Northeastern University

Salvatore D'Oro

Tommaso Melodia

Francesco Restuccia

Leonardo Bonati

Lorenzo Bertizzolo

Zhangyu Guan

Michele Polese

Hai Cheng

Provided herein are systems for controlling a network of distributed non-terrestrial nodes including a control framework operative to train and control a plurality of the non-terrestrial nodes, the control framework including a control interface in communication with a network operator to receive one or more specified control objectives, and a learning engine operative to train a virtual non-terrestrial network, wherein the control framework is further operative to transfer knowledge gained through the training of the virtual non-terrestrial network to the network of distributed non-terrestrial nodes as data-driven logic unit configurations tailored for the specified control objectives.

Distributed deep reinforcement learning framework for software-defined unmanned aerial vehicle network control

Methods and systems are provided for controlling an unmanned aerial vehicle (UAV) swarm network including receiving, at a control terminal, an operator input defining a plurality of network performance directives, constructing a network control model representing the plurality of network performance directives, decomposing the network control model into a plurality of sub-models, generating, via the control terminal, a distributed UAV control model corresponding to each of the distributed sub-models, transmitting the generated distributed UAV control models to each of a plurality of UAVs of the UAV swarm network, executing, at each of the plurality of UAVs, at least one of the distributed UAV control models, and changing, responsive to the execution of the distributed UAV control models, an operating condition of at least one of the plurality of UAVs to satisfy the network performance directives.

Software defined drone network control system

A self-optimizing operating system for next-generation cellular networks is provided. The system provides telecommunications operators with an efficient and flexible network control platform that hides low-level network details through a virtual network abstraction and allows them to define centralized and high-level control objectives with no need for in-depth and detailed knowledge of cross-layer optimization theory or the underlying wireless protocol stack implementation.

Operating system for software-defined cellular networks

A radio access network (RAN) intelligent controller (RIC) and corresponding method may be implemented within RAN and in next-generation cellular networks to improve performance. The RIC comprises an interface to a RAN and further comprises a data-driven logic unit. The data-driven logic unit (i) produces, based on data received from the RAN via the interface, a representation describing a state of the RAN and (ii) based on the representation describing the state, instructs an action associated with at least one network element. The interface transmits a message based on the action instructed. The message is to be routed to the at least one network element. The representation is based on a context of the RAN. The message transmitted enabling re-configuration of the at least one network element. The re-configuration improves performance of the at least one network element within the context.

Intelligence and learning in O-RAN for 5G and 6G cellular networks

Methods and systems are provided for allocating resources to users in a wireless network including a plurality of edge nodes that provide wireless network access and multi-access edge computing functions. Slice requests are received from the users for a type of resource, including one or more of networking resources, storage resources, and computation resources. A set of slice requests to be admitted is determined based on resource availability constraints among one or more of the networking resources, the storage resources, and the computation resources at each edge node.

Methods for multi-access edge computing network slicing in 5G networks

A steganographic communication system and method are provided. A covert packet generator can embed a stream of covert data as covert data symbols within primary I/Q symbols of a primary data stream in a covert packet. The covert packet has a data structure having a header, a payload, and a payload error detecting code. The header includes information on how to demodulate the covert packet by a receiver. The covert packet generator can also determine if a number of primary I/Q symbols is large enough to generate the header and can generate displacements in the primary I/Q symbols in a constellation diagram randomly in a plurality of transmissions to mimic channel noise. A transmitter and receiver can provide mutual authentication for covert transmissions.

Private 5G cellular connectivity as a service through full-stack wireless steganography

Methods and systems for allocating radio access network (RAN) spectrum resources among a plurality of mobile virtual network operators (MVNOs) of a network of base stations. The methods and systems include determining a slicing enforcement policy that assigns resource blocks (RBs) of frequency units and time slots of spectrum resources to each MVNO according to a slicing policy in which each MVNO is allocated an amount of the spectrum resources on at least one base station in a determined time span. The slicing enforcement policy minimizes overlap between each MVNO's set of RBs with another MVNO's set of RBs on a same base station, and interference between each MVNO's set of RBs with another MVNO's set of RBs on an interfering base station.

Methods for the enforcement of network slicing policies in virtualized cellular networks

A method of determining a response of a radio frequency wireless communication system to an adversarial attack is provided. Adversarial signals from an adversarial node are transmitted to confuse a target neural network of the communication system. An accuracy of classification of the incoming signals by the target neural network is determined.

Neural network for adversarial deep learning in wireless systems

A machine learning (ML) agent operates at a transmitter to optimize signals transmitted across a communications channel. A physical signal modifier modifies a physical layer signal prior to transmission as a function of a set of signal modification parameters to produce a modified physical layer signal. The ML agent parses a feedback signal from a receiver across the communications channel, and determines a present tuning status as a function of the signal modification parameters and the feedback signal. The ML agent generates subsequent signal modification parameters based on the present tuning status and a set of stored tuning statuses, thereby updating the physical signal modifier to generate a subsequent modified physical layer signal to be transmitted across the communications channel.

Device and method for reliable classification of wireless signals

A communications system provides for accurate fingerprinting of devices across a communications channel. A transmitter device modifies a physical layer signal prior to transmission as a function of signal modification parameters. A receiver device classifies a received physical layer signal and to outputs a classification indicator and a score. The receiver further analyzes the classification indicator and the score to produce an updated set of signal modification parameters provides the parameters to the transmitter device. The transmitter device, in turn, updates its signal modification parameters accordingly, thereby generating subsequent communications that more clearly indicate a fingerprint of the transmitter device.

Growing community of inventors

Allston, MA, United States of America

Salvatore D'Oro