Wednesday, August 19, 2015

Advancing Autonomous Systems: Rough Seas Ahead for Command & Control

by Prof Mark Nissen, NPS, mnissen(at)nps.edu

Command & control (C2)[1] is quintessentially important to military endeavors. As Joint Publication 6-0[2] elaborates authoritatively (I-1): “Effective C2 is vital for proper integration and employment of capabilities.” Further, our contemporary and informed understanding of C2 indicates that it applies to much more than just the technologic underpinnings of command and control systems. As Naval Doctrine Publication 6[3] reinforces: “… technology has broadened the scope and increased the complexity of command and control, but its [C2] foundations remain constant: professional leadership, competence born of a high level of training, flexibility in organization and equipment, and cohesive doctrine.”

Joint Publication 6-0 expounds (I-2): “Although families of hardware are often referred to as systems, the C2 system is more than simply equipment. High-quality equipment and advanced technology do not guarantee adequate communications or effective C2. Both start with well-trained and qualified people supported by an effective guiding philosophy and procedures.” Indeed, the first element of C2 is people: “Human beings—from the senior commander framing a strategic concept to a junior Service member calling in a situation report—are integral components of the C2 system and not merely users.” This concept is embedded deeply within our research, engineering, leadership, command, control and operation of manned systems (e.g., airplanes, ships, networks), forces and operations.

In contrast, however, a great many researchers, engineers, leaders, commanders, controllers and operators of autonomous systems (e.g., unmanned vehicles, robots, cyber applications) concentrate principally—if not exclusively—on technology, paying scant attention to the people, processes and organizations required for command, control and mission efficacy. This leaves a dearth—if any—residual attention to C2 of autonomous systems today. Given the quintessential importance of C2, such technologic focus is problematic, particularly where large-scale, joint or coalition operations are considered.

Moreover, today’s problems portend to become tomorrow’s vulnerabilities. As our research looks five to ten years into the future, not only do we foresee ever increasing technologic advancement of autonomous systems, but we preview it combined with ever increasing integration of manned and unmanned systems. In a great many circumstances, teams of autonomous systems and people (TASP) will become the norm, with both manned and unmanned systems and operators integrating their complementary attributes and capabilities for outcomes more effective and successful than possible through either manned or unmanned alone.

Especially together, the technologic advancement of autonomous systems and the manned-unmanned integration through TASP imply rough seas ahead for C2. Using the state-of-the-art simulation system POWer [4]  we see, for instance, how current C2 organizations and approaches strain already with multiple unmanned aircraft systems (UAS) in common airspace. Consider a joint task force (JTF) environment today, for example, say with only two UAS launched from different ships. Who is the lowest level person in the JTF organization with authority over both unmanned aircraft? It is probably the CTG, who may not even be onboard either ship, which signals a serious issue.

Now exacerbate this issue with multiple UAS— perhaps even a swarm—from multiple ships and shore facilities, flying in common airspace. Then exacerbate it still further with multiple UAS (maybe even operated and controlled by a set of diverse coalition partners) flying in common airspace with manned systems. Far from just the physical control issues (e.g., collision avoidance), how do we integrate manned and unmanned systems and missions to leverage their complementary attributes and capabilities? How do we institute and optimize joint manned-unmanned training? How can we expect for the different people and machines from manned and unmanned squadrons to cohere seamlessly when an integrated mission begins? How, when and to which levels do we delegate TASP mission authority and control? What are the major impediments to effective TASP missions, and what should we be doing now to prepare for and overcome them? These are all important C2 questions that we will need to answer well in advance of TASP missions becoming commonplace in the coming half decade.

Through CRUSER sponsorship and guidance, our POWer[5] research is beginning to answer some of these questions. Examining UAS in use today within the CTG mission environment, as an important place to begin, we’ve identified many troublesome C2 problems already. For one instance, the C2 organization reflects a tall, functional hierarchy, with considerable centralization, substantial formalization and frequent staff rotation. This makes for relatively long information flows and decision chains, coupled with perennial battles against knowledge loss from personnel turnover and challenges with cross-functional (and even more so with joint and coalition) interaction. Many organization experts would argue that the correspondingly long decision chains, information flows and staffing turbulence militate against efficient—or even effective—C2.

As another instance, the formalization inherent within this C2 organization reflects strong dependence upon written standards, rules and procedures (e.g., SOPs, TTPs, PPRs, work standards, job qualifications). However, the continuing technological advance and integration of UAS suggests that formalization through written documents may have a hard time keeping up with rapid and local knowledge onboard various ships and across diverse crews. Although this is a knowledge management problem, as with the long decision chains, information flows and staffing turbulence noted above, many organization experts would argue here that the correspondingly high dependence upon standardization and written documentation militate against efficient—or even effective—C2.

Moreover—and perhaps somewhat counter intuitively—for many years to come, unmanned missions will likely require more planning, monitoring, intervening and like control activities than their manned counterparts. Hence greater numbers of C2 staff—or more skilled and experienced staff members—will be required for unmanned than for manned missions, and such missions will be expected to take more time, suffer from more mistakes, and generally tax the C2 organization more greatly. (Overall, many unmanned missions are still more economic, but they exact greater demands in terms of C2 coordination load.) This eventuality will exacerbate for integrated manned-unmanned events, particularly as we expand across joint and coalition operations.

As a third instance, problematic issues are highly likely to arise also in terms of different skill levels, lack of common training or co-operational experience, and very low—or no—trust between manned and unmanned aircraft operators. A great many manned and unmanned systems personnel are members of different tribes—with distinct cultures and status—that recruit, train, operate and promote separately for the most part. TASP requires manned and unmanned mission integration, flying together in common airspace, and relying integrally upon one another. Imaging telling a Fleet aviator that he or she will have an unmanned wingman!

Of course the simple solution is to keep manned and unmanned systems separate: in separate organizations, in separate airspaces, with separate skill sets, with separate procedures. Such simple solution negates the integrative power and efficacy of TASP, however. Where teams of autonomous systems and people can be more effective than either manned or unmanned systems alone, an adversary can potentially become victorious with C2 sufficiently advanced for TASP. This can be the case even where the technology of our manned and unmanned systems is superior. In other words, advances in C2 may trump superior technology.

So where do we go from here? Our ongoing research continues to employ POWer to project and analyze the comparative performance of different missions, technology degrees, levels of manned-unmanned integration, and approaches to C2 organization. We compare performance metrics across an array of measures including time for effective mission completion, mission errors and corresponding rework, C2 communication and coordination load, along with mission cost, risk and others.

We also examine UAS across a wide range of technology degrees: from operational UAS in the current inventory, through those undergoing test and evaluation today, to future systems envisioned with performance levels matching—and even surpassing—those achievable only through manned systems today. This enables us to examine a correspondingly wide array of C2 organizations and approaches, mis- sion scenarios, technology degrees and levels of manned-unmanned mission integration, from those taking place in current operations through counterparts likely five to ten years hence.

Further, POWer supports computational experiments that allow us to examine this wide array in a very systematic and precise manner. Changing the level of only one variable at a time—or analyzing suites of level changes across multiple variables simultaneously—we can ascribe resulting performance differences specifically and unambigu- ously to each such level change, and we can explain precisely how each variable—independent, dependent or control—is defined, operationalized and manipulated. This supports exceedingly high reliability and internal validity through our experiments.

Moreover, the cost of computational experiments is exceptionally low, and the speed is exceptionally high, so we can assess hundreds or thousands—even millions—of different scenarios in short periods of time, with no risk of losing valuable equipment or people (e.g., as can occur through lab and field experiments) in the process. This capability equips us to peer well into the future, to make informed decisions, and to take dominating actions, not only regarding which alternate futures to select, but also regarding how to achieve each future in a competitively advantageous way.

In light of the issues identified above, we’re looking in particular at how to address the long decision chains, information flows and staffing turbulence that militate against efficient or effective C2 at present, and we’re concentrating on managing the kinds of fast-changing local knowledge that challenges even the best efforts in terms of written standards, rules and procedures. We’re considering further how to decrease the coordination load on C2 of unmanned systems, and we continue to envision alternate approaches to the integrated recruiting, training, promotion and performance of manned and unmanned operators.

In the near future, we anticipate laying out a list of highly promising, agile approaches to adapting C2 in response to such issues, with a set of milestone markers to signal when each will likely become most appropriate, and a set of plans for how to effect each of them. The idea is to peer sufficiently far into the future so that we can provide leaders, policy makers and technologists today with the time and guidance needed for them to prepare for and navigate the rough seas ahead.

[1] The term C2 as discussed here subsumes and largely replaces the myriad extension of “Cs” (e.g., C3, C3I, C4, C4I, C4ISR, C5I).
[2]  JP6-0, Joint Publication 6-0: Joint Communication System Washington, DC: Joint Chiefs of Staff (2015).
[3]  NDP6, Naval Doctrine Publication 6: Naval Command and Control Washington, DC: Department of the Navy (1995).
[4]  POWer derives from the VDT Group at Stanford and has been tailored and validated to simulate the qualitative and quantitative behaviors of C2 organizations, approaches, personnel and systems.
[5]  See, for example, Nissen, M.E. and Place, W.D., “Computational Experimentation to Understand C2 for Teams of Autonomous Systems and People,” Technical Report NPS-14-007, Naval Postgraduate School, Monterey, CA (December 2014).

Reprinted with permission from the Naval Postgraduate School's CRUSER News. All opinions expressed are those of the respective author or authors and do not represent the official policy or positions of the Naval Postgraduate School, the United States Navy, or any other government entity.