Autonomous Systems as an Integral Part of Military Operational Capabilities

It is widely accepted that autonomous weapon systems and robotics will play a major role in shaping warfare in the future. All major countries have accepted this truth and yet, only USA has actively and officially started work towards development and induction of autonomous systems in active duty. USA has set the benchmark in autonomous systems induction thanks to increased budget in relevant R&D fields.

Benefits Of Autonomy

The benefits of autonomy in weapon systems are many. In a nutshell, it will make warfare more precise, deadly and quick. Moreover, it will prove extremely useful in mobilization and logistics which often are the most challenging tasks in war.

Speed is one of the most attractive products of autonomy. No human can execute the orient, observe, decide, act (OODA) loop as fast and effectively as a machine. This is why time critical missions in particular make extensive use of autonomous systems. Electronic and cyber warfare, as well as their countermeasures are almost entirely dependent on autonomy and in both attack and defensive capabilities.

Autonomy also provides possibilities of improved command and control as these systems are capable of collating, collecting and interpreting data, often from places where troops on the ground can’t safely or rationally reach. Agility is another benefit of autonomy. It gives weapon systems a lot more agility from an operational and command and control perspective but reducing the requirement of human contact. Allowing the forces to scale down the involvement of man power in data collection, communication and processing for overseas assignments.

The new generation of warfare, often occurring in urban areas has seen a huge requirement of extremely accurate and precise weapon systems and payload delivery, especially because of the presence of civilians in the areas of operations. It is for this reason that the accuracy provided by autonomy is a substantial benefit for modern military and counter terrorism operations across the globe, greatly reducing collateral damage. These weapon systems are also persistent and remain unaltered over time. They don’t get tired, or sick or don’t lose motivation in the battlefield. Commanders can easily send these weapon systems into dangerous situations without having to worry about any loss of life. Reach of autonomous systems is another advantage or conventional weapon systems. They give military commanders the unparalleled capabilities of sending weapon systems to areas of extreme danger as well as those which will require extensive human effort to access. UAVs (Unmanned aerial vehicles) can have a greater range than aircrafts with pilots, increasing fuel efficiency and decreasing costs.

How Much Autonomy Should Weapon Systems Have?

It is with general consensus that autonomous systems while extremely beneficial, should not exist without human supervision. From a military standpoint, lack of human supervision neither is, nor should be the objective. Rather, militaries wish to have complete and full control over autonomous systems. This narrative suggests that the goal of developing autonomous systems is not to remove human involvement altogether, rather to develop capabilities of troops and weapon systems such that they complement each other. It is no secret that while humans might not be physically capable of achieving what a machine might, machines do not and might never possess the morality and decision making capabilities of our troops.

Technical Obstacles Facing Autonomous Systems

The biggest and most obvious challenges or obstacles to development of autonomous weapon systems are the technical limitations themselves. It is observed that existing autonomous systems lack flexibility in operations. That is, autonomous systems are capable of dealing with only those situations which the programmers may perceive or foresee and understand. Overlooking greatly, the fact that warfare is extremely unpredictable and no playbook exists on how to act or behave in situations of extreme unpredictability. The safety and effectiveness of autonomous systems can only be ensured in situations which can be predicted by programmers and are non-adversarial. Moreover, existing technology has greatly limited machine perception capabilities. That is, autonomous systems have very limited capabilities in terms of perceiving and making sense of their environments and adjusting their actions therein. Therefore, targeting systems that depend on this limited technology do not cope with the principles of international laws; distinction, proportionality and precaution in attack.

Another technical issue is that posed by hardware limitations, keeping autonomous systems from fully realizing their potential. An example is Boston Dynamic’s pack mule, a mobile autonomous logistic system which is capable of carrying weapons and other equipment into the battlefield, to aid soldiers. The pack mule operates on a diesel engine and is simply too noisy to be effective in a battlefield. Boston Dynamic did introduce a quieter electric system instead of diesel, but it simply couldn’t carry the same amount of weight. This is a common problem in almost all robotic systems, the limitations of their propulsion systems. Finding the right human-machine ratio poses extremely diverse problems. One of these problems is also quite ironic.

Over dependence of human operators on their machine counterparts can be dangerous and result in reduced cognitive focus in each all aspect of their operation, this is also known as automation complacency. Under-trust is quite the opposite and also a major issue. How can you make your troops trust an autonomous system enough to put their lives on the line alongside them? Under-trust in human operators sometimes results in them ignoring relevant information. In a nutshell, a fundamental problem with autonomous weapon systems is; how to properly calibrate sufficient human control over autonomous systems and ensure that is remains reliable.

The technical limitations of existing V&V (Validation and Verification) procedures is another hurdle we face. These procedures refer to various tests and evaluation processes aimed at providing assurance that the latest autonomous systems can adequately meet and satisfy the user’s demands. Therefore, these procedures are crucial in determining the weapon system’s effectiveness, reliability and whether it is safe to be deployed. As autonomous systems become more complex, dynamic and intelligent, it becomes impractical to put them through a whole range of inputs to test their performance in the wide variety of functions they can carry out. This problem in fact, is not just exclusive to autonomous weapon systems, rather, to all autonomous systems under development today. The existing V&V procedures can only test for and validate static and predictable environmental conditions, which are simply not enough to test the complete capabilities of these systems.

Militaries across the globe have multiple reasons to accept autonomy as an integral part of their operational capabilities. Research in autonomy will make weapon systems far more precise, mobile and in some cases, even cost effective than they are today. Is has also been established that complete autonomy is not the desire, rather, the aim to have effective and complete control over these autonomous systems to make them an extension of existing capabilities. Autonomous systems will alter the relationship of humans with machines and the operations carried out in the battlefield, but they will not replace the value of human involvement.