AI/ML for Spectrum-agile Communications

Smartnet Communications is a pioneer in the integration of AI and ML solutions in the design and operation of wireless communication systems and networks in shared, congested, and/or contested spectrum environments. Our team has extensive expertise in designing and training AI/ML algorithms for channel access protocols, considering heterogeneous coexisting wireless technologies (e.g., LTE, 5G NR, Wi-Fi, radar, DSRC/C-V2X, etc.). Our solutions target both mid-band spectrum (e.g., CBRS, C-band, 3.1-3.45 GHz and 3.45-3.55 radar bands, unlicensed UNII bands in the 5 GH and 6 GHz regions) as well as high bands (millimeter-wave and sub-THz spectrum). Sample technological innovations made by our team include MAMBA (a low-overhead solution for accurate beam tracking at mmWave bands using online reinforcement learning techniques), SmartLink (a solution that exploits the multi-cluster scattering phenomenon to establish multi-link directional communications), and FastLink (a rapid approach for initial access and cell discovery in mmWave systems).

Signal Intelligence for Secure Wireless Systems

The open nature of the wireless medium leaves it exposed to adversarial attacks, including active attacks (e.g., jamming and signal manipulation) and passive attacks (eavesdropping). Using commodity radio hardware, unauthorized adversaries can easily intercept packet transmissions. Although encryption can be applied to ensure information secrecy, eavesdroppers can still perform low-level RF and traffic analysis, and capture with different degrees of certainty several communication attributes, such as the packet size, its duration, inter-packet times, traffic directionality, channel state information, etc. Through knowledge of the protocol semantics, these attributes can be correlated to create transmission fingerprints, from which an adversary can derive a wealth of contextual information, and subsequently use such information to launch sophisticated active attacks (e.g., selective jamming, data poisoning, etc.). The Smartnet Commmunications team provides significant contributions in this domain, targeting both defense and commercial needs. Examples of our technologies include secure protocols for FH (frequency hopping) based satellite links, attack-resilient signal classifiers for modulation detection and protocol identification, novel CNN/RNN classifiers for identifying rogue signals and other anomalies, and signal obfuscation techniques for hiding meta-attributes and low-layer headers of wireless transmissions.

Next-Generation Wi-Fi

Smartnet Communications expertise spans all generations of Wi-Fi, from the first 802.11/802.11b based Wi-Fi devices all the way to the upcoming Wi-Fi Generation 7 (based on the IEEE 802.11be standard). Our solutions address channel efficiency as well security aspects of Wi-Fi, including Physical-layer and MAC-layer security. An example of our secure Wi-Fi technological advances is the patented CryptoSafe software, which ensures fingerprint-free secure Wi-Fi communications through modulation obfuscation and embedding of encryption keys within the preamble of transmitted Wi-Fi frames. Implemented at the digital modulation stage, Friendly CryptoJam enables transmitted devices to hide unencrypted PHY/MAC headers and obfuscate the frame payload’s modulation scheme. Our design maintains backward-comptability with legacy systems. Our engineers and computer scientists design ML-based techniques for detecting malicious attacks on Wi-Fi systems, including SIG tampering and selective jamming attacks (e.g., attacks on frequency offset, frame detection, pilots, etc.). Smartnet Communications also provides solutions for harmonious coexistence between Wi-Fi systems and other systems that operate over unlicensed bands, including LTE LAA, 5G NR-U, ZigBee, Bluetooth, and DSRC/C-V2X. Our extensible preamble modulation (eP-Mod) design for 802.11n/ac/ax based systems allows devices to embed user-specific information in the preambles of Wi-Fi frames, which can be used to facilitate paiwair encryption of a full frame, exchange secret keys, convey frame-specific “traveling pilots”, adapt the PHY-layer modulation scheme, facilitate dynamic switching of the MIMO/FD mode, and others.