DARPA Offers Advanced Planning System to the Air Force
A command and control tool is under evaluation by the service’s Kessel Run Laboratory.
Air Force officers in charge of creating air tasking orders have long developed mission plans at air operations centers, known as AOCs, or centralized hubs in a specific command. The Air Force is looking at diversifying and decentralizing how and where those plans are created to add depth and resiliency to the process. This may be needed as designing air battle plans against potential peer threats will only grow in complexity in the future, experts say.
To ease this process, the Defense Advanced Research Projects Agency, or DARPA, has created a distributed planning software tool, the Resilient Synchronized Planning and Assessment for the Contested Environment, known as RSPACE, which includes two classes of decision aids: a planning tool and an execution monitoring application.
RSPACE presents a different way to orchestrate mission planning and evaluation, spreading out planning roles from the centralized AOC model, explains Craig Lawrence, outgoing program manager in DARPA’s Strategic Technology Office. Lawrence is heading to the University of Maryland’s Applied Research Laboratory for Intelligence and Security, as deputy director for systems research.
The command and control tool is designed to overlay existing Air Force software programs, giving operational planners distributed, scalable and automated planning capabilities, he says.
Before developing RSPACE, Lawrence and other officials at DARPA spoke to leaders at Air Combat Command; Pacific Air Forces (PACAF) Joint Base Pearl Harbor–Hickam, Hawaii; and the Life Cycle Management Center (LCMC) at Hanscom Air Force Base, Massachusetts, and found that a common concern was that the AOCs are a single point of failure.
“It’s one big ops [operations] center with potentially up to a couple of thousand people trying to manage an air battle,” Lawrence says. “But if comms [communications] are cut off, or if there’s a cyber attack or any other kind of attack, they might lose continuity of operations.”
Moreover, U.S. adversaries know that the planning is being done at centralized AOCs. “For instance, in the Pacific theater, INDOPACOM [Indo-Pacific Command], one of the AOCs is at the 613th in Hawaii at Hickam Air Force Base, and although it is kind of far from the fight, that’s where everything is managed, and [adversaries] know that,” he continues. “That’s the way the [planning] architecture works now. So the service is trying to look at alternative architectures to use when needed. They still would rather do it centralized. That’s what they’re trained to do. But when needed, the Air Force can push that planning out and be able to do pieces of it at other locations all throughout the battlespace, even at wings that are closer to the fight.”
To enable a more distributed process, RSPACE coordinates across the service’s network and allows mission planning and assessment not just at a single point, but across various planning operation locations or nodes, Lawrence offers. RSPACE aids duty officers’ day-to-day scheduling and assembly of mission packages, while also providing an assessment of the mission packages with replanning and re-tasking capabilities. And underneath, DARPA’s middleware allows RSPACE to overlay existing Air Force systems.
Meanwhile, the execution monitoring decision aid monitors simultaneous activities and compares them to the designed plan to evaluate how a commander’s orders are followed. “RSPACE [works] in real time as things are actually happening, as plans are being executed, and it monitors what’s happening,” he says. “It does some analysis of how well am I following plans—usually not well—and then how do I correct that if I need to, or throw alerts, or identify if there’s perhaps a cascading effect that might have implications on future missions, and then potentially do repairs.”
Additionally, DARPA officials wanted to provide automation in the planning tool, given the complexities planners have to take into account when developing a mission plan, such as thousands of sorties a day, along with coordinated strike assets, tanker support for refueling, suppression of enemy air defenses, electronic warfare support or intelligence, surveillance and reconnaissance (ISR).
DARPA also built automation into the execution monitoring tool, to harness real-time tactical data feeds coming in from the missions and digitally assess that information. For that, it was important to strike a balance between the human-computer roles, incorporating effective automation and more capabilities for the computer to perform, while keeping the creative and insightful human a part of the process, he shares.
“What we tried very hard to do is to give those planners a tool where they could direct the planning process, but let the automation do as much as possible to make it easier on them, and do it in a way that they still understand the plan that they’re building, and they can still direct what they have,” Lawrence relays. “It’s not just ‘press the easy button’ and out comes your plan, because they never end up liking that,” he said of the airmen DARPA worked with.
The agency also worked with several industry partners to develop RSPACE. Northrop Grumman Mission Systems was the integrator. Vern Boyle, vice president of the company’s advanced technologies, explains that for their role, getting input from airmen was important to the development operations, or DevOps, environment. “[Our] focus [in] the integrator role is to provide a DevOps environment for rapid prototyping with continuous integration, continuous deployment and continuous testing to shorten development lead times and enhance software quality,” Boyle states. The company also harnessed two software integration labs hosting an Amazon Web Services region running on Stratoscale Symphony—a converged cloud product—that supports testing and demonstration of capabilities program partners develop.
To provide resiliency to RSPACE users, the company employed techniques to ensure system and data availability in a distributed architecture even when users are faced with a highly constrained, disrupted or disconnected communications channels.
“It accomplishes this by treating data as an enterprise resource, rather than an application-specific resource, and manages data according to enterprise needs and priorities, ensuring distribution of data uses the available communications resources in accordance with the criticality of the missions supported,” Boyle clarifies. It allows the data distribution of both legacy and emerging applications and avoids having to rely on hosting legacy applications at all of the sites where the data is required to perform enterprise functions, and it does so without modification to legacy applications, he adds.
Meanwhile, BAE’s Autonomy Controls and Estimation Group focused on the assessment functionality of RSPACE and created software called the Distributed, Interactive, Command-and-Control Tool (DIRECT), says Michael Schneider, chief scientist, who led the company’s contribution to RSPACE. DIRECT provides a visual interface for operators and generates real-time alerts so that they can evaluate areas of concern during the planning and execution of a mission. To improve mission continuity, it also automatically adjusts to minimize bandwidth when communications are limited, he suggests.
“BAE’s technology takes an operational air plan and projects out under that plan what the Air Force would expect to see in terms of where aircraft are going to be, what their status is, in terms of fuel and weapons consumption, and what issues there might be in terms of how they might be interacting with different threat systems,” Schneider says. “We can continuously update our projections of the plan from the current state and re-ingest data from tactical data links that indicate the current state of platforms, and then we update what our projections are going forward.” The tool compares real-time activities to the original plan, highlights any deviations and allows users to see where the mission is going well and where it might need adjustments, he notes.
A post-mission assessment feature of the software gives planners insights at the end of an operation on which tasks and mission objectives laid out by the commander have been met. “It also allows users to review what’s happened in the plan in terms of going backwards, looking at where things went off plan and understanding what the circumstances were when that happened,” Schneider states.
And while RSPACE can enable distributed planning, Lawrence confirms that the Air Force would have to restructure its guidance for any alternative decentralized planning process. It is a tactics, techniques and procedures, and concepts of operation, matter for the Air Force to examine.“It is how you carve that out, and who is responsible for what,” he suggests.
As part of the RSPACE effort, DARPA explored with the service the idea of functional decomposition where “one node might be responsible for planning a lot of the strike missions or ISR, and another is responsible for tankers and EW support,” Lawrence says. “Or it could be geographic decompositions, where maybe you carve out the battlespace into regions. So we’ve tried to be consistent with what they’ve been thinking about in how they might distribute these types of operations.”
In August, DARPA and the industry partners will get to test the capabilities of RSPACE in an operational setting at one of the service’s facilities as part of a demonstration exercise. Then the RSPACE program will end in September, with completion of the third phase of the project, integration and demonstration, Lawrence says.
As such, DARPA is looking to transition the tool to the Air Force. But to be picked up as a program of record, RSPACE has to be selected by the Air Force’s Kessel Run Experimentation Laboratory, he explains.
Coincidently, RSPACE is one of the first projects to run through the laboratory’s evaluation process, since the Air Force stood up the new software innovation center last May in Boston with the aim of getting information technology out to the service faster.
“The Air Force declared a year or so ago that the path into the Air Operations Center–Weapons System (AOC–WS) program of record is through Kessel Run,” Lawrence says. “This evolved as we were working on the program. It’s been a work in progress, and we’ve been working very closely with them.”
For their first application, DARPA submitted to Kessel Run the execution monitoring and assessment part of RSPACE that was developed by BAE. For that, RSPACE is a bit of a test case for the lab, Lawrence observes. “It is one of their first [projects] that they are actually evaluating from the [science and technology] community to come into the program of record through the Kessel Run process.” Officials at Kessel Run have spun up a team to examine the capability over several months to see if the Air Force could incorporate it into planning operations.
“We’re in the door at Kessel Run and they are evaluating us, and no promises until they have fully evaluated it,” Lawrence states. “And it’s very important for us to do well on this capstone because we want to push the other capabilities over the finish line as well.”
BAE’s Schneider adds, “I think one of the main things is that we’ve put together a tool that’s really focused on supporting the role of the duty officer and the combat operations division, and RSPACE compliments a lot of the existing technology that’s provided to their operations center.”