Exploiting Intelligence Mission Data: an option for autonomous advantage

In his latest book, the soldier-scholar David Kilcullen lays out his ideas of how state and non-state actors adapt, including a discussion on the recent evolution of drones in war. The early chapters neatly document recent advances in commercially-derived drone technology, armaments, and tactical employment, presciently predicting many features of the current Nagorno-Karabakh conflict. One of the more troubling conclusions he draws is the increasing ability of these systems to operate without command signals, thus rendering many current Electronic Warfare (EW) based counter-drone technologies and tactics obsolete. This observation is potentially disconcerting for those charged with countering this increasingly capable and available threat. However, it also points to a potential emerging advantage for western forces to leverage their extensive collections of terrain datasets for tactical advantage.

As unmanned systems become increasingly autonomous (i.e. able to operate without command input whilst in mission), they necessarily rely on increasingly comprehensive data models of their environment, friendly and adversary systems, and other relevant criteria (e.g. weather effects). To be truly fire-and-forget these systems also require pre-determined Rules of Engagement, targeting criteria, and even tactical manoeuvres. Western militaries are more likely to prioritise these longer range, more sophisticated and tactically useful systems over the sensationalist ‘slaughterbots’ and personnel focused drones often noted by the media. The survivability and lethality of these systems will be in direct proportion to the quality of their information and the algorithms that exploit it; for example, simultaneously saturating a priority target with multiple missiles or drones is likely to be more effective in overwhelming the target’s defences than a sequential or linear attack. This reliance on quality data models and algorithms has important implications for how Australia might seek advantage in this field. In particular, it greatly increases the need for quality Intelligence Mission Data (IMD)[1] on the physical environment, target and threat characteristics.

Defence’s understanding of IMD is well advanced, but primarily through an EW lens; the need to understand the characteristics of how friendly, neutral and threat systems transmit and receive in the Electromagnetic Spectrum (EMS). Maximising the advantage offered by autonomous systems will require significant increases to the quality of EMS data, but also to the resolution and accuracy of physical terrain data, and other information about the battlespace. The battlespace is—after all—where the engagement occurs, and the force that has a better understanding of how to exploit its various features possesses a significant advantage. For autonomous systems this information permits covered or concealed navigation to battle positions or target areas; the capacity to recognise when and where targets are most likely to appear; and the ability to exploit the most advantageous attack profiles independently or in conjunction with other systems. The quality and resolution of this information is critical; elevation data suited for terrain avoidance by manned aircraft is unsuitable for a small Unmanned Aerial System trying to remain behind a tree canopy, or a Unmanned Ground System trying to move through rocky terrain. Having the right quality of battlespace data is therefore essential to generating future combat advantage.

For the autonomous systems of the future, quality battlespace data likely means centimetric resolution—that is pixels or point spacings of less than 10cm—with commensurate accuracy. Such accuracy and precision of data allows precise navigation and position with aid of computer vision and other passive sensing techniques. It also requires a shift towards Three Dimensional (3D) data. Current terrain information is typically limited to documenting only the surface data as viewed from directly above. 3D data goes beyond this to include information on multiple vertical and horizontal surfaces. For example, the thickness and composition of building walls, or the diameter and spacing of tree trunks. Better data allows for more precise targeting and understanding of weapon effects, but also better route selection or battle positions. This in turn allows improved concealment of system launch and control positions, improving a force’s physical or functional stand-off. The collection, storage and availability of this data is possible presently, but would likely need a new approach to the way it is distributed and consumed by the joint force.

These problems are not new. When Intercontinental Ballistic Missiles (ICBMs) first emerged to replace crewed bomber fleets in the 1950s they were necessarily massive to carry the very large nuclear warheads of the era. The warheads were large in part because hitting even a city-sized target with a missile fired from the far side of the earth is a difficult navigational problem. A significant contribution to the error of these warheads arose from local variations in gravity and the shape of the earth, particularly as the warhead descended on an essentially uncorrected ballistic trajectory towards the target. As identified in a now released CIA report from 1957, these factors were significant, and the US advantage in mapping the earth’s true shape and gravitational variations permitted the development of smaller warheads and thus more potent nuclear capabilities. More recently, a failure to account for the quality and format of terrain data may have resulted in an embarrassing miss for an Indian precision strike; a fact that nation has perhaps now attempted to address. The collection and exploitation of data on battlespace characteristics thus confers a significant advantage now and into the future.

In summary, there are considerable tactical advantages to be derived from high quality datasets on the battlespace. Australia should exploit these advantages by ensuring that a comprehensive approach to IMD includes the collection and exploitation of this data by all elements of the future force.