Army 17. 1 Small Business Innovation Research (sbir) Proposal Submission Instructions

TECHNOLOGY AREA(S): Biomedical

OBJECTIVE: The objective of this topic is to research, develop, and demonstrate a computer vision system capable of accurate and robust geometric pose estimation for the human body that would support the use of emerging robotics and autonomous systems for applications in combat casualty care in future operational environments. This computer-vision and perception system will need to accurately determine the position and orientation of major body parts (limbs, torso, head, etc.) in real-time and be invariant to environmental conditions typically seen in the far-forward operating environments.

DESCRIPTION: The application of unmanned/robotic technologies has the potential to perform timely extraction of casualties in man-denied environments and reduce the risk of injury to medics and other personnel during casualty extraction attempts in high threat areas. Current robotic systems are prohibitively slow and rely heavily on tele-operation. In order for autonomous robotic systems to be utilized effectively for causality extraction missions, the perception system must identify and locate a casualty at a distance. Once the casualty is located, the robotic system requires a reliable and robust means for autonomously determining the location and orientation/posture of the body of the casualty at close range. This capability is critical for enabling future unmanned systems to perform the autonomous dexterous manipulation tasks required during fully autonomous casualty extraction or semi-autonomous casualty extraction with the human operator providing high-level commands and supervisory control. This system should include the sensor/computing hardware and software required to perform the perception tasks with the ability to easily integrate with future mobile robotics platforms. Therefore, an open architecture and compliance with existing interoperability standards is required. This capability should be extensible to a wide range of robotic applications that require close interaction with humans.

As outlined in the “Unmanned Systems Integrated Roadmap FY2013-2038”, the Department of Defense expects unmanned systems to play a key role in many of the Tier One Joint Capability Areas (JCAs), and specifically mentions the capabilities of unmanned systems to provide medical resupply and casualty extraction and evacuation as future systems are enabled through greater autonomy with respect to both navigation and manipulation. In order for autonomous systems to be used for casualty extraction, and other medically-relevant operations, a number of key technology areas need to be further developed and matured in order for these robotic systems to effectively and safely interact with casualties and their care providers in the field. One of these technology areas that needs to be further matured and tailored for the combat environment is a robotic perception system that is capable of precisely determining the position/pose of a soldier, and more specifically a casualty or patient, in real-time. The observed position and orientation of the casualties’ body, acquired from this perception system, is needed for any robotic system to interact safety with casualties because it is required to effectively plan the kinematics of the robotic system.

There are several computer vision systems in the commercial arena, focused on human pose estimation for various different applications. These commercial available systems are not robust enough for the combat environment and are often highly sensitive to environmental factors, e.g. lighting and background imagery. They also usually require a large standoff distance to work effectively. Existing systems typically focus solely on the mapping of joint locations, but mapping the exterior surface of the body is also important if this perception system is to be used effectively for applications in which the robotic system needs to physically engage with a human. There are also considerations unique to the combat casualty care environment that would needed to be addressed in a human/casualty pose estimation system, e.g., casualties in the prone position sometimes in close proximity to the robotic system, camouflaged clothing, partial occlusion of the body by gear, medical equipment, or the environment, etc. It is the intent of this topic to research and develop a robust human pose estimation system, consisting of both the sensors and computational components, that addresses the specific needs of the combat casualty care environment in the field that can be leveraged in future robotics systems to enable the capability of autonomous casualty extraction, and evacuation.

This human pose perception system needs to be developed with a strong emphasis on ease of integration with future robotic and autonomous systems, which would use this perception system as a modular component, using a system of systems approach. As the perception technology matures for combat casualty care, it could be extended to future robotic systems with dexterous manipulation capability that could provide some level of remote monitoring, imaging, or intervention to support enroute care, or prolonged field care, in future operational environments in which traditional medical resources are unavailable or denied access.

PHASE I: Conduct a feasibility study of the proposed approach for applying the state-of-the-art in human pose estimation systems and techniques for use in a far-forward combat casualty care environment. The feasibility study shall take into account the challenges likely to be imposed by the system in a far-forward operational environment and should be extensible to combat casualty care applications, when integrated into robotic systems with dexterous manipulators, that would enable capabilities like remote patient monitoring, and remote imaging (e.g. ultrasound) and examination. The feasibility study of the proposed approach shall be used to inform the conceptual design of a human/casualty pose estimation system. The conceptual design of the casualty identification and pose estimation system should take into account the ease of integration into existing and emerging robotic platforms, both in terms hardware and software interfaces.

PHASE II: Demonstrate the functionality of the system by accurately identifying and locating a casualty at a stand-off distance, and mapping of a casualties’ pose and 3D boundary volume at close-range and in real-time in an environment representative of far forward field care. Because this perception system is targeted for applications of autonomous casualty extraction, it needs to not only accurately identify the pose of major body components, but also estimate weight and center of gravity of individual components and the body as a whole. The perception system should demonstrate robustness to different types of poses (e.g. crouching, sitting, laying, and standing) and variation in Soldier height and body type. Modularity of the casualty pose perception system needs to be demonstrated by providing an open application program interface that would allow for a robotic manipulator system to easily ingest the body pose mapping information for use in kinematic/trajectory planning for applications involving close proximity human-robot interaction in real-time. The perception system shall be prototyped, in both hardware and software, as a modular system that shall be demonstrated by integration with an existing robotic platform with a manipulator. The integrated robotic prototype system shall demonstrate the robot’s ability to use the 3D pose and body mapping provided by the perception system in order to execute a relevant task.

PHASE III DUAL USE APPLICATIONS: Further develop these capabilities to TRL-7 or 8 and demonstrate the application of this casualty pose estimation system through a robotic system designed for either casualty extraction, or assisted or autonomous en route care capabilities. While the primary intended use for this system is for stand-off casualty extraction on the battlefield, an alternate use could be as an integral component to a medical device with a target application to provide stand-off casualty care capabilities. Once validated conceptually and technically, the dual use applications of this technology are significant in both civilian emergency services, and other military operations. The technology could be used in applications, in which standoff interaction with soldiers or civilians is of paramount importance, e.g. environments with CBRNE exposure with potential applications to mortuary affairs operations. This technology could be extended in the future to allow medical experts remote care capabilities to far-forward casualties that would otherwise be inaccessible at the Role 1 level of care.

REFERENCES:

1. “Unmanned Systems Integrated Roadmap FY2013-2038”, Department of Defense

2. “U.S. Army Roadmap for UAS 2010-2035”, U.S Army

3. “Early Entry Medical Capabilities Concept of Operations” Capabilities Development Integration Directorate, U.S. Army Medical Department Center and School, U.S. Army Health Readiness Center of Excellence, August 2015

4. “Safe Ride Standards for Casualty Evacuation Using Unmanned Aerial Vehicles” NATO STO Technical Report TR-HFM-184, December 2012

5. “UAS Control Segment Architecture”, https://ucsarchitecture.org/

6. “SAE AS-4 JAUS Standard”, http://www.sae.org/standardsdev/jaus.htm

KEYWORDS: Combat Casualty Care, Human Pose Estimation, Robotics, Autonomous Systems, Robotic Perception, Casualty Extraction, En Route Care, Casualty Evacuation, Unmanned Systems, Human-Computer Interaction, UAS, UGV, CASEVAC

TECHNOLOGY AREA(S): Biomedical

OBJECTIVE: The objective of this topic is to investigate, develop and demonstrate a biometric identity and access management capability on Android-based platforms that enable rapid authentication of medical personnel into medical devices and software, role-based access to medical personally identifiable information (PII), Health Insurance Portability and Accountability Act protection, and the rapid identification of casualties provisioned before mission execution. This research will incrementally advance the state of the art biometric mechanisms for authentication of mobile medical devices and software at the point of injury (POI), such that the final demonstration shows proof-of-concept feasibility of biometrics identity and access management in medical systems used in the operational environment.

DESCRIPTION: This topic is designed to address the imbalance between security requirements for protection of information and operational requirements for mobile electronic health record systems. The security requirements for protection require that access mechanisms to information prevent unauthorized access. This is in contrast to the operational requirements to meet user expectations for easy and rapid mechanisms for employment of a capability that does not detract focus from their mission of providing care to traumatically injured personnel. Through the use of biometric identity and access management, android-based systems can facilitate rapid user access authentications during combat that do not require the use of additional authentication mechanisms such as Common Access Card (CAC), pins, passwords, or other physical authentication tokens. This would allow for information protections by enabling role-based access controls for user specific "need to know" access to medical personally identifiable information (PII) subject to Health Information Privacy Protection Act (HIPPA) protection as well as tactical, operational, intelligence, and situational awareness information.

The capability for the end user (Medic) to log into the Android and into Android applications, such as the electronic DD Form 1380 Tactical Casualty Card (eTCCC), is a mandatory requirement. The capability should be implemented to facilitate the integration into any suitable current and future application. The capability to use biometrics to identify a casualty at the point of injury is optional, but preferred. Proposals that address both capabilities with solution strategies that are feasible under combat conditions will be more favorably considered.

The biometric identification and access mechanisms would be utilized in the austere conditions on a battlefield that is dirty, noisy, extremely stressful, and require the users to wear protective gear (i.e. body armor, gloves, goggles/eye protection). The protective gear may limit ease of access to some biometric identification features. Additionally, for casualty identification, some biometric features may not be available based upon the injuries and condition of the patient.

The biometric identification and access management mechanism would ultimately be employed on the U.S. Army’s Nett Warrior platform, and integrated into the eTCCC application (minimum requirement). The biometric identification and access management system should have mechanisms for the provisioning and management of both biometric user authentication and biometric identification of soldiers. The biometric identification and access management mechanism would be utilized in systems that are connected to the network through the Tactical Radio Network and would need to minimize network traffic requirements.

PHASE I: Develop system design and Proof of Concept, to include system hardware and software architecture for an Android-based biometric identification and access management capability that allows for user authentication into both an Android device and an Android application (mandatory), and also allows for biometric identification of casualties (optional) that were previously provisioned. The Proof of Concept should include the basic properties of software architecture, underpinning hardware technology concepts and description of the system concept that addresses feasibility and benefit in an austere (battlefield) environment with intermittent low-bandwidth communications. The Proof of Concept should identify key capability parameters and provide predictions that will later be validated in Phase I and II activities. Conduct feasibility testing of system components and provide results of analytical and experimental results that validate the assertions about the key capability parameters identified in the Proof of Concept. Seek partnerships within government and private industry for transition and commercialization of the production version of the product.

PHASE II: From the Phase I design, develop a ruggedized prototype of a biometric identification and access management capability that could be used on any Android mobile device and demonstrate the capability with soldier medical attendants and simulated casualties in a relevant environment; such as at a US Army TRADOC Battle Lab. The capability will be used with Nett Warrior End User Devices, which currently are Samsung S5 phones with cybersecurity hardening. Demonstrate the biometric identification and access management capability for the medic to log into the Nett Warrior device, eTCCC app and other applications. The biometric identification and access management capability for casualty identification should be integrated into a government provided eTCCC app. Access to the eTCCC app source code will be provided.

PHASE III DUAL USE APPLICATIONS: Incorporate system improvements resulting from Phase II evaluation results and user feedback. Execute system evaluation in a suitable operational environment (e.g. Advanced Technology Demonstration (ATD), Joint Capability Technology Demonstration (JCTD), Marine Corps Limited Objective Experiment (LOE), Army Network Integration Exercise (NIE), etc.). Present the prototype project, as a candidate for fielding, to applicable Army, Navy/Marine Corps, Air Force, Cost Guard, Department of Defense, Program Managers for Combat Casualty Care systems along with the government and civilian program managers for emergency, remote, and wilderness medicine within state and civilian health care organizations, and Departments of Justice, Homeland Security, Interior, and Veteran’s Affairs. Execute further commercialization and manufacturing through collaborative relationships with partners identified in Phase II.

REFERENCES:

1. Edwards, John and E. Keiser, “DoD rethinks Identity and Access Management strategy”, C4ISR & Networks, 17 May 2015. http://www.c4isrnet.com/story/military-tech/cyber/2015/05/17/dod-identity-access-management/27167901/

2. Project Manager, Department of Defense Biometrics (DoD Biometrics) Website. http://www.eis.army.mil/programs/biometrics

3. NSTC Subcommittee on Biometrics and Identity Management, “NSTC Policy for Enabling the Development, Adoption and Use of Biometric Standards”, 7 September 2008. http://www.biometrics.gov/standards/NSTC_Policy_Bio_Standards.pdf

4. White House National Science and Technology Council, “Registry of USG Recommended Biometric Standards Version 5.0”, September 2014. http://www.biometrics.gov/standards/Registry_v5_2014_08_01.pdf

5. Department of Defense Integrated Data Dictionary Version 5.0, 8 December 2011, DIN: BIMA-STB-STD-11-001.2

6. Manemeit, Carl and G.R. Gilbert, “Joint Medical Distance Support and Evacuation JCTD, Joint Combat Casualty Care System Concept of Employment” NATO HFM-182-33 paper, 21 April 2010. http://ftp.rta.nato.int/public/PubFullText/RTO/MP/RTO-MP-HFM-182/MP-HFM-182-33.doc

KEYWORDS: android, authentication, automation, biometric, combat casualty care, combat medic, identification, mobile device

TECHNOLOGY AREA(S): Biomedical

OBJECTIVE: The objective of this topic is to research and study the effects of maneuvering/ acceleration G-Forces and altitude have on casualties during flight and ground evacuations, and develop a constraint system that will restrict and limit the vehicles movement/maneuver capabilities during casualty extraction and evacuation missions to within casualty tolerable limits. The movement/maneuver restrictions and limitations need to be established on autonomous Unmanned Systems, namely Unmanned Aerial Vehicles (UAVs) and Unmanned Ground Vehicles (UGVs) when performing Casualty Evacuation (CASEVAC) missions. Also, the preliminary research and development look at hypoxia and hypothermia effects at altitude on casualties during evacuation via a UAV.

DESCRIPTION: Currently, casualties are air evacuated via manned helicopters from the battlefield with altitude, acceleration and g-forces in manned piloted flights limited to the pilot’s human performance, and ground evacuated casualties are transported by manned vehicles. Drones and UAVs/UGVs are being deployed with operational forces and being unmanned, UAVs without a human onboard, they are only bound by their structural limitations. An unpressurized UAV can still fly at a high altitude and could make serious G-Force maneuvers that normally may not occur in an unpressurized aircraft with human pilot onboard. So with the operational forces looking to deploy with UAVs/UGVs for logistical medical resupply missions, it is only a matter of time before a commander decides to use an UAV/UGV asset for transporting a casualty from the battlefield back to a rear facility. If a human is put on a UAV, the UAV may exceed human-tolerable g-forces and altitude, and further injure a casualty during an extraction mission.

“The U.S. Navy’s Chief of Naval Operation Instruction (OPNAVINST) 3710.7U (Naval Air Training and Operating Procedures Standards (NATOPS) General Flight and Operating Instructions) restricts manned piloted flight for unpressurized aircraft; paragraph 8.2.4.1 - Unpressurized Aircraft governs the use of oxygen equipment in unpressurized aircraft with oxygen systems. In unpressurized aircraft with oxygen systems, the pilot at the controls and aircrew participating in physical activity (loadmasters) shall use supplemental oxygen continuously when cabin altitude exceeds 10,000 feet. When oxygen is not available to other occupants, flight between 10,000 and 13,000 feet shall not exceed 3 hours duration, and flight above 13,000 feet is prohibited. In aircraft where oxygen systems are not available (such as helicopters), it must be determined that it is mission essential for flight altitude to exceed 10,000 feet. Time above 10,000 feet shall not exceed 1 hour and altitude shall not exceed 12,000 feet.”

Casualty evacuation/extraction limitation have not been fully developed for UAVs in this instruction, so instructions may need to be changed or waived to reflect the realities of UAV/UGV operations for casualty extraction. UAV/UGV guidelines are still being developed; this research can aid the personnel writing doctrine to institute guidance on UAV/UGV operations that will be used for casualty evacuation.

This research should investigate the limits a casualty can handle during an air/ground evacuation for G-Force maneuvers, acceleration and altitude. The research needs to focus and look at a variety of different wounds: open and closed head trauma; internal abdominal bleeding and chest wounds (lung and heart damage); and closed long bone fractures.

PHASE I: Conduct preliminary G-Force and altitude investigation on human performance on healthy individuals and casualties during air evacuation in unpressurized and pressurized flights that are manned piloted. Investigate human performance on G-forces (vibration and repeated mechanical shock) dealing with ground and air transport. This initial research will be the baseline going into Phase II evaluating the effects of G-Forces and altitude has on casualties during medical evacuation in a UAV/UGV. At the end of the Phase I, provide a feasibility study and a plan to develop a concept design on software and/or design a breadboard/proof of concept constraint solution that can be further developed into a test bed product that can be field tested from the Phase II research. Also, prepare Human Use documentation to be submitted for Use of Humans in Research (IRB) trials in Phase II, if needed.

RESEARCH INVOLVING ANIMAL OR HUMAN SUBJECTS: The SBIR/STTR Programs discourage offerors from proposing to conduct Human or Animal Subject Research during Phase 1 due to the significant lead time required to prepare the documentation and obtain approval, which will delay the Phase 1 award.

All research involving human subjects (to include use of human biological specimens and human data) and animals, shall comply with the applicable federal and state laws and agency policy/guidelines for human subject and animal protection.

Research involving the use of human subjects may not begin until the U.S. Army Medical Research and Materiel Command's Office of Research Protections, Human Research Protections Office (HRPO) approves the protocol. Written approval to begin research or subcontract for the use of human subjects under the applicable protocol proposed for an award will be issued from the U.S. Army Medical Research and Materiel Command, HRPO, under separate letter to the Contractor.

Non-compliance with any provision may result in withholding of funds and or the termination of the award.

PHASE II: From the results of the Phase I feasibility study and proof of concept constraint system demonstration, continue development to a production prototype that further investigates the effects of acceleration, g-forces, and altitude has on casualties during flight and ground evacuations. Design and develop a mission constraint solution to restrict and constrain a current or future UAV/UGV during a casualty evacuation mission is required. The focus should be on how the g-forces and motion affect a variety of different wounds: open and closed head trauma; abdominal and chest wounds; and long bone fractures. The prototype constraint solution that is to be developed should mitigate these forces on the human physiology to ensure the casualty does not incur more harm while being transported via a UAV/UGV. At the end of Phase II, a field able prototype constraint system needs to be developed to be used in field evaluation testing in a simulated operational mission. Also, a commercialization plan needs to be developed.