Artificial Intelligence In Military Application Information Technology Essay

Since the dawn of civilizations, humans have endeavored to be in control of their environments and surroundings. This quest resulted in many discoveries and inventions, most notably among them are machines. Human used machines as an aid to make one’s life comfortable, effective and efficient and aimed to develop machines capable of working like human beings, if possible. Computer is one of the most important machines which has not only raised hopes in this regard but has also contributed significantly in every sphere of human endeavor. Human approach to problem solving is one of its kinds. It is based on abstract thought, logic, reasoning and recognizing of pattern. Computers and humans are different. A computer is yet to understand all situations and simultaneously adapt to an evolving situations.

The military systems including weapons will be smart; too fast, too small, too many, and will create a complex environment for humans to monitor, control and direct them. Information-based systems will lead to a data overload that will make it a challenge for humans to directly intervene in decision making. Weapons and other military systems already under development will function at increasingly higher levels of complexity and responsibility, without meaningful human intervention and control.

In future military conflicts, norm of engagement will be to act rapidly. The military architectures of tomorrow will consist of a new array of sea, ground and space based sensors, unmanned combat aerial vehicles (UCAV), and missile defence technologies. These will take advantage of directed energy weapons. Military forces will be both faster and agile. Opponents will take advantage by operating faster than a defender can observe, orient, decide how to respond and act on that decision. The attacker will thus place himself inside the defenders Observe, Orient, Decide and Act (OODA) loop, destroying an adversary’s ability to conduct an active defence [1] . To execute the OODA process faster than the enemy is at the core concept of future digital and information warfare. Automated systems, assisted by artificial intelligence in some form or the other, may be a way out for this problem. The advances gained in the field of artificial intelligence technology can be utilized by unmanned systems to be able to assess operational and tactical situations and decide an appropriate action. Information will drive success of command and control. These systems will collect data, have the ability to analyze data and provide recommendations to the commander. The difference between providing a recommendation and acting on a recommendation may be only a software twist.

Artificial Intelligence (AI) is the branch of computer science focusing on creating machines that can engage on behaviour that humans consider intelligent. AI aims to improve machine behaviour in tackling complex tasks. Smart machines have now become a reality and researchers are creating systems which can mimic human thought, understand speech, beat the best human chess player, and achieve many other advantages. With the introduction of web-enabled infrastructure rapid developments have been made in the application of Artificial Intelligence techniques in the recent past. AI is the key technology in many of today’s applications in all field including military.

AI methodologies are being applied to support decision making at all levels of military operations such as assessment of force readiness, reliability and capability, complex missions planning and integration of data from multiple sources [2] . Research in the field of AI is also addressing the challenges presented by supporting such decision making in rapidly changing environments. The use of such technology opens up endless possibilities in the military. This paper aims to trace the contours of AI, examine current efforts to utilise Artificial Intelligence and explore its potential applications in military.

Genesis and Recent Past             

The first Electronics computer was developed in 1941; however, the field of AI research was formally founded at a conference at Dartmouth College in 1956 only. Early work in AI focused on using cognitive and biological models to simulate and explain human information processing skills. In the 1990s and early 21st century, AI achieved its greatest successes. AI has advanced rapidly in the past decade. This happened due to greater use of the scientific method in experimenting with and comparing approaches that systematises and automates intellectual tasks. Thus it is relevant to any sphere of intellectual activities of human.

The success can be attributed to the incredible power of modern computers, a greater emphasis on solving specific sub-problems, the creation of new ties between AI and other fields working on similar problems. AI is seen and perceived by different people or groups of people differently. The definitions of artificial intelligence can be broadly put into two approaches: one centred around humans and other centred on rationality [3] . The human centred approach must be an empirical science involving hypothesis and experimental confirmation while a rationalist approach involves a combination of mathematics and engineering.

There have been efforts to introduce new creative approaches and refine the best one. Recent progress in understanding the theoretical basis for intelligence has gone hand in hand with improvements in the capabilities of real systems. Various subfields of AI have become more integrated. AI has found some common ground with other disciplines. A better understanding of the problems and their complex properties, combined with increased mathematical sophistication has led to workable research agendas.

Fields of AI

In AI the problem of intelligence simulation is generally divided into a number of specific sub-problems. These consist of particular capabilities that researchers like an intelligent system to display. For difficult problems, most of the algorithms require large computational resources and the amount of memory or computer time required become very high. With rapid strides in computer technology, research and utilization, the field of AI witnessed new frontiers. Russell and Norvig explains, “AI encompasses a large variety of subfields ranging from general purpose area such as learning and perception to such specific tasks as playing chess, proving mathematical theorems, writing poetry, and diagnosing diseases. AI systematises and automates intellectual tasks and is therefore potentially relevant to any sphere of human intellectual activity. In a sense, it is a truly universal field” [4] The various fields and subfields that received more attention in order to solve larger problems are: –

Learning The centrality of learning was discussed by Turing in 1950. From the beginning itself machine learning has been central to AI research. The ability to find a pattern in a stream of input is called unsupervised learning where as supervised learning includes classification and numerical regression both. Classification is used to determine what category something belongs in. This is done after seeing a number of examples of things from several categories. Regression takes a set of numerical input and output examples and attempts to discover a continuous function that would generate the outputs from the inputs. In case of reinforcement learning, the agent is rewarded or punished based on good or bad responses.

Natural Language Processing It gives machines the ability to read and understand the languages spoken by human beings. Text mining and machine translation are example of some basic applications of natural language processing.

Perception Perception provides agents with information about the world in which they exist. Perception is initiated by sensors. Machine perception is the ability to use input from various sensors such as cameras, microphones, sonar etc to deduce aspects of the world. Computer vision is the ability to analyse visual input. Facial recognition, object recognition and speech recognition are some of the selected sub-problems

Social Intelligence In order to obtain social intelligence capability, Artificial intelligence has to establish able human interaction and also possess the emotions that people have during their everyday lives. Social skills and emotion play two important roles for an intelligent agent. First, it should be able to foresee the actions of others, by knowing their motives and state of emotions. This involves elements of game theory, decision theory, as well as the ability to model human emotions and also the perceptual skills to detect emotions. Also, it is expected that for good human-computer interaction, emotions need to be displayed by an intelligent machine also. It must appear polite and sensitive to the humans it interacts with. At best, it should have normal emotions and at least it should appear polite.

Creativity Artificial Intelligence that deals with the development and exploration of systems that exhibit creativity. It includes systems capable of such things as scientific invention, visual artistry, music composition and story generation etc. A section of AI addresses creativity both theoretically from a psychological perspective and practically via specific implementations of systems that generate outputs that can be considered creative. Artificial Intuition and Artificial Imagination are the areas related with computational research.

General Intelligence Many of the researchers hope that their work will finally be included into a machine with general intelligence, combining all the other skills and exceeding human abilities at most of them.

Knowledge Representation knowledge representation is one of the important and most familiar concepts in AI. Most of the problems that machines are expected to solve will require extensive knowledge about the world. Among the things that AI needs to represent are objects, properties, categories and relations between objects, situations, events, states and time, causes and effects, and many other less well researched domains. Knowledge representation and knowledge engineering are central to AI research

Planning Planning are the subfields of AI devoted to finding action sequences that achieve the agent’s goals. Intelligent agents should be able to lay down goals and accomplish them. They need a way to imagine the future and be able to select choices that maximize the value of the available choices. They should have a representation of the state of the world and be able to make predictions about how their actions will change it. Multi-agent planning uses the cooperation and competition of many agents to achieve a given goal.

Tools Used to Solve Problems of AI

In the course of years of research, AI has developed a large number of tools to solve the difficult problems in computer science. A few of the most common of these methods are mentioned: –

Search and Optimisation Search is the subfield of AI devoted to finding action sequences that achieve the agent’s goal. Several problems in AI can be solved in theory by intelligently searching through many possible solutions. Reasoning can be reduced to simply perform a search operation. Planning algorithms search through trees of goals and sub-goals, attempting to find a path to a target goal. This process is called means-ends analysis. Robotics algorithms for moving limbs and grasping objects use local searches in configuration space. Several learning algorithms use search algorithms based on optimization. In case of most of the real world problems, simple exhaustive searches are rarely sufficient. Therefore, heuristics supply the program with a best guess for the path on which the solution lies on.  Also in case of many problems, it is possible to begin the search with some form of a guess and then incrementally refine the guess until no more refinements can be made.

Logic Logic is the primary vehicle for representing knowledge. It is used for knowledge representation and problem solving. However, it can be applied to other problems also. AI uses several different forms of logic research. Propositional or sentential logic is the logic of statements which can be true or false. There is well developed technology for reasoning in proportional logic First-order logic also permits the use of quantifiers and predicates. It can express details about objects, their properties, and their relations with each other. Fuzzy logic is a version of first-order logic. It allows the truth of a statement to be represented as a value between 0 and 1, rather than simply True or False. Fuzzy systems have been widely used in modern industrial and consumer product control systems and can be used for uncertain reasoning. Several extensions of logic have been intended to handle many domains of knowledge. Other forms of logic designed to help with default reasoning and the qualification problem include default logics, non-monotonic logics and circumscription

Probabilistic Methods For Uncertain Reasoning A large number of problems in AI such as learning, reasoning, planning, perception and robotics call for the agent to operate either with uncertain or incomplete information. A number of powerful tools using methods from probability theory and economics have devised by AI researchers to solve these problems. Bayesian networks are a very common tool that can be used for a large number of problems likes learning, reasoning, planning and perception etc. Probabilistic algorithms can also be used for filtering, prediction and finding explanations for streams of data, helping perception systems to analyse processes that occur over time. Mathematical tools have been developed that analyse how an agent can make choices and plan, using decision theory, decision analysis, information value theory.

Classifiers and Statistical Learning Methods Classifiers and Controllers are two types of AI applications. Classification forms a central part of many AI systems however, controllers do also classify conditions before inferring actions. Classifiers are functions that make use of pattern matching to determine a closest match. They can be tuned as per examples, making them very attractive for use in AI. These examples are known as patterns or observations. Each pattern belongs to a certain predefined class in case of supervised learning. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on earlier experience. A classifier can be trained in many ways.

Neural networks A neuron is a cell in the brain whose principal function is the collection, processing, and dissemination of electrical signals. The brain’s image processing capacity is considered to emerge mainly from network of neurons. A neural network is an interconnected group of nodes, similar to the large network of neurons in the human brain. Neural networks are designed to distinguish patterns in data and forecast an output from a given set of data. Neural networks need to be trained on the data before it can predict or learn. In this way they learn from examples similar to the way a child learns. Computer learning skills that are developed for neural networks are used in a class of computer programs called ‘expert systems’. Expert systems also learn from experience and get better at their job, the longer they are doing it.

Control Theory Control theory is the foundation of AI and deals with designing devices that act optimally on the basis of feedback from the environment. Initially, the mathematical tools of control theory were quite different from AI, but the fields are coming closer together. Control theory, has many important applications, especially in robotics.

Languages AI researchers have developed several specialized languages for AI research, including Lisp and Prolog. In AI, the automation or programming of all aspects of human cognition is considered from its foundations in cognitive science through approaches to symbolic and sub-symbolic AI, natural language processing, computer vision, and evolutionary or adaptive systems. It is inherent to this very complex problem domain that in the initial phase of programming a speci¬c AI problem, it can only be speci¬ed poorly. Only through interactive and incremental re¬nement does more precise speci¬cation become possible. This is also due to the fact that typical AI problems tend to be very domain speci¬c therefore, heuristic strategies have to be developed..

Applications of AI

General

The probable applications of Artificial Intelligence are plenty. Application of AI is possible in all fields, requiring intelligent analysis, precision and automation. They stretch from the military to the entertainment industry, to big establishments dealing with large amount of information such as banks, hospital and insurances. AI can also be used to predict customer behavior and detect the trends. There are many general fields where AI can be very usefully utilized for Autonomous planning and Scheduling, Autonomous Control, Medical Diagnosis, Logistics Planning, and Language Understanding, Problem Solving, Game Playing [5] . Some of the important applications are appended below: –

Pattern Recognition Pattern recognition is the area of research that studies the operation and design of systems that identify patterns in data. When a program makes observations of some kind, it is often programmed to compare what it sees with pattern e.g face, fingerprint or handwriting recognition. Important application areas are image analysis, character recognition, speech analysis, man and machine diagnostics and person identification.

Bio-Informatics Bioinformatics is the application of computer technology for the management of biological information. AI provides several powerful algorithms and techniques for solving important problems in bioinformatics. Approaches like Neural Networks, Hidden Markov Models, Bayesian Networks and Kernel Methods are ideal for areas with more data but very less theory. The goal in applying AI to bioinformatics is to extract useful information from the wealth of available data by building good probabilistic models.

Data Mining An AI powered tool that can discover useful information within a database that can then be used to improve actions. Data mining is the process of extracting patterns from data. Data mining is seen as an increasingly important tool by modern business to transform data into business intelligence giving an informational advantage. It is currently used in a wide range of profiling practices such as marketing, surveillance, fraud detection, and scientific discovery.

Expert Systems An expert system is a computer program that represents the reason with knowledge of some specialist subject with a view to solve problems or give advice [6] . It is the knowledge-based applications of artificial intelligence that have enhanced productivity in almost all fields such as business, science, engineering, and the military. With advances in the last decade, today’s expert systems clients can choose from dozens of commercial software packages with easy-to-use interfaces. Diagnosis and Trouble-shooting explain the development and testing of a condition-monitoring sub-module of an integrated plant maintenance management application based on AI techniques. It is mainly knowledge-based systems, having several modules, sub-modules and sections.

Computer Vision It is essential for computer to perceive the objects. Vision includes the acquisition and processing of visual information both. AI enabled technologies have made possible many amazing achievements. Vehicles that are able to steer themselves safely along highways, and computers that can recognize and interpret speech or facial expressions. AI supported vision technology has made many applications possible. Some of them are like 3D modeling, image stabilization, image synthesis, surgical navigation, handwritten document recognition, and vision based computer interfaces. While explaining success of an autonomous system trial supported by computer vision Russel and Norvig mentioned:

“The ALVIN computer vision system was trained to steer a car to keep it following a lane.  It was tried in the Carnegie Melon University (CMU) NAVLAB computer controlled minivan and used to navigate across the United States for 2850 miles it was in control of steering the vehicle 98% of the time. A human took over the other 2% mostly at the exit ramps. NAVLAB has video cameras that transmit road images to ALVIN, which then computes the best direction to steer, based on experience from previous training runs.” [7] 

Image Processing Perception appears to be an effortless activity for humans however, it requires significant amount of sophisticated computation. The team associated with image formation and processing is concerned with research issues related to the acquisition, manipulation, and synthesis and distribution of images. In AI, applications include video phone, video conferencing, teleconferencing, and multimedia databases. Progressively, this research has combined image or vision with audio or speech. For example in the video indexing project, the group is using both visual and audio cues to derive semantic labels for video shots.

Robotics Robots are physical agents that perform tasks by manipulating the physical world. Robots are comprised of several systems working together as a whole. Robots are widely used in assembly plants, space stations, and hospitals and now in homes also. Other type of mobile robots includes unmanned land vehicle, unmanned aerial vehicles and autonomous underwater vehicle.

Knowledge Representation and Reasoning The representation of knowledge and the reasoning processes that bring knowledge to life are the two concepts, central to the field of AI. The knowledge representation means encoding real world, commonsense etc in a format that is both readable and understandable by the computer. Logical Agents is the representation of knowledge and the reasoning processes that bring knowledge to life. Logic is the primary vehicle for representing the knowledge throughout and Semantic webs describe things in a way that computer applications understand.

Gaming Game playing was one of the first tasks taken by AI. Games, unlike other problems, are interesting because they are too difficult to solve. Games like the real world require the ability to make some decisions even when calculating the optimal decision is not feasible. Games also penalize inefficiency severely. Game playing research has contributed in many ideas on how to make best possible use of time. AI techniques are used in computer and video games to produce the illusion of intelligence in the behaviour of non-player characters. The techniques used typically draw upon existing methods from fields that include control theory, robotics etc. IBM’s Deep Blue became the first computer programme to defeat the world champion in a chess match when it performed better than Garry Kasparov in an exhibition match.

Military Applications

The military applications of Artificial Intelligence are spread over large areas of military functions. Some of the military functions where AI techniques are being used or have potential use are given below.

Operations

Command and Control Command and Control is considered as one of the most important functions of military operations. In a networked centered scenario of battlefield with host of sensors deployed at different stages and the amount of data flowing between various centres, the time available for decision making is at premium. The information overload sometimes can impair the decision making hence require intelligent filtering of information to take timely and appropriate decisions. AI is used in playing a decisive role in reducing the load on the human beings in the loop and at times taking autonomous decisions as and when warranted.

Navigation The availability of Global Positioning Satellites (GPS), Inertial Navigation Systems (INS) and autopilot along with host of sensors and On-board computers has helped in overcoming certain human limitations and resulted in safe and efficient management of flying aircrafts.

ISR Intelligence, Surveillance and Reconnaissance are the key elements of battlefield management. Over time the battlefield scenario has undergone dramatic change and so as the means of identification, surveillance and reconnaissance. Advancement in technology in many spheres has offered sensors with high sensitivity, small sizes and better visibility. AI has contributed significantly in this regard in terms of ground, aerial, space and underwater ISR capabilities.  Unmanned Aerial Vehicles (UAV) using AI offer tremendous potential as intelligence, surveillance and reconnaissance platforms for early detection of security threats and for acquisition and maintenance of situational awareness in the crisis condition. Using their capabilities effectively requires addressing a range of practical and theoretical problems. Developments in the field of hardware and software technologies, as well as economies of scale, make UAVs feasible for increasingly diverse airborne observation missions. Expert systems are promising technologies that manage information demands and provide required expertise. Thus they are well suited to many of the tasks associated with environmental impact assessment. While highlighting the contribution of artificial intelligence in battlefield surveillance using geographical information system, Maj Jagmohan Singh of Project Management organization, Battlefield surveillance system , Army HQ concludes:

“Transparency of the battlefield is a critical factor influencing the outcome of future battles. Battlefield transparency would provide a framework for `scientific’ and deliberate decision making. The dependence of commanders on paper maps and sand models for operational planning will have to be replaced by the latest GIS tools. These tools permit dynamic visualization of a 3D terrain model for seamless access, query and analysis across multiple types of military geographical data. Mapping and analysis is done using various GIS technologies incorporating satellites and aerial imagery, and photography of the target area. The future technologies would further enhance the visualization techniques and enable the commanders to take timely decisions to defeat the adversaries. However, emphasis needs to focus on refinement of some critical technologies such as Multi Sensor Data Fusion (MSDF), Artificial Intelligence and Interoperability issues. ” [8] 

Weapon System The weapon technology has seen constant change and has gained more lethality and effectiveness in its evolution. A host of modern weapons are in use or in process of development which can change the landscape of the battlefield. Missiles, Directed energy weapons, Standoff weapons, autonomous weapons etc are few examples of intelligent weapons and have even greater potential in future.

Communications and Computers Communication is the core of all activities. In the age of modern communication, the geographical boundaries have come closer and visibility has improved to a great extent. The advent of satellite and availability of internet has revolutionized the communication. In future the success of battlefield will depend on maintenance of network connectivity and management of information from a large variety of sources. This will also made real time communication more important. The non availability of real time information can hamper the decision making ability of soldiers fighting the war. It may result in the failure of mission and even danger to personal survival. Wren, Ichalkaranje and Jain commented on the contribution and maturity of AI:

“Intelligent Decision Support Systems have the potential to transform human decision making by combining research in artificial intelligence, information technology, and systems engineering. The field of intelligent decision making is expanding rapidly partly due to advances in artificial intelligence and network-centric environments that can deliver the technology. Communication and coordination between dispersed systems can deliver just-in-time information, real-time processing, collaborative environments, and globally up-to-date information to a human decision maker. At the same time, artificial intelligence techniques have demonstrated that they have matured sufficiently to provide computational assistance to humans in practical applications.” [9] 

Network centric environment facilitates leveraging Artificial Intelligence to allow soldiers to access and share information throughout the entire network. Network centric environment provides coordination, where each node in the network helps provide a flawless, decentralized organization of intelligent resources.

Maintenance

Repair and Overhaul The fighting capability of the forces depends upon the serviceability and availability of the range of equipment held in its inventory. Most of the modern days equipments used for military application have certain defined life span. Also the demand of battlefield has ensured that highly sophisticated equipments should be made available for combat. This demands a system with quick fault diagnostic capability, easy maintainability and highly trained human resources along with modern ground facilities. Presently the expert systems are in use to analyze the faulty Printed Circuit Boards (PCB) of Radars or aircraft avionics using Automatic Test Equipments (ATE). Built-in Test systems are encouragingly being used with modern development of weapons. The techniques such as expert system and robotics are fairly in use in military application however has the potential to be exploited at much greater scale to expedite automation.

Logistics Logistics is the life line in case of military operations. The various models of operations research have been employed in effective management of the logistics operation. The system of simulation has helped in optimizing the operation and AI has a great potential in assisting in planning and keeping the supply chain effective and efficient. Russell and Norvig highlighted that:

“During the Persian Gulf crisis of 1991, US forces deployed a Dynamic Analysis and Replanning Tool (DART) to do automated logistics planning and scheduling for transportation. This involved up to 50000 vehicles, cargo and people at a time and had to account for starting point, destinations, routes, and conflict resolution among all parameters. The AI planning techniques allowed a plan to be generated in hours that would have taken weeks with older methods. The Defence Advanced Research Project Agency (DARPA) stated that this single application more than paid back its 30 year inv