HR001118S0035.pdf

HR001118S0035 SIGMA+ SENSORS 1

Broad Agency Announcement SIGMA+ Sensors

Defense Sciences Office

HR001118S0035

April 5, 2018

HR001118S0035 SIGMA+ SENSORS 2

Table of Contents

PART I: OVERVIEW INFORMATION

PART II: FULL TEXT OF ANNOUNCEMENT

I. Funding Opportunity Description A. Introduction B. Background C. Program Description/Scope D. Program Structure – Sensors Thrust E. Technical Area Descriptions – Sensors Thrust F. Schedule/Milestones G. Deliverables H. Other Program Objectives and Considerations

II. Award Information A. General Award Information B. Fundamental Research

III. Eligibility Information A. Eligible Applicants B. Organizational Conflicts of Interest C. Cost Sharing/Matching

IV. Application and Submission Information A. Address to Request Application Package B. Content and Form of Application Submission C. Submission Dates and Times D. Funding Restrictions E. Other Submission Requirements

V. Application Review Information A. Evaluation Criteria B. Review and Selection Process C. Federal Awardee Performance and Integrity Information (FAPIIS)

VI. Award Administration Information A. Selection Notices B. Administrative and National Policy Requirements C. Reporting

VII. Agency Contacts VIII. Other Information

A. Frequently Asked Questions (FAQs) B. Collaborative Efforts/Teaming

IX. Appendix A

BAA ATTACHMENTS:

A. ABSTRACT SUMMARY SLIDE TEMPLATE

B. ABSTRACT TEMPLATE

C. PROPOSAL SUMMARY SLIDE TEMPLATE

D. PROPOSAL TEMPLATE – VOLUME 1 TECHNICAL & MANAGEMENT VOLUME

E. PROPOSAL TEMPLATE – MILESTONES AND DELIVERABLES TABLE

HR001118S0035 SIGMA+ SENSORS 3

F. PROPOSAL TEMPLATE – VOLUME 2 COST VOLUME

G. PROPOSAL TEMPLATE – COST SUMMARY

H. PROPOSAL TEMPLATE – VOLUME 3 ADMINISTRATIVE & NATIONAL POLICY

REQUIREMENTS VOLUME

I. TEAMING PROFILE TEMPLATE

HR001118S0035 SIGMA+ SENSORS 4

PART I: OVERVIEW INFORMATION

Federal Agency Name: Defense Advanced Research Projects Agency (DARPA), Defense Sciences Office (DSO)

Funding Opportunity Title: SIGMA+ Sensors

Announcement Type: Initial Announcement

Funding Opportunity Number: HR001118S0035

Catalog of Federal Domestic Assistance (CFDA) Number(s): 12.910 Research and Technology Development

Dates (All times listed herein are Eastern Time.)

o Posting Date: April 5, 2018 o Teaming Profile Deadline: April 11, 2018, 4:00 p.m.

o Abstract Due Date: April 18, 2018, 4:00 p.m.

o FAQ Submission Deadline: May 24, 2018, 4:00 p.m. See Section VIII.A.

o Full Proposal Due Date: May 31, 2018, 4:00 p.m.

Anticipated Individual Awards: DARPA anticipates multiple awards

Types of Instruments that May be Awarded: Procurement contracts, cooperative agreements or other transactions

Agency contacts o Technical POCs:

Dr. Vincent Tang, Program Manager, DARPA/DSO – SIGMA+ program lead Dr. Anne Fischer, Program Manager, DARPA/DSO – chemical/explosive sensors lead Col. Matt Hepburn, M.D., Program Manager, DARPA/BTO – biological sensors lead o BAA Email: SigmaPlus@darpa.mil o BAA Mailing Address:

DARPA/DSO

ATTN: HR001118S0035

675 North Randolph Street Arlington, VA 22203-2114 o DARPA/DSO Opportunities Website:

http://www.darpa.mil/work-with-us/opportunities

Teaming Information: See Section VIII.B for information on teaming opportunities.

mailto:SigmaPlus@darpa.mil http://www.darpa.mil/work-with-us/opportunities

HR001118S0035 SIGMA+ SENSORS 5

Frequently Asked Questions (FAQ): FAQs for this solicitation may be viewed on the DARPA/DSO Opportunities Website. See Section VIII.A for further information.

HR001118S0035 SIGMA+ SENSORS 6

PART II: FULL TEXT OF ANNOUNCEMENT

I. Funding Opportunity Description

This Broad Agency Announcement (BAA) constitutes a public notice of a competitive funding opportunity as described in Federal Acquisition Regulation (FAR) 6.102(d)(2) and 35.016 as well as 2 CFR § 200.203. Any resultant negotiations and/or awards will follow all laws and regulations applicable to the specific award instrument(s) available under this BAA, e.g., FAR 15.4 for procurement contracts.

A. Introduction The Defense Sciences Office at the Defense Advanced Research Projects Agency (DARPA) is soliciting innovative research proposals in the area of advanced chemical and biological detector technologies to support a networked system of sensors for the detection, interdiction, and deterrence of clandestine weapons of mass destruction (WMD). Proposed research should investigate innovative approaches that enable revolutionary advances in science, devices, or systems. Specifically excluded is research that primarily results in evolutionary improvements to the existing state of practice.

B. Background The United States faces challenges from increased adversarial access to low-cost enabling technologies for the production and deployment of weapons of mass destruction. Additive manufacturing and small-scale chemical reactors, for example, reduce the infrastructure and economic investment required to develop and manufacture weapons with destructive power disproportionate to their size and weight. These developments increase the probability of covert WMD manufacture and deployment, either for terrorist or rogue state purposes, and reduce the observable signatures required to detect their presence.

The SIGMA program began in 2014 as an effort to significantly advance scalable detection capabilities against radiological and nuclear (RN) WMD threats from these non-traditional, clandestine attack vectors. SIGMA developed thousands of high-capability, low-cost detectors and networked them to demonstrate large-scale, continuously-streaming physical sensor networks for the RN interdiction mission. SIGMA capabilities have been tested and operationalized with federal, state, and international partners. Further information about the SIGMA program can be found at the following links:

https://www.darpa.mil/program/sigma https://www.darpa.mil/news-events/2017-03-01 https://www.darpa.mil/news-events/2016-10-11 https://www.darpa.mil/news-events/2016-08-23

Additionally, the SIGMA network technical capabilities are described in Appendix A.

The SIGMA+ initiative will build on SIGMA’s successes by developing a persistent, real-time, early detection system for the full spectrum of chemical, biological, radiological, nuclear, and explosive (CBRNE) WMD threats at the city-to-region scale. Specific targeted capabilities for https://www.darpa.mil/program/sigma https://www.darpa.mil/news-events/2017-03-01 https://www.darpa.mil/news-events/2016-10-11 https://www.darpa.mil/news-events/2016-08-23

HR001118S0035 SIGMA+ SENSORS 7

each threat mode will focus the envisioned sensor network, advanced intelligence analytics, and adversary modeling developments under one shared network backbone infrastructure and a ubiquitous mobile, adaptive sensing strategy.

This BAA focuses on research, development, and scale-up of chemical and biological sensors as well as their associated algorithms for the SIGMA+ system. Responses outside of this focus will be deemed non-responsive. Current plans include a subsequent BAA, anticipated to be released in the first half of the 2019 Government Fiscal Year, to address the developments for automated intelligence analytics and advanced adversary modeling, as well as further developments for the SIGMA network backbone that are expected to be required to perform full fusion of these data and methods.

C. Program Description/Scope SIGMA+ will advance, integrate, and scale emerging sensor and data analytics technologies to demonstrate and transition a detection system that fundamentally changes how we detect, interdict, and deter clandestine WMD. The SIGMA+ initiative will leverage and build on sensor capabilities to enable fully-networked scalable, high-capability detectors in the chemical, biological, and explosive (CBE) threat space, similar to what was accomplished in the RN threat space under the SIGMA program. (Although the SIGMA+ system will address RN threats, new RN sensor capabilities are not solicited in this BAA.) Beyond incorporating these additional sensor modalities, SIGMA+ will fuse CBRNE sensor data with new automated intelligence analysis and other contextual data. Furthermore, advanced social science techniques will be leveraged for adversary modeling and integrated into SIGMA+ to maximize detection and interdiction effectiveness. This holistic development and integration of physical sensing, automated intelligence and contextual data analysis, and advanced adversary modeling will result in a transformative and practical early detection system for the full spectrum of CBRNE WMD threats. Proposers to this BAA should focus only on the CBE sensor network domain; other areas mentioned will be incorporated into the complete system through subsequent SIGMA+ solicitations and integration efforts.

For chemical and explosives threats, the existing SIGMA network will be extended to include scalable chemical detection technologies that enable identification of a broad range of species and precursors at the 10 parts-per-billion (ppb) (or better) level to identify illicit production of harmful threats in complex urban environments. The focus on detecting threat production will help enable interdiction prior to an attack.

For biological threats, SIGMA+ will develop novel methods, either environmental or human-sensing based, for improved real-time detection of attacks. This effort aims to provide days earlier attack detection and geolocation of a much wider range of attacks, enabling more effective countermeasures and mitigation strategies.

For radiological and nuclear threats, the incorporation of large-scale automated intelligence analytics into SIGMA+ will allow prioritization of detections near statistical limits to enable interdiction of heavily shielded threats, increasing effective system sensitivity by up to an order of magnitude.

HR001118S0035 SIGMA+ SENSORS 8

Federal and state partnerships will be established for joint testing and large-scale field trials of SIGMA+ capabilities to enable interagency and operational transition.

One example concept of operations (CONOP) for the final SIGMA+ system, for the case of detecting an illicit lab, is as follows: A system of mobile, automated chemical sensors (potentially leveraging existing vehicle fleets as carriers) will continuously scan the streets and buildings of a wide urban area, reporting results in real time via the SIGMA+ network. The incoming data stream will be continuously assessed against, and will help to build, high-resolution spatiotemporal backgrounds to maximize sensitivity for species of interest (e.g., threat and volatile precursors). Automated intelligence analytics, guided by adversary models and contextual data to help extract the most relevant and actionable intelligence, may also be used to cue and direct the sensor network. In that example case, the chemical sensors could be automatically reconfigured for increased detection sensitivity of a particular set of precursor chemicals and/or re-directed to a prioritized search area for further scanning. Similar scenarios can be imagined for the detection of a biological attack. While these examples represent some of the most challenging of applications, the technical advances of the SIGMA+ program will have broad applicability throughout the DoD.

SIGMA+ will consist of three integrated thrusts: sensors, network, and test and evaluation. The program will be conducted in two 2.5-year (30-month) phases. The initial phase (Phase 1) will focus primarily on sensor, network architecture, and automated analytics development, while in the subsequent phase (Phase 2), the emphasis will shift towards full integration into a holistic detection network. System modeling, test and evaluation, and field trials will occur throughout the program.

The simplified timeline below (Figure 1) summarizes the three thrust and two phases involved, as well as the key program milestones.

Figure 1: Simplified timeline and key milestones of the SIGMA+ initiative.

HR001118S0035 SIGMA+ SENSORS 9

This BAA covers only the sensors thrust, described fully in Sections D and E. There exist, however, dependencies between each of the three thrusts for the construction of the SIGMA+ system. The current SIGMA network backbone capabilities and proposed SIGMA+ network developments are described in sufficient detail below for proposers to understand the system into which their sensors and associated algorithms will be integrated. Appropriate application programming interfaces (APIs) and other interface documentation will be provided to selected performers.

Planned tests and evaluation are also discussed below to provide proposers with sufficient context for the delineated goals. DARPA expects sensors BAA performers and their teams to work with other performers to conduct system-level integration, experiments, and tests with their newly developed detectors. DARPA will work with individual performers to coordinate this in-development testing cycle. Performers should not expect this testing to replace their own internal device testing, and should be prepared to share the results of their internal tests with DARPA-designated performers on the SIGMA+ team.

Summary of Network Thrust This thrust will develop the SIGMA+ network architecture and incorporate the individual sensor, multi-sensor, and system-level algorithms. Beginning in the second year, physical sensors will be integrated into the network, and system algorithms will be developed by the third year.

Additionally, adversary characterization will be combined with automated intelligence analytics algorithms. Further refinement of algorithms and adversary modeling will occur in the remaining years.

Algorithms, network architecture, and user interfaces will be continuously refined throughout the program. The first integrated version of the system (alpha system) will occur at the end of Phase 1, the beta version of the system will be in Year 4, and SIGMA+ version 1.0 will be completed near the end of Phase 2 in Year 5.

The network thrust will be covered by a separate BAA at a later time. Information provided here is purely for guiding sensor proposals covered under this BAA.

The SIGMA+ network backbone is intended to retain capabilities of the SIGMA network backbone while incorporating new features to achieve SIGMA+ objectives. An overview of the baseline SIGMA network backbone capabilities can be found in Appendix A. Proposers are strongly encouraged to review this information in order to understand the capabilities of the current network and how their sensor might integrate into it, and also to clearly state what new network backbone capabilities might be required to fully support their proposed sensor as well as the proposed CONOP.

Summary of the System Modeling, Test and Evaluation, and Field Trials Thrust Work in this thrust will combine extensive government-led testing, red teaming, and integrated system modeling to validate capabilities and provide optimal deployment plans for transition.

System models will be developed and tested in Years 1 and 2, with sensor testing, red team experiments, and small field trials providing validation data. These data sets, along with system model analyses, will be used to evaluate program progress at the end of Year 2.

HR001118S0035 SIGMA+ SENSORS 10

Additionally, this thrust will focus on field trials of the SIGMA+ system and engagement with transition and operational partners. Transition partners will be identified early such that the system hardware, algorithms, and network can be tailored with these partners as exemplar cases throughout the program. As the system scales up in Years 3-5, larger, iterative system field trials will be conducted with federal and local/state operational partners to test, demonstrate, and further refine system capabilities. Field trial data will be combined with red-team experiments and model simulations to produce validated system receiver operating characteristic (ROC) curves for each version of the system. Transition partners will be engaged throughout the program, maximizing the potential for transition and operational deployment of the SIGMA+ system.

This work is not being solicited under this BAA. Information provided here is purely for guiding sensor proposals covered under this BAA.

D. Program Structure – Sensors Thrust This BAA solicits proposals to the sensors thrust, which focuses on developing and incorporating chemical and biological sensors, as well as sensor algorithms, into the SIGMA+ architecture.

Proposers are required to address both the sensor hardware and data processing algorithms that will be needed to meet program metrics.

The first 30 months (Phase 1) of the sensors thrust will focus on the development, optimization and validation of sensors, as well as building and integrating corresponding sensor-level algorithms into a sensor system.

“Sensor system” is defined here as a collection of sensors with algorithms, and its performance within a proposer-defined CONOP should satisfy the system metrics and constraints as described in Section E and the high-level objectives described in Section C. It is anticipated that the advanced sensors developed will require two types of algorithms to meet the system metrics and to maximize the desired CBRNE early detection capabilities discussed in this BAA: algorithms that operate at the individual sensor level, and higher-level algorithms that operate across multiple sensors in order to, for example, localize the emission of detected precursor effluents.

Proposers should make clear if additional contextual data is required for, or would greatly benefit, their detection, identification and tracking algorithms (e.g. real-time weather information, available public health reports of infection rates, etc.); if and how that information will be collected by the proposer’s sensor system or via an external feed; and what nominal requirements might be imposed on the SIGMA+ network to incorporate and process that contextual data. Algorithms for automated intelligence analysis as described earlier are outside the scope of this BAA; proposers should, however, ensure that their sensor algorithm and analysis framework can easily incorporate cues and output stemming from such external analyses. For example, intelligence analysis could indicate that a particular type of chemical threat might be in a particular operating area, and this would result in a desired, automated tuning of the sensor algorithms to be more responsive to that type of threat for any detectors entering that operating area.

HR001118S0035 SIGMA+ SENSORS 11

Proposals must demonstrate that the development of the proposed new detection concepts and technologies can meet the aggressive program timeline for technology maturity for deployment.

Full-scale production and integration of custom chemical and biological sensors will begin in Phase 2, along with development of any additional advanced sensor concepts that are identified to close a capability gap. Unit-level testing of sensors and algorithms will be performed continuously throughout the program in both Phase 1 and Phase 2. The government will conduct independent performance tests, but performers should not expect this testing to replace their own internal testing. Performers should document their internal testing and be prepared to share these results with DARPA-designated performers on the SIGMA+ team.

Proposals must describe the projected ability of the proposed sensors, algorithms, and manufacturing R&D plan to meet Phase 2 program goals and metrics, described in Section E, for each technical area. A treatment of Phase 2 program plans, including projected costs, is thus required in response to this BAA. Proposals that do not address Phase 2 goals may be deemed non-responsive.

The Government will use a phased acquisition approach for the SIGMA+ program under this Broad Agency Announcement. At this time, DARPA is soliciting detailed proposals for the Phase 1 effort only, consisting of a 30-month base period. Proposers are also required to submit a rough order of magnitude (ROM) proposal for Phase 2, which must include Phase 2 ROM costs, a Phase 2 draft statement of work, and any additional information on anticipated Phase 2 program plans.

DARPA intends to request updated technical and cost proposals for Phase 2 approximately six months prior to the completion of Phase 1. Competition for Phase 2 will be limited to only Phase 1 performers. Participation for Phase 2 will be optional and proposal guidance for Phase 2 will be provided towards the end of Phase 1 to Phase 1 performers. Associated proposal preparation costs for Phase 2 will not be reimbursed under Phase 1 awards.

Evaluation of Phase 2 proposals will be based on evaluation criteria to be specified in the Phase 2 proposal requests, and Phase 2 proposal evaluations will be conducted through a scientific and technical review process. The Phase 2 evaluation criteria will be consistent with the evaluation criteria in this solicitation, and may be tailored to the Phase 2 requests for updated proposals.

The Government reserves the right to change the award instrument or issue a new solicitation for Phases 2 if programmatic circumstances dictate.

Program continuation beyond Phase 1 will depend on funding availability and promising Phase 1 results.

An initial assessment of progress will take place at the end of the first year, when all sensors will be assessed against agreed-upon Year 1 metrics. A more rigorous assessment will take place at the end of the second year. Performance will also be assessed against agreed-upon Year 2 metrics.

HR001118S0035 SIGMA+ SENSORS 12

The first demonstration of SIGMA+ system capabilities will take place at the end of Phase 1 with the alpha version of the system. At this point, chosen sensors types and sensor algorithms will be fully integrated into the network along with baseline intelligence analytics. The beta version of the system will be completed in the following year, with refined and further scaled sensors and algorithms. In the final year of the program, SIGMA+ version 1.0 will be completed, including fully scaled sensors and operationalized algorithms. Capabilities of this system will be iteratively demonstrated through large field deployments with operational partners.

E. Technical Area Descriptions – Sensors Thrust The sensors thrust consists of two technical areas (TAs). TA1 encompasses environmental chemical sensing. TA2 involves biological sensors and is further divided into new detectors for environmental sensing of aerosolized biological threats (TA2.1), and new human-based sensors that may indicate illness before the onset of significant symptoms (TA2.2).

The following subsections provide more information on the sensors of interest along with performance metrics, cost goals, and the requested performance period. While performers may propose to one or all technical areas, separate proposals must be submitted for each technical area proposed. It is strongly recommended that proposers review Section E in its entirety even if they are only proposing to one technical area.

In each subsection, the overall objectives for each TA and its detection system are described quantitatively, with both threshold and objective metrics. The threshold metrics are expected to be reached in Year 2, and the objective metrics are expected to be reached in Year 4. A nominal timeline for field testing and demonstration is also provided. Proposers should note that the number of sensors they project to meet the proposed sensor network configuration will be required for full-scale field trials. For example, if a sensor network of 10 sensors is projected to meet SIGMA+ goals, 10 sensors must be available for field trials. If 1000 sensors are projected, 1000 sensors must be available for field trials. Proposals should provide technical rationale for the proposed size of the sensor network, ensuring the projected final sensor count will be capable of providing enough relevant information to demonstrate the desired capabilities at scale.

Lastly, a nominal system and CONOP, with derived requirements for the sensors that make up the system, are provided as a baseline and example. Performers may propose a different technical approach and CONOP to satisfy the overall system goals and objectives.

Proposed sensor technologies and scanning concepts that can most effectively use existing infrastructure, people, and vehicle fleets with minimum procurement, operational burden, and operations and maintenance costs to meet SIGMA+’s wide-area monitoring objectives are strongly encouraged. For example, existing vehicle fleets were leveraged under the SIGMA program to carry completely automated and networked RN detectors to continuously monitor the region for RN materials.

Additionally, proposed technologies that can provide additional daily operational or commercial value to help sustain the capability are encouraged. For example, chemical sensors may also

HR001118S0035 SIGMA+ SENSORS 13

provide drug production detection capability for daily operational value beyond the CBRNE mission.

Further, local processing and display of information is desirable to maintain capability for the local user during disconnected operations.

In each proposal, performers must clearly describe, as quantitatively as possible:

proposed detector technology and associated algorithms, at the individual sensor and multi-sensor fusion levels, and the nominal CONOP for satisfying the defined system requirements and testing and field trials cycle for each technical area expected system detection, identification, tracking, and interdiction capabilities within and beyond the system metrics, and any other auxiliary capabilities that can provide daily operational value network and local computational requirements as well as network data streaming requirements for proposed sensors and algorithms treatment of disconnected operational scenarios including data management, local display capability as well as a clear description of the differences and limitations between local and network algorithm capabilities any required contextual data and sources of those data, as well as requirements for the network to ingest and process these data for sensor algorithms proposed detector and algorithm performance/metrics as a function of program year in terms of the DARPA-provided metrics described in the subsections below, and any other additional metrics appropriate to the proposed sensor concept (or propose expected performance as a function of program year in terms of proposed alternate metrics that can still achieve system goals); describe highest priority and highest risk metrics risks in the proposed concept, and the proposed R&D path and intermediate milestones to retire the highest risks as soon as possible expected system design trades to be studied how sensors and sensor algorithms will be internally tested and validated, and how results will be presented expected device characterization, stability, and device-to-device variability requirements given defined performance metrics and expected algorithm performance given sufficient investment, the scaling path to approach the proposed price levels prior and ongoing efforts that will be leveraged for the proposed work.

available ROC curves and confusion matrixes from prior work on proposed sensors and sensor system expected overall size, weight, and power of sensors that make up the sensor system prior experience in developing and deploying large scale sensor systems

In addition, performers should justify and break down the prototype and final system costs, and how the proposed R&D will result in achieving the cost goals.

Raw and processed detector data (i.e., spectra and associated information) must be shared to a DARPA-maintained database beginning at month six and throughout the remainder of the program. Control documents for data formats and required metadata will be established with input from both network IT and sensor performers. The government

HR001118S0035 SIGMA+ SENSORS 14

shall have unlimited rights to all raw and processed sensor data, and unlimited rights to the developed algorithms are strongly preferred. Note that fielded detector systems and associated hardware are program deliverables and therefore will become government property.

Teaming is encouraged to leverage different capabilities that may allow a detector concept that meets or exceeds the desired metrics, and/or provide a realistic path for scale-up for manufacturing the advanced concept.

Considering the objectives of the SIGMA+ program, selected performers must comply, as applicable, with International Traffic in Arms Regulations (ITAR). Proposers must state in their proposal if the work proposed is expected to be ITAR-controlled, and, if so, if their organization has ITAR certification. It is the responsibility of the proposer to make all determinations about what is and is not ITAR-controlled regarding their work. All selected SIGMA+ performers shall comply with all applicable laws and regulations regarding export-controlled items, including, but not limited to, the requirement for contractors to register with the Department of State in accordance with the ITAR. Proposers may consult with the Department of State regarding any questions relating to compliance with the ITAR and may consult with the Department of Commerce regarding any questions relating to compliance with the Export Adminstration Regulations.

Technical Area 1: Chemical Detection DARPA seeks to develop a chemical detection system that enables persistent stand-off monitoring of large urban areas (~10 km2) with multi-story structures to detect and identify production of chemical threats (e.g., explosives, chemical warfare agents, narcotics, toxic industrial chemicals, etc.). This will require simultaneous detection and identification of multiple trace species of interest, such as precursors for specific threats.

DARPA seeks to demonstrate these capabilities in an iterative and agile approach, with the following nominal timeline and accomplishments. Proposers must address their proposed development and deliverable cycles relative to this nominal timeline:

By end of Year 1: Lab tests, initial deployment of prototypes outdoors for background collection

By end of Year 2: Outdoor testing covering > 0.3 km2 with threshold metrics and integration into SIGMA+ network

By end of Year 3: Field trials covering > 1 km2 with better-than-threshold metrics By end of Year 4: Field trials covering 10 km2 with objective metrics By end of Year 5: Large-scale field trials > 10 km2 with objective metrics or better

Proposers are required to propose an integrated system that, in addition to hardware, includes algorithms for real-time data analysis and network integration of the individual sensors to improve detection and identification capabilities, reduce false alarms, and provide spatial and temporal localization of production dynamics and potential plume tracking. These algorithms will be scaled for deployment in a cloud environment, and successful performers will be expected to work with the SIGMA+ network backbone team for integration. The system must be

HR001118S0035 SIGMA+ SENSORS 15

capable of automated, real-time and streaming network operation, but individual sensors should also have stand-alone analysis capabilities in the event of a communications outage or a communications denied environment.

Given the urban environments in which the system will be operating, the ability to detect and determine the concentrations of trace amounts of molecules of interest in the presence of dynamic and chemically complex backgrounds will be essential. Backgrounds will not only include interferences, but may also include background concentrations of precursor chemicals of interest (e.g., acetone) that swamp trace signals attributable to nefarious activity. Approaches that include methods of, for example, deconvolution, signal separation, pattern recognition, machine learning, or non-negative matrix factorization that are able to dynamically assess changes in concentration that might be indicative of a pattern of interest are encouraged.

Chemical Sensor System Goals The chemical detection system goals are shown below in Table 1, Table 2, and Table 3.

Proposers should review these metrics in detail, including the notes.

Table 1: Chemical detection system-level goals

SENSOR NETWORK SYSTEM GOALS

Parameter Threshold Objective

Scan rate, long-range* 0.5 km2 in 1 hour, 3 story structures 1 km2 in 1 hour, 3 story structures

Scan rate, point-sampler mounted in vehicle*

0.5 km2 in 1 hour: equivalent to ~10 km/hour. < 50 m resolution

1 km2 in 1 hour: equivalent to ~20 km/hour. < 30 m resolution

*Nominal urban city block: 100m x 100m, standard sidewalk and 2 vehicle lanes in each block *Proposers should state clearly what spatial resolution and localization capability their system can provide when sensors are operated in long range scanning mode

Chemical sensitivity and identification capability**

> 5 agents plus precursors (e.g., sulfur mustard, TATP, fentanyl)

Adaptable without hardware modification > 20 agents plus precursors

(e.g., sulfur mustard, TATP, fentanyl)

Identification sensitivity**

Long range, absolute levels:

< 1.25 ppm-m, < 60 s integration

Long range, change in level:

< 0.1 ppm-m

Point sampling mode:

< 5 ppb, < 60 s integration

Long range, absolute levels:

< 0.25 ppm-m, < 30 s max integration

Long range, change in level:

< 0.05 ppm-m

Point sampling mode:

< 1 ppb, < 30 s integration

**Clearly list expected sensitivity and identification capability, including targeted chemical species, expected common confusers, and provide any prior ROC curves and/or confusion matrixes

Probability of ID 85% 95% Probability of false alarm 10-5 10-6

HR001118S0035 SIGMA+ SENSORS 16

Table 2: Chemical detection system computational goals

SYSTEM COMPUTATIONAL GOALS

Parameter Threshold Objective

Lag time to identify event after signal integration, detector level (i.e.

individual sensor) < 5 seconds < 2 seconds

Lag time to identify/track event after signal integration, system level (i.e. requiring analysis of multiple sensors and integration of contextual data on network)

< 20 seconds < 10 seconds

Network computational processing requirements (outside of any on-board processing local to sensor) < 1 core per sensor

< 0.1 core, (max 10 per km2) per sensor Network update interval > 1 Hz with GPS location

Data transfer rate < 10 kb/s per sensor

Table 3: Chemical detection system cost goals

SYSTEM COST GOALS

Parameter Threshold Objective

System cost < $600k per km2 of coverage < $300k per km2 of coverage Operations cost per year < 5% of initial procurement cost < 2.5% of initial procurement cost

These system and scanning goals could be met, for example, by a system of versatile, mobile (vehicle-mounted) detectors that can operate as both long-range detectors as well as highly sensitive point detectors. These detectors might be mounted in the cargo spaces of existing vehicle fleets, and normally operated in highly-sensitive point-sensor mode to continuously scan a region for threats and to learn the background. When triggered either by an anomalous detection or by other means, a subset of the detectors might then be converted to long-range scanning mode to further identify the precursor chemicals of interest and spatially isolate the building emitting the effluent. A set of about five such long-range, mobile detectors capable of 300 m sightlines could achieve the desired clearing rate of 1 km2 in one hour of an area with buildings averaging 3 stories high, assuming each detector satisfies the other objective metrics above.

Such a system could be made of sensors described by the additional sensor-level goals in Table 4 below; however, proposers may suggest alternate sensors types and CONOPs that meet the overall system goals. Proposers must fully describe the CONOPs and provide appropriate derived goals.

Table 4: Additional chemical detection system goals for example CONOPS

SENSOR GOALS FOR EXAMPLE CONOPS

Parameter Threshold Objective

Type semi-mobile, point or long-range vehicle-transportable, long range, convertible to point sensor

Standoff distance > 200 m > 300 m Size, weight, power < 400 L, 50 kg, 1 kW < 200 L, 25 kg, 0.5 kW

Environmental tolerance -20 to +55 deg C, 20-80% RH -30 to +65 deg C, 0-100% RH

HR001118S0035 SIGMA+ SENSORS 17

DARPA is seeking systems with specified sensitivities and identification capabilities for the broadest number of species, which may be updated for expanded target databases without hardware modification, and that may be converted to point detection operation with equivalent or better detection sensitivities than when in standoff configuration. Proposers may assume detection of a single species for the purposes of detection limits in this context. However, selected performers must demonstrate the same detection limits in complex backgrounds/mixtures.

The government will administer iterative trace species detection challenges that may include specified or unknown target species, detection sensitivities, and environments/backgrounds.

Continued participation is contingent on successfully completing each challenge and meeting agreed-upon metrics for each year.

Technical Area 2: Biological Detection The biological detection TA is divided into two separate focus areas: TA2.1 focuses on environmental monitoring solutions for aerosolized bio threats, and TA2.2 focuses on human-based sensors. TA2 proposers may propose to either TA2.1, TA2.2, or both. If both areas are proposed, each must be submitted as a separate proposal.

a. Technical Area 2.1: Environmental Sensors For TA2.1, DARPA seeks to develop a system of scalable, versatile, and mobile (ex: vehicle-mounted) air monitoring detectors for continuous surveying of cities for early detection of biological attacks, as well as for monitoring and characterizing benign backgrounds for a wide variety of species. The mobile and real-time network nature of the system could additionally allow external contextual data, such as local weather data, to be taken into account to provide maximum system sensitivity while minimizing false alarms, and to provide rapid secondary screening and adjudication.

DARPA seeks to demonstrate these capabilities in an iterative and agile approach, with the following nominal timeline and accomplishments. Proposers must address their proposed development and deliverable cycles relative to this nominal timeline:

By end of Year 1: Lab tests, initial deployment of prototypes outdoors for background collection

By end of Year 2: Outdoor testing covering > 3 km2 with threshold metrics and integration into SIGMA+ network

By end of Year 3: Field trials covering > 10 km2 with better than threshold metrics By end of Year 4: Field trials covering 100 km2 with objective metrics By end of Year 5: Large-scale field trials > 100 km2 with objective metrics or better

Proposers are required to propose an integrated system that, in addition to hardware, includes algorithms for real-time data analysis and network integration of the individual sensors to improve detection and identification capabilities, reduce false alarms, and provide spatial and temporal localization of an attack. These algorithms will be scaled for deployment in a cloud environment, and successful performers will be expected to work with the SIGMA+ network backbone team for integration. The system must be capable of automated, real-time and

HR001118S0035 SIGMA+ SENSORS 18

streaming network operation, but individual sensors should also have stand-alone analysis capabilities in the event of a communications outage or a communications denied environment.

The biological detection system goals for environmental sensing are shown below in Table 5, Table 6, and Table 7.

Table 5: Biological environmental detection system-level goals

SENSOR NETWORK SYSTEM GOALS

Parameter Threshold Objective

System scan rate*

1 km2 in 4 hours 100 m & 0.5 PPL resolution

< 30 s per scan, with ability to localize signal peak to 5 s

1 km2 in 4 hours 100 m & 0.1 PPL resolution

< 30 per scan, with ability to localize signal peak to 5 s

Adaptable w/out hardware modification

Adaptable w/out hardware modification

ID > 40 pathogens w/Cat A threats ID > 40 pathogens w/Cat A threatsAgent sensitivity**

< 1500 CFU or PFU sensitivity < 300 CFU or PFU sensitivity Probability of ID 85% 95%

False positive rate < 10-6 < 10-7

*1 km2 is assumed to contain ~20 km of roadway that must be scanned at ~100 m resolution, assuming a vehicle mounted detector. A 5-second localization time provides ~100 m resolution for a vehicle at 20 m/s.

**Clearly list expected sensitivity and identification capability, including targeted species, expected common confusers, and provide any prior ROC curves and/or confusion matrixes.

**For air sampling approaches, concentrators might be used; clearly define expected flow rates and required size, weight, and power.

Table 6: Biological environmental detection system computational goals

SYSTEM COMPUTATIONAL GOALS

Parameter Threshold Objective

Lag time to identify event after signal integration, detector level (i.e.

individual sensor) < 5 seconds < 2 seconds

Lag time to identify/track event after signal integration, system level (i.e. requiring analysis of multiple sensors and integration of contextual data on network) < 20 seconds < 10 seconds

Network computational processing requirements (outside of any on-board processing)

< 1 core per sensor

< 0.1 core, (max 10 per km2) per sensor Network update interval 1 Hz with GPS location

Data transfer rate < 10 kb/s per sensor

Table 7: Biological environmental detection system cost goals

SYSTEM COST GOALS

Parameter Threshold Objective

System cost < $60k per km2 of coverage < $30k per km2 of coverage Operations cost per year < 5% of initial procurement cost < 2.5% of initial procurement cost

HR001118S0035 SIGMA+ SENSORS 19

These goals are geared towards a system made of versatile, mobile and continuous air-monitoring detectors for a wide urban area. Such a system could be made of vehicle mounted sensors described by the additional sensor-level goals in Table 8 below; however, proposers may suggest alternate CONOPs that meet the overall system metrics and desire for an adaptable and mobile system.

Table 8: Additional biological environmental detection system goals for example CONOPS

SENSOR GOALS

Size, weight, power* < 50L, 30 kg, 200 W Vehicle mounted

< 25L, 15 kg, 100 W Vehicle mounted

Environmental tolerance -20 to +55 deg C, 0-100% RH -30 to +65 deg C, 0-100% RH System cost < $60k/sensor < $30k/sensor

* Not including concentrators

In all cases, the proposers must fully describe the CONOPs and suggest appropriate derived goals. Proposers should also state if their sensors could be adapted for or have capabilities beyond air monitoring, e.g., for liquids.

The government will administer iterative trace species detection challenges that may include specified or unknown target species, detection sensitivities, and environments/backgrounds.

Continued participation in the program is contingent on successfully completing each challenge.

b. Technical Area 2.2: Human-Based Sensors For TA2.2, DARPA seeks to develop a two-tier biological detection system that identifies respiratory pandemics up to three weeks earlier than the state of the art, and provide days earlier detection of large-scale biological attacks. DARPA will field and test this system at scale in order to demonstrate these early warning capabilities. The goal is a system that will measure physiological attributes and biomarkers (defined by the proposer) of an individual, identify deviations from normal to determine if an individual is getting sick, and provide an accurate prediction of the severity of illness before or at the onset of symptoms (tier 1). Suspicion of illness, coupled with predicted severity, will then trigger routing of the individual to obtain accurate point-of-care testing for pathogen identification and reporting (tier 2). Portable point-of-care technology is encouraged, as ultimately diagnostic testing ideally occurs in the home / at the bedside, instead of expecting the patient to travel to a fixed location.

The government will conduct a series of field trials for system capability demonstration. These trials will take an iterative approach as in TA1 and TA2.1; proposers should describe a proposed development and test schedule. As a validation of biological sensing capabilities, the proposed human-based sensor systems should demonstrate detection of a natural influenza outbreak three weeks earlier than the current state of the art in Year 4 of the program. This is envisioned to require deployment of more than 1,000 sensors in Phase 2.

Proposers are required to propose an integrated system that, in addition to hardware, includes algorithms for real-time data analysis and network integration of the individual sensors to improve detection and identification capabilities, and to reduce false alarms. These algorithms will be scaled for deployment in a cloud environment, and successful performers will be

HR001118S0035 SIGMA+ SENSORS 20

expected to work with the SIGMA+ network backbone team for integration. The system must be capable of automated, real-time and streaming network operation, but individual sensors should also have stand-alone analysis capabilities in the event of a communications outage or a communications denied environment. The proposal must describe and justify the physiological attributes and biomarkers being sensed, as well as the frequency of sensing required to detect aberrations from a healthy state.

Such a system consists of multiple subsystems for tier 1 and tier 2 as follows. First, real-time monitoring of relevant biochemical analytes, such as oxygen, are measured to correspond with tissue-level effects, and deviations from homeostasis are accurately detected. This information can be integrated with measurements of vital signs and activity level, to augment the sensing accuracy of biomarkers. The readout subsystem might consist of a small wearable device that reads out the biomarker assay results, plus a smartphone that performs some level of local data processing and transmits to the network; alternatively, both of these functions could be contained in a single device similar to a smart watch. A suite of data analysis algorithms processes the data (locally, on the network, or a combination of both) to assess whether the wearer is getting sick.

Early signs of illness, sensed in real-time as deviations from normal patterns by these advanced algorithms, lead to notifications to the person and encouragement to proceed to a point of care location for further assessment. All these data are collected by the readout subsystem and transmitted to the network backbone for early detection of a potential outbreak or attack.

At the point of care location screening for relevant pathogens and appropriate treatment are conducted. Pathogen information from the point-of-care subsystem is then also transmitted in real time to the network to identify the potential outbreak. It is expected that the point-of-care subsystem is largely based on a COTS platform that may require additional development for additional pathogens and integration into the network.

Proposers must describe how all data will be secured and anonymized, so that there is no risk that potential health information is unprotected.

The goals for this system are shown in Table 9 below, and the derived subsystem goals shown in Table 10, Table 11, and Table 12 below.

Proposers may suggest alternate CONOPs that meet the overall system objectives described above, but must explain how their alternate CONOPs would quantitatively achieve equivalent or better early detection capability within the context of the system metrics and goals provided below. In all cases, proposers must fully provide appropriate derived goals and metrics at the same or greater level of detail as provided below.

HR001118S0035 SIGMA+ SENSORS 21

Table 9: Biological human-based detection system goals

INTEGRATED SENSOR SYSTEM GOALS

Parameter Threshold Objective

Sensor types Biomarker sensor + readout system + point-of-care platform

Pathogen sensitivity, selectivity 95% for Influenza A and B, RSV, MERS-CoV, one additional relevant respiratory pathogen*

(minimum set)

95% for Influenza A and B, RSV, MERS-CoV, plus 4 additional relevant respiratory pathogens*

Probability of detection, days prior to onset of symptoms 80%, 0 days prior 95%, 3 days prior

False alarm rate** <10-2

Severity prediction accuracy, days after onset of symptoms*** 80%, 3 days after 95%, 1 day after

Cost, biomarker sensor plus readout < $1k/system (1000 units) < $500/system (1000 units) Cost, point-of-care platform < $10k/system, $50 per test < $4k/system, $10 per test

*pathogens transmitted by the respiratory route that is contagious and responsible for acute pulmonary pathology **this rate is likely dominated by the false alarm rate of the biomarker sensor and readout system to detect and predict severity of illness

***Levels of severity are defined as follows:

Level Oxygen level Clinical evidence of pneumonia Care required Mild > 90% No None

Moderate 85-90% No Emergency room visit Severe < 85% Yes Hospitalization

Table 10: Biomarker sensor subsystem goals

BIOMARKER SENSOR SUBSYSTEM GOALS

Parameter Threshold Objective

Size and weight minimally invasive Power self-contained, no power or maintenance required

Additional biomarker sensor subsystem goals:

• The sensor technology shall be of sufficient maturity to be ready for clinical evaluation

(in the context of a clinical research protocol) within 3 months of contract award for initial data collection in humans.

• The sensor shall be scalable to thousands within 4 years.

• The sensor shall have already been demonstrated to function in vivo for at least three months’ duration.

HR001118S0035 SIGMA+ SENSORS 22

Table 11: External readout sensor subsystem goals

EXTERNAL SENSOR SUBSYSTEM GOALS

Parameter Threshold Objective

Operating system Android…