Deep Learning based Intelligent Scheduling for Tactical Adaptive Networks with Cognitive Engine (DISTANCE)

TDMA-based scheduling is widely used to coordinate transmissions for the MAC, which is typically centralized and not adaptive and therefore is not flexible in terms of allocating dynamic time slots among transmitters to adapt to spectrum and traffic dynamics. Tactical networks need to operate in contested and congested environment subject to various forms of impairments in communications due to in-network interference, out-of-network interference, channels and traffic congestion. Therefore, it is essential for distributed nodes to operate cohesively as a network and build together cognitive network knowledge. To address this critical need, IAI proposes to design and implement the DISTANCE to provide a unified framework of novel modules, including data management, deep learning based spectrum analysis, and distributed scheduling for tactical networks in contested environments. DISTANCE supports tactical nodes to sense the spectrum, characterize the communication impairments, disseminate this information over the network, and utilize distributed scheduling protocols to mitigate or alleviate these impairments and improve tactical communications.