King Kong QA 26 May 2022.pdf

Cognitive Electronic Warfare (EW) Software Development for In- Flight, In-Field, and In-Squadron Applications (King Kong) Call 01

Questions (Q) and Answers (A) 26 May 2022

CALL 01 * King Kong SPECIFIC QUESTIONS

Solicitation Q&As

FA8650-22-S-1054

1) Q: In section 5: Other Information, in the King Kong Solicitation lists that both GFP and GFI are not anticipated under Call 1 – however, in the SOO multiple objectives imply Government furnished information

(Ex. SOO Section 4.1 lists several GFI algorithm sources etc.). Is it the intention of the Government to support GFI/GFE requests under the Kaiju Call 1 effort?

A: Call 01 will include GFI/GFE (Call 01 Solicitation has been amended). This will include both data, software infrastructure, description documents, and data. Contractors will only utilize GFI/GFE data exclusively. The awardee will not be expected to create their own data. If additional data is needed, the contractor can request new data to be made to meet program objectives.

2) Q: In Para I.5.c. of the Call 01 Solicitation: Is this work primarily intended to be performed at government facilities?

A: It is expected that work will be performed at both government and contractor facilities, as stated in the Call 01 Statement of Objectives (SOO)There will likely be a fluctuation of where work is performed based on demonstrations, experiments, etc.

3) Q: In section 1.0: Background and Motivation of the King Kong SOO, please elaborate on the “DevOps software infrastructure” and “open architecture” within the current AFRL portfolio

A: AFRL’s ARCEM lab established a DevOps environment that houses all of our algorithms, analysis scripts, and software processes. All future software developed under Call 01 would need to comply with ARCEM’s documented software standards and interfaces. Call 01 will be adopting the Big Iron standard for all software developed.

4) Q: In section 2.0: Primary Objective of the King Kong SOO, the stated objective is to “perform both applied and advanced R&D…to aid platform cognitive capability and next sortie reprogramming”. Should the delivered product feature the ability to “perform both applied and advanced R&D….” to enable the customer to execute this process beyond the POP?

A: Yes. All 6.2 and 6.3 R&D will be developed and delivered to be executed as a formal process to aid in platform reprogramming.

5) Section 4.1: Advanced Cognitive EW AI/ML Concepts:

Q: The section describes several efforts, including “closed loop software component control and tuning”. We believe this could be interpreted in several ways, so additional elaboration would be appreciated.

A: Closed loop meaning that a software component could alter its response based on new input data. This could simply be achieved via expert system rules or by machine learning. The intent is to have flexible, data-driven processing – i.e., software can be tuned online to maximize performance goals.

Q: Typically, when we talk about “closed loop” we think of Electronic Warfare (EW) agents under test that progress through a cycle of emitting, sensing, reacting, etc. We could extend this to include learning, which then results in tuning. Is this the context in which we should understand “closed loop software component control and tuning”?

A: Yes.

Q: Performance comparison has been cited throughout the sections as well as human/machine learning trust.

Will the government provide any truth data in terms of sample recorded data (e.g. PDWs, I/Q, EDWs)?

A: Yes, reference Kaiju Call 01 (FA8650-22-S-1054) Amendment 02.

6) Section 4.2: Software-Centric System Design

Q: This section specifically designates Big Iron as a foundational element of all development. This clearly applies to any apps required to “mimic EW system OFP within a controlled environment”. Does this hold for all software developed in support of this effort? If tangential utilities are created to aid in the development/testing of the aforementioned EW apps, do those utilities (which are not EW apps) also need to be Big Iron compliant?

A: Yes. All software will be Big Iron compliant.

Q: Does AFRL possess a Big Iron framework for missionware generation?

A: Yes, a baseline framework exists. Concurrent development on the framework will occur to keep compliance with the Big Iron standard as well as specific architectural needs for reprogramming applications.

Q: Is there a specific aircraft platform being targeted?

A: Call 01 is focusing on tactical platforms.

7) Q: In section 4.3, does the Government intend for contractor to utilize existing instantiation of these tools at

AFRL facilities or expect them to be available at contractor facilities?

A: All software tools can be utilized at both government and contractor facilities.

8) 4.4 Data Collection and Data Generation

Q: Is it the Government’s intention to provide existing data collects (I/Q, PDWs etc.) as source data in support of Kaiju Call 1?

Q: Is it the Government’s intention to provide synthetic data sources (emitter models etc.) in support of Kaiju

Call 1?

Q: Is it the Government’s intention to provide reference mission data (mission data files etc.) in support of

Kaiju Call 1?

9) Q: In Section 4.5: Data Analysis Tools and “Big Data” Community Collaboration of the King Kong SOO, the section reads as follows: “Consolidate community tools for data analysis, verification, and data sharing. To accomplish this (my emphasis), advanced receiver models are needed to comb over output data.” Please elaborate on the intended role of advanced receiver models with respect to the task of consolidating community tools. Is the Government’s intention with this objective to build a capability to compare and verify similarity of disparate sets of collected data (simulated vs. real, real vs. real etc.) and/or similarity score simulated or collected data against a known truth data set?

A: Yes. Recorded data and synthetic data are colored by receiver processing (e.g., channelization architecture, adjacent channel logic, bandwidths). Downstream processes are affected by the fluctuation in error profiles imparted on measurements. These error profiles must be analyzed and accounted for as part of this effort with the eventual goal of automating aspects of sensor engineering.

10) Q: Para 6.0: Hardware and software deliverables are very broadly described. Fixtures could range from tens of dollars to hundreds of thousands. Software deliverables could range from small utilities to millions lines of code. Is the intent here to capture the type of work that the contractor would be performing and not to use the description for pricing?

A: Correct, the intent is for vendors to propose to the type of work to be performed.

11) Q: Given the competitive nature of Call 01 - does the requirement to submit subcontractor cost and pricing data still apply?

A: Yes, Subcontractors are required to submit their sanitized proposal (cost and pricing data) directly to the prime to be incorporated into the prime contractors’ proposal submission. If the subcontractor would like to provide an un-sanitized copy to the Government, they may do so by requesting a DoD Safe Drop Off from the

Contractual POC listed within the solicitation, however this is not required.