News and Media

Media Center

Latest news and media releases

Recent News


Recent News

DS2 and Belle AI Develop ATAK Plugin for SOFWERX Virtual Sprint

Dynamic Software Solutions (DS2), as a team member under Belle Artificial Intelligence Corporation, participated in a four-week Automated Analytics Virtual Sprint. SOFWERX, in concert with USSOCOM Joint Acquisition Task Force (JATF), hosted this event to identify technologies that are relevant to assist with automated analytics during Phase 0 Irregular Warfare (IW) operations.

Operators at the tactical edge require a wide range of relevant information to make sound tactical decisions. Although the operators frequently collect the raw data that this information is derived from, the process of integrating, analyzing, and interpreting the data to produce usable information is frequently the job of analysts who sit far from the tactical edge. As the number of data sources increase, the amount of data to be analyzed grows exponentially, and the distillation of useful information is laborious and time consuming, resulting in intelligence products that are often stale and cannot be easily actioned after-the-fact. This information exchange is also dependent on reliable communications between the analysts and the operator. Moreover, in the contested and denied environments that will characterize Special Operations Forces’ (SOF) future operating environment, operators are not be able to readily communicate outside of the tactical bubble and will be forced to make more decisions rapidly and organically to effectively contend with peer and near-peer competitors. To meet these challenges, USSOCOM sought to take advantage of the capabilities of artificial intelligence running on edge computing platforms to provide SOF teams with the organic capacity to distill mission-relevant information from operator-born sensor data.

During the four-week Automated Analytics Virtual sprint, Belle AI and DS2 built an ATAK plug-in capable of identifying buildings in overhead imagery and numbering them in accordance with standard Gridded Reference Graphic (GRG) practice. In addition to automatically annotating the buildings using a convolutional neural network (CNN) model, the plug-in provided integration with named areas of interest (AOI), phase lines, and cellular/wi-fi infrastructure information from other vendors. The resulting GRG was able to be shared among TAK devices running all versions of TAK.

After successfully building the ATAK plugin, our team focused on integrating the other companies’ data that were participating in the event to stream into ATAK. The combined common operating picture, along with other geolocated data developed by other companies, were presented during the demo as the final product. Several of these firms sourced data from various sources (e.g., Twitter, Instagram, news sources) to draw Key Leader Engagement (KLE) profiles. KLE profiles are needed to meet the demands of complex contemporary conflicts. Civilian and military actors need to work together and reaching the population and acquiring their support is often vital to mission success. KLE is an important element of C2 that the commander can use to achieve this. KLE is not a new phenomenon. Military commanders and diplomats have been meeting with important local officials for decades in different countries and mission areas. However, the nature of contemporary conflicts and the need for collaborative C2 have renewed the interest in this concept.

Companies that participated in this event included Belle Artificial Intelligence Corporation, Amazon, Cognovilabs, Reveal Global Consulting, Premise, Parasanti, and Alteryx. They were selected from nearly 80 firms submitting proposals.

Showing 0 Comment

Comments are closed.