Military vehicles are generally classified into two broad categories, tracked and wheeled. Vehicles
like tanks have excellent cross country mobility due to presence of tracks over its wheels.
However vehicles having wheels but without tracks do need careful route planning before cross-
country movement can be attempted. Vehicles carrying essential war stores like ammunition; fuel
and other supplies are all wheeled vehicles.

The need for a knoowledge based approach

The activities mentioned above are just a few undertaken by the defence forces while planning for
a military operation. Most of these require good interpretation skills to understand the terrain.
These skills may vary from person to person and hence the interpretation is also likely to vary.
This may adversely affect the battle plan therefore, there is a need to standardize procedures and
incorporate systems, which use the existing knowledge acquired by experts, intelligence agencies
and other means. This knowledge base (KB) can be effectively used to make accurate decision
making tools which can easily be used by military commanders at all levels.

KB expert systems can be developed, which can take into account the experience and knowledge
of terrain analysts and other experts to convert them into a set of rules, which can then be applied
to digital data to derive a number of thematic maps that can in turn be used in war planning.

A typical KB system comprises of a set of interrelated and interconnected components such as

knowledge base, inference mechanism (IM), user-interface, mechanism to update KB, and the
explanation of the rules applied (Nikolopoulos, 1997). The KB is a systematic collection of
information from various sources and experts in the area of application (e.g., war planning here).
It is organizes the information into rules, which are generally written in the form of IF-THEN-ELSE
statements. The IM is the work center as it provides the deductions or solution to a particular
problem based upon the rules framed. The user interface is the link between the user and the KB
such that a non-expert can also use it comfortably. With passage of time, the KB has to be
updated in terms of informations and rules, thus Mechanism to update KB is an essential facility
to upgrade and check the validity of KB. Generally, a KB system shell is built that houses the IM,
the user interface, an explanation system and a knowledge base editor. There are numerous
commercial KB system shells, each one appropriate for a slightly different range of problems.
Using shells to prepare a KB system generally reduces the cost and time of development.

In this paper, the knowledge based classifier in the well known Image Processing and raster GIS
software namely ERDAS Imagine has been used as a shell to develop a knowledge-based
approach for the military operations mentioned above.

Study area
The KB presented here has been designed keeping in mind its effective use in the western
borders of our country. In the western region, the general terrain conditions are plain with the soil
type varying from desert sands of Rajasthan to the marshy areas of Gujarat. The landscape in
this region is interspersed with rivers and canals.

However, due to the security restrictions in use of topographical sheets of the border areas, an
area having somewhat similar ground conditions has been selected here to test the knowledge-
based approach developed. The area lies south of Saharanpur city, UP, between Latitude 29°45’
to 29°50’ and Longitude 77°30’ to 77°35’. Survey of India (SOI) map sheet 53 G/9 having a scale
of 1:50,000 has been used. The area covered is approximately 8 Km by 8 Km. This area has one
major river Hindan flowing from north to south and few canal systems as shown in Fig 1.
 Fig. 1 Topographical map of study area selected

GIS Data Layers

The rules to be formed are based upon the spatial information about the terrain as desired by a
particular military operation. A number of spatial raster data layers are prepared. The data for the
study area has primarily been collected from the topographical map as above. However, due to
non-availability of certain information, some hypothetical data has also used for the preparation of
layers. Following data layers have been prepared by on-screen digitizing the scanned
topographic map in ERDAS Imagine,

i. Water Bodies
In the study area one major river, River Hindan, is flowing from north to south with a
network of canals, to the north west of the area and some small lakes around villages.
Thus, three categories of water bodies are considered as rivers, canals and lakes with
their raster values assigned as 1, 2 and 3 respectively (Fig. 2).
 Fig. 2 Thematic layer of water bodies

ii. Road network

Existing roads and tracks are an important aspect as it ensures the smooth movement of
both fighting columns and logistical supplies. The area has a good network of roads and
tracks both along the river and perpendicular to it. The roads and tracks are therefore the
automatic choice for the categories in roads layer and are assigned a raster value of 1
and 2 respectively (Fig. 3).

Fig. 3 Thematic layer of road network

iii. Power Lines
Existence of power lines is an important factor while deciding the location of helipads.
Hence, power lines have also been digitised and assigned a raster value as 1 and rest of
the area as 0 where no power lines exist (Fig. 4).

Fig. 4 Thematic layer of power lines

iv. Slope Map

The area is in general flat with heights varying from 252 m to 267 m. Only one contour at
20 m interval passes in the vicinity of the area. Therefore, spot heights and benchmarks
available on the topographical sheet have also been digitized to generate a Digital
Elevation Model DEM. 3D surfacing tool of ERDAS Imagine has been used to generate a
raster DEM from the digitized contour and the spot heights (Fig. 5). From this DEM, a
slope map has been prepared showing four categories as, no slope (0% slope), plain
(slopes less than 5 percent), moderate (slopes between 5 to 10 percent) and steep
(slopes greater than 10 percent) (Fig. 6).
 Fig. 5 Raster DEM of study area
Fig. 6 Thematic layer of slopes
v.
vi. Land use land cover map
IRS-LISS III data has been used for classification of various land use and land cover of
the area. Five land use land cover categories namely built-up areas, forests, rivers,
canals and cultivated areas have been selected and given raster values as 1, 2, 3, 4, and
5 respectively (Fig. 7).

Fig. 7 Thematic layer of land use

In addition to the above data layers, two more data layers have been prepared based upon the
hypothetical data,

i. Soil map
Soil conditions play an important part in the process of site selection for various military
tasks like bridging, helipad location etc. The factors like bearing capacity and the soil
moisture are critical for operations like bridging. Approach to site and area near a bridge
must be able to take on heavy vehicular traffic. Due to lack of authentic soil map of the
area, a hypothetical soil data has been generated and is classified into three categories
namely sand, silt and clay with their raster values as 1, 2 and 3 respectively (Fig. 8).

Fig. 8 Thematic layer of soil types

ii. Water depth

Wet Bridges are built using pontoons having decked girders, which float on water. Hence
there is need for minimum water depth, which is generally kept as 1m. In absence of
authentic data, hypothetical water depth data for the river in the study area has been
generated and divided into two broad categories as adequate or inadequate with raster
values assigned as 1 and 2 respectively (Fig. 9).
 Fig. 9 Thematic layer of water depth

Generation of rules for the KB

Knowledge acquired in the form of GIS data layers converted into rules that are transformed into
a knowledge base using the Knowledge Engineer (KE) shell of ERDAS Imagine. Each data layer
provides a parameter or condition, which can be used for the formation of rules to get the final
hypothesis. Separate knowledge base has been prepared for each military operation selected in
this paper. For brevity, the KB for the selection of wet bridging site has been described here.
However, all the graphical representations of each KB are shown in Fig. 10 to Fig. 15.

Fig. 10 KB for Selection of Wet Bridging Sites

Fig. 11 Knowledge Base for Selection of Dry Bridging Sites

Fig. 12 Knowledge Base for Selection of Ferry Sites

Fig. 13 Knowledge Base for Selection of Helipad Sites

Fig. 14 Knowledge Base for Identification of Tactically Important Roads

Fig. 15 Knowledge Base for Preparation of Vehicle Mobility Maps

For launching a wet bridge, the parameters that are likely to be considered are type of water
body, ground slope, type of soil, land use and availability of adequate water depth. The KB for the
selection of wet bridge site consists of the following rule written in the text form as:
IF WATER BODY == RIVER (1)
AND SLOPE < = PLAIN (5)
AND SOIL TYPE > = SAND (1)
AND LAND USE >= CULTIVATED AREA OR RIVER OR CANAL (3)
AND WATER DEPTH == ADEQUATE (1)
THEN SUITABLE SITE FOR WET BRIDGING
The numbers in the bracket show the respective raster values. On executing this KB, the
hypothesis gives the output class (colour coded as red) as the possible site locations for the wet
bridge (Fig. 16).
Similarly, once the KB for each military operation is executed, outputs are a set of thematic maps,
which are shown in Fig. 17 to Fig. 20. These thematic maps have been visually analysed with the
input images in relation to the rules applied. The broad areas as identified by each KB have been
checked using topographical map and military data regarding parameters for the various military
uses, and have been found to be correct. Thus, the KB approach can be effectively used for
military operations. The thematic maps thus produced can be also used as overlays to carry out
an accurate planning for various military tasks.
Conclusions
Accurate and timely terrain analysis is the key for today’s fast paced mobile battlefield.
Conventional techniques need to be updated due to availability of data products like maps in
digital form and high-resolution satellite imagery. The knowledge base approach for the
interpretation of terrain features will prove to be very useful for modern day war planning. This
approach combines the experience and knowledge of experts and delivers this to the soldier in
the battlefield.