An Industry Day in support of Steel Vulture has been planned and will be held on 21 OCT 2021 from 14:30-16:00 ET using Zoomgov.com.
Questions received by the published deadline of 15 OCT 2021 will be addressed during the event. A limited time for Q&A will be allotted during the event, and there is no guarantee you will have an opportunity to pose your question(s) during the Industry Day event. To insure you have the opportunity to have your question officially submitted, it must be submitted to email@example.com (with subject line of “TW-22-0002 Inquiry”) no later than 2400 ET on 21 OCT 2021.
To register for the Industry Day, please use the following link:
Tradewind SCO Steel Vulture (TW-22-0002) Industry Day
SUMMARY. The TRADEWIND OTA Enterprise Office, in conjunction with the Joint Artificial Intelligence Center (JAIC), will host a competitive process to develop artificial intelligence (AI) object detection, classification, and tracking algorithms applicable to Wide-Area Motion Imagery (WAMI) collected in the Mid Wave Infrared (MWIR) band. The AI algorithms developed will be evaluated by an existing USAF Test and Evaluation (T&E) process. The AI algorithms developed are intended to be deployed with the MWIR WAMI sensor.
INTENT. The JAIC seeks to enter into Other Transaction Authority (OTA) agreement(s) with academic and industry partners whose solutions are favorably evaluated. As such, this opportunity is considered competitive, and solutions will be evaluated independently of one another for merit.
BACKGROUND AND CONTEXT. DOD’s Strategic Capabilities Office (SCO) is developing a next-generation Mid Wave Infrared (MWIR), Wide-Area Motion Imagery (WAMI) sensor. The size of the area covered by this sensor is 10x the current capability of a single airborne sensor. Due to the large amounts of data collected (over 32TB in a single mission), the diversity of target sets, varying phenomenologies, and data complexities; human analysis is not possible for the near-real time requirements of the mission. AI is required to analyze the data. As such, this development effort will advance the state-of-the-art for artificial intelligence and machine learning (AI/ML) algorithms, facilitating near-real-time target detection, classification, and tracking across the field of view (FOV) of the sensor.
PROBLEM DESCRIPTION. Due to the large amount of data generated by the MWIR WAMI sensor, human analysis of data is impractical and not timely enough to meet mission needs.
TECHNICAL CAPABILITIES PURSUED. TRADEWIND is seeking to fund a competition of multiple vendors, or teams of vendors, in the development of AI/ML algorithms to detect, classify, and track targets of interest (TOI) from the imagery collected by the MWIR WAMI sensor in near-real time. The algorithms will be trained against data sets from the actual sensor along with surrogate imagery from an existing full motion video (FMV) sensor flown on the same airborne platform. The training data will be provided by the government. TRADEWIND envisions an Agile development effort and after each sprint the algorithms developed will undergo test and evaluation at AFRL’s COMPASE Center. The T&E methodology, metrics, and measures will be described to the participating performers at Industry Day (21 October 2021); written descriptions of the T&E process will be provided at that time.
The complete announcement is available in the accompanying document.
Deadline: 24 NOV 2021, 1500 ET
Participation requires that you register in the Vulcan platform using the associated link below.