SAM Public Posting DTC JA Virtual Testbed V9 Redacted Revised v2_Redacted.pdf

JUSTIFICATION AND APPROVAL

FOR OTHER THAN FULL AND OPEN COMPETITION

SUBJECT: Request for Approval of Justification for Other than Full and Open Competition for the DARPA Triage Challenge Virtual Testbed

PART I – JUSTIFICATION

1. Agency and Contracting Activity Identification: This Justification and Approval for other than full and open competition is executed by the Defense Advanced Research Projects Agency (DARPA), Contracts Management Office (CMO).

2. Nature and/or Description of Action: DARPA proposes to award a sole source contract to Applied Research Associates (ARA) of Orlando, Florida for development, delivery, and support of a high-fidelity open-source medical triage virtual testbed for use by the organizers and competitors of the DARPA Triage Challenge, supporting technology development efforts as well as the execution of the virtual and real-world challenge events scheduled between September 2024 and December 2026. The challenge kickoff is expected in Fall 2023. The targeted period of performance start date is July 2023 through December 2026.

3. Description of Required Supplies/Services: The DARPA Triage Challenge (DTC) aims to drive breakthrough innovations in identification of “signatures” of injury that will help medical responders perform scalable, timely, and accurate triage. Of particular interest are mass casualty incidents (MCIs), in both civilian and military settings, when medical resources are limited relative to the need. To advance progress towards this vision, the DTC will bring together multi-disciplinary teams and industries that will identify physiological signatures and develop sensor and algorithm strategies for complex MCI settings. The DTC includes a Primary Triage Competition (injury signatures from stand-off sensors on robotic platforms) and a Secondary Triage Competition (injury signatures from non-invasive sensor data). Teams participating in the DTC will be tasked with developing and demonstrating strategies for capturing high-value signatures for either primary or secondary triage, or for both.

The DTC virtual platform will support this goal by providing a resource for training, testing, and evaluating algorithms in simulated mass casualty scenarios. The contractor will be required to develop a virtual testbed capable of simulating large MCIs for primary triage scenarios while supporting numerous high-fidelity casualty models, diverse sensor phenomenology, and multiple heterogeneous autonomous agents (e.g., aerial and/or ground robots). Because this virtual testbed is required to support the execution of the challenge events of the DTC, the contractor must clearly demonstrate a high degree of maturity and fidelity in their existing capability with the ability to rapidly extend the performance and features of the virtual testbed to meet the needs of the challenge. The first delivery of a functional capability is required within 3 months of contract start to meet the requirement to provide competitors in the DTC a working virtual testbed for development and testing.

The primary objective for the virtual testbed and virtual challenge is to develop algorithms that will provide scalable, timely, and accurate triage in real-world environments. In order for the algorithms to translate from the virtual environment to the real world, the maturity and fidelity of the virtual testbed is of critical importance. Every aspect of the virtual testbed is required to meet this high bar of realism including the casualty models, sensor models, robot models, and environment models. In particular, the testbed must include an extensive library of diverse and high-fidelity casualty models that are integrated with a dynamic, time variant human physiology engine capable of modeling realistic casualties, including variation in type and severity of injuries.

In addition to these strict requirements, the testbed is further required to be capable of running in real-time on a standard desktop computer as well as faster-than-real-time when run without a front end..

The contractor must have prior experience developing and maintaining software on an open-source basis, creating code and support documentation including tutorials and user guides, and supporting competitor engagement with a broad community of researchers. The contractor will make the virtual testbed available on an open-source basis, with no additional paid tiers or features, and will host the simulation so that participants in the DTC can access and utilize it freely.

Specifically required features and capabilities of the open-source virtual testbed include the following:

• Fully featured virtual testbed package with no additional paid tiers or features.

• Open-source, extensible architecture that supports third-party development for virtual testbed features, system models, and virtual environments.

• Capable of supporting many simultaneous and/or heterogeneous agents (e.g., aerial and/or ground robots) employing stand-off sensors (e.g., UAS-borne sensors, ground robot-borne sensors, diverse sensor phenomenology).

• Capable of supporting large-scale three-dimensional virtual worlds that realistically represent the real-world challenges of mass casualty incidents, including elements that partially obscure casualties (e.g., debris, buildings, overlapping bodies), complex terrain and surface textures, and additional environmental elements (e.g., low light, smoke/mist, mud, water, wind).

• Capable of supporting a wide range of stand-off sensors including (but not necessarily limited to): RGB camera, audio, lidar, radar, thermal, multi- and hyper-spectral, and infrared.

• Include an extensive library of realistic casualty models, including variation in type and severity of injuries with variation in pose, skin tone, sex, size, and age.

• Capable of modeling multi-modal stand-off sensor capture of casualty features.

• Uses physics-based models of inertia, actuation, contact, and environment dynamics to simulate agent’s motion and interactions.

• Supports multiple physics engines capable of both maximal-coordinate and reduced-coordinate representations.

• Supports the Linux operating system, the Robotics Operating System (ROS), and other common robotics programming languages and robot control middleware.

7. Determination of Fair and Reasonable Costs: Prior to contract award, the Contracting Officer will conduct a detailed cost and price analysis of ARA’s final cost proposal in accordance with FAR Subpart 15.4 to ensure the proposed price to the Government is fair and reasonable.

8. Market Survey: Prior to issuance of the Broad Agency Announcement (BAA) for the DARPA Triage Challenge program on November 17, 2022, informal market research was conducted by DARPA’s Biological Technologies Office personnel by contacting other DARPA Program Managers, Program Managers in other government agencies, and the medical simulation community to identify existing companies that could potentially provide the virtual testbed requirements described herein.

DARPA also conducted a more formal market survey and documented it in writing. The survey methodology included literature reviews supplemented by informal discussions with organizations that have medical, casualty, and robotics simulation capabilities. Because DARPA has sponsored a wide variety of development in these areas, the agency is well aware of the availability of technical capabilities, experience, and resources present in the market to fulfill the objectives of this effort.

Of the capabilities statements that DARPA has reviewed, the survey found that Applied Research Associates is the only source capable of meeting the requirement of the DTC virtual testbed. No alternative solution exists with the open-source software infrastructure, level of maturity, and physiology engine capabilities as demonstrated by Applied Research Associates.

As stated in Paragraph 6 above, no sources expressed interest in response to DARPA’s notice of intent to sole source posted on March 13, 2023.

9. Other Facts Supporting the Use of Other Than Full and Open Competition: Not applicable

10. Written Interest in Acquisition: None. See Paragraph 6 above.

11. Actions to Overcome Barriers to Competition: This effort is for Applied Research Associates to deliver an open-source virtual testbed for use by the DARPA Triage Challenge and will not be a repetitive acquisition. In addition, the open-source virtual testbed, once developed and delivered to DARPA, inclusive of software, data, and documentation developed by Applied Research Associates for this project will be made available with no limitations, which will allow anyone to use the code in any manner, including in closed-source commercial and other government products.

12. Delivery Requirements: Applied Research Associates will perform the tasks identified in Paragraph 3 above. Additionally, ARA will deliver milestone reports required to receive payment, and submit quarterly technical summaries during the period of performance as well as a final technical report.