«ARL-SR-0327 ● JUNE 2015 US Army Research Laboratory Visualizing the Tactical Ground Battlefield in the Year 2050: Workshop Report by Alexander ...»
Included among the “existence proof” are the following: application-to-application autonomous sharing is a current feature of groups of applications; learning applications and machine-assisted learning also exist; and the “intelligence” that enables “plug and play” is a forerunner of self-organization. These capabilities are widely accepted and even sought out by many consumers. In addition, selforganization is necessary in order to de-conflict the many entities that will share the battlespace of 2050.
To counter the effectiveness of distributed, self-organizing, and self-synchronizing collectives, adversaries could inject “moles” that behave in mischievous ways and attack cohesion and trust. Attacks on communications and processing capabilities are also an attractive option. The possible counters to a distributed adversary include preventing the development of shared awareness by sowing distrust to reduce or prevent information-sharing. Counters to attempts to prevent accurate shared awareness from being developed or maintained include the establishment of tamper-proof identification and embedded information provenance.
2.2.6 Cognitive Modeling of the Opponent A vastly improved capability to understand an opponent and predict their actions will enable a new and potentially disruptive capability in this time frame. Directed at both populations and key adversary decision makers on an individual basis, this targeting capability is based upon an understanding of population and individual motivations, biases, cognitive processes, and decision-making styles. In addition, physical and physiological states will be capable of being known. In terms of both individuals and populations, it will be possible to sense their moods and whether or not they are vulnerable to deception or primed to act in a certain manner (resist or be passive).
In addition to the enablers of micro-targeting previously discussed that make this capability likely, by 2050, sensors of various kinds will be ubiquitous and include sensors on and inside humans that can provide the information to support individual, dynamic cognitive modeling (physical state has an impact on cognitive abilities and processing). In addition to having the information available to vastly improve individual cognitive modeling, such models offer the opportunity to disrupt adversary organizations and operations in a cost-effective manner when compared to existing capabilities. Development of these models is already underway as business are trying to better understand us as consumers and microtarget us with advertisements. This “neuro-marketing” is getting more sophisticated every day.
Counters are difficult as we are constantly creating data with every movement we make and every product we research and purchase. However, if key individuals can build a public profile that is misleading, adversaries will not have accurate information to use and their models will be incorrect. Even by 2050, cognitive modelling will be relatively resource intensive, and thus, be limited to a small number of key individuals and somewhat homogeneous populations. Organizations that delegate decision rights more widely could spread micro-targeting resources too thin and make an adversary turn to the more complex problem of understanding and predicting collective behaviors.
Counters to actions that make it more difficult to identify and model key individuals and population centers include being able to react in near real time, reducing the need to predict.
2.2.7 Ability to Understand and Cope in a Contested, Imperfect, Information Environment After a discussion of all of the improvements that could be expected in the quality of information, the workshop participants came to the conclusion that 2050 would not see the realization of the long-heralded era of perfect information. The attributes associated with the quality of information in this discussion included correctness, completeness, relevance, timeliness, precision, authenticity, secureness, uncompromisability, availability, trustedness, and ease of use. Participants concluded that on the battlefield of 2050 there would still be a significant amount of noise mixed in with information. Therefore, it would remain difficult to extract key information and identify misinformation, as well as identify unverified, unattributed, unsourced, and incorrect and/or out-of-date information.
What would be “new” would be individuals’ (and organizations’) ability to cope with an imperfect information environment and the ability to extract value in the form of actionable information. A major aspect of this new capability would be the ability to more accurately understand the imperfections and risks associated with the available information. Put another way, individuals would be far better consumers of information that is currently the case.
The participants felt this development was likely for a number of reasons. First, between now and 2050, individuals will have been born into this challenging information environment and, of necessity, will have developed information survival skills. Second, there will be increased metadata tagging. Third, the multiplicity of sources assisted by information processing agents will help in verifying and cross-checking sources. Fourth, visualization tools will be vastly improved and help individuals deal with increased information. Fifth, organizations and processes will be more distributed and collaborative and thus be better able to bring a diverse set of eyes to the information.
The counters to this include more effective misinformation campaigns and attacks on information sources, processing, and communications capabilities.
The counters to these counters involve more training and better tools, as well as better defenses against information-related attacks.
2.2.8 Other Observations While most of the discussion involved 1 of the 7 developments discussed above,
the participants made a number of additional observations:
Humans will take on a number of different roles vis a vis automated systems • to include not only the traditional role of “human-in-the-loop,” but also a new role “human-on-the-loop.’ Micro-management will be rare, as mission command will be fully • understood and practiced.
It will be possible, at times, to take control of local or even adversary • communications infrastructure and sensors.
Situational awareness will include an understanding of the state of both • friendly, local, and adversary networks and present a dynamic picture that can be used for a variety of purposes
3. Warfare in 2050: Moving, Surviving, Effecting, and Sustaining
This section summarizes the workshop discussions and presents related findings with respect to what we called “moving, shooting, effecting, and sustaining.” Included in these are the battlefield capabilities of tactical maneuver, delivery of a diverse set of effects (lethal and non-lethal, kinetic and non-kinetic, and information and cyber-related), force protection, and logistics.
3.1 Point of Departure for Discussion Workshop participants initiated their discussion of “moving, shooting, effecting,
and sustaining” by considering the following questions:
Will the trend toward highly intelligent weapon systems and munitions • continue? Perhaps creating swarms of robo-munitions?
Perhaps this trend would result in a convergence of robots and munitions: a • smart killer that executes an assigned attack autonomously or in collaboration with distant human controller; these can be ground-mobile, air-mobile, or mixed-mode?
Would the force consist largely of robotic shooters (vs. humans)?
• Would there be few humans manning platforms (e.g., tank with just 1 • human pilot or 1 human to handle multiple tanks)?
Will many (most?) of these robo-munitions be dedicated to hunting and • killing the opponent's robo-munitions?
What would this say about the required level of computational intelligence?
• Collaborative intelligence? Is it feasible by 2050? What would be feasible?
Will such proliferation of machine intelligence in weapons and munitions • engender massive use of electronic and cyber countermeasures? What forms might it take?
Will the role of human Soldier change? How would Soldiers survive in this • environment? What would be nature of their C2 or interactions with these munitions, counter-munitions, and electro-cyber weapons?
If such “see” and “shoot” capabilities emerge, how would anything/ • anybody move and survive on the battlefield? Would this imply short jumps between prepared strong points? A new era of trench warfare? Or mass use of stealth and deception in mobility, e.g., with informational smoke-screen?
Perhaps hundreds of simulated movers, decoys, for each real one?
3.2 Major Aspects of the Tactical Battlefield of 2050 The discussion of what major changes we could expect with respect to our ability to move, shoot, effect, and sustain our forces on the tactical battlefield of 2050 was predicated upon a shared view that this battlefield would be characterized by the
For each of these developments, the group offered their reasons why they felt that these potential transformative capabilities would be found on the tactical battlefield of 2050 and discussed the ways that adversaries could counter or mitigate the effectiveness of these capabilities, as well as counters to these countermeasures.
3.2.1 Ubiquitous Robots Workshop participants envisioned the battlefield of 2050 as being populated by large numbers of autonomous entities of all kinds. These entities were referred to by participants as simply “robots.” Many of these robots would be fairly similar to the systems that exist today, such as unattended ground sensors, small unmanned aerial vehicles (UAVs), and fire-and-forget missiles. However, their 2050 versions would possess significantly greater capabilities of machine reasoning and intelligent autonomy than those existing today. Participants saw these robots operating and maneuvering on the battlefield (in the battlespace) in a large variety of ways. They would move over the ground and in the air at low altitudes. Some of these would have locomotion that was bio-inspired. These robots would range in size from insect-sized entities to robotic vehicles capable of transporting a team of humans. They would also be “virtual” and able to navigate and “act” in cyberspace.
The collection of battlefield robots would be robustly networked and capable of communicating and collaborating with one another, with a variety of systems, and with humans.
Many of these robots would have ISR-related roles with many (perhaps most) of them possessing autonomous sensors that would provide nearly continuous coverage of every inch of the battlefield. Other robots would act as intelligent, single-use munitions. These could operate in “teams,” like wolf packs of fire-andforget missiles and ground-crawling or jumping intelligent mines. Some of these robots would be employed in cyber/network defense, including defending electronic components resident on/in a human; serving as intelligent defense assistants able to prevent or warn about incoming threats, or acting in the role of advisors in complex decision-making tasks, such as performing a detailed course of action analysis prepared for local conditions observed in real time. These deployed robots would be capable of operating in a variety of “control” modes from total autonomy to active management by humans.
Adversaries could counter the effectiveness and utility of robots in a number of ways. An adversary could create physical obstacles to mobility and other robot capabilities. These particular counters could be mitigated by deploying a heterogeneous collection of robots with diverse forms of locomotion or providing a given robot with multiple modes of adaptive locomotion.
Another counter to battlefield robots involves engaging robots with conventional kinetic weapons or DEWs. To counter this counter, robots could be hardened in a variety of ways from the mundane to the very sophisticated with more sophisticated defensive measures involving the protection of high-value robots with a “force field” (discussed later in this section of the report).
Another counter involves attacks against robot power sources. To counteract such attacks, these power sources could be spatially dispersed, designed to have backup alternative sources, involve wireless delivery of power, or have organic power sources. In the case of attacks on power sources, another counter would involve the robot’s ability to detect the attack and adjust its behaviors in a variety of ways that reduce its power requirements.
Other counters to robots may be cognitive in nature. For example, spoof attacks that misdirect, degrade, or take control of a robot. To prevent spoofing attacks, robots could be deployed with bio- or behavior-based authentication. Another type of cognitive attack on robots would involve eroding trust, either the trust the human controller has in the robot or a robot’s trust in information and orders received.
These sorts of attacks could be mitigated or avoided by employing forensic introspection of robot behavior.
3.2.2 Swarms and Teams Robots will commonly operate in teams or swarms in the battlespace of 2050 in the same way Soldiers act in teams today. These self-organized and/or collaborative collections of robots would operate with varying degrees of freedoms (from being actively managed to being autonomous) under dynamically established rules of engagement/priorities. Robot swarms and teams (as well as individual robots) would be assigned a variety of tasks. For example, as independent attack forces or as part of an orchestrated attack using a variety of weapons, as a collective defensive shield, and as a sensing field. Among the less obvious roles for a robot team, the participants of the workshop envisioned a team of robots warning civilians (e.g., in a battle raging in a mega-city environment) to keep away from dangerous areas and even acting as a defensive shield for the civilians against any stray projectiles.