Underlay Dynamic Spectrum Access (DSA) addresses the challenge of radio spectrum scarcity by allowing Cognitive Radios (CRs) in a secondary network (the "secondary users" - SUs) to share a spectrum band with an incumbent primary network (PN). In underlay DSA, CRs transmit over the same portion of the spectrum being used by the PN and at the same time, by limiting their transmit power level so that the interference they create on the PN remains below a tolerable threshold. Despite having been extensively studied, the main challenge in underlay DSA remains on how to establish the interference threshold for the PN links and how the SUs autonomously become aware of the interference they create on the PN, especially when following an ideal operating setup where there is no exchange of information between primary and secondary networks. In contrast to previous works that assume that SUs could access the control feedback channel in the PN, we will present a technique that does not rely on any information exchange between the networks but, instead, takes advantage of the use at the primary network of adaptive modulation and coding (AMC). More specifically, in the presented technique we propose an underlay DSA mechanism that only requires the estimation of the adapted modulation scheme being used at the nearest PN link (this can be achieved by performing modulation classification on the radio waveform from the PN link that the CR is passively listening to during spectrum sensing). Since the full AMC mode (modulation order and channel coding rate) chosen for transmission depends on the link signal-to-interference-plus-noise ratio (SINR), by estimating this mode, it is possible for a CR to learn the SINR experienced by a primary link. However, in a general setting with no exchange of information between the primary and the secondary network, while a CR can apply signal processing techniques to infer during spectrum sensing the modulation order being used by a primary transmitter, there is no previous known practical way to infer the experienced throughput of the primary link. The technique to be presented applies a non-linear autoregressive exogenous neural network (NARX-NN) cognitive engine in the SUs to infer, without tapping into feedback or control channels from another network, the full AMC mode used in the transmission of the other network, and to leverage this inference in the realization of a fully autonomous and distributed underlay DSA scheme. For this, the NARX-NN cognitive engine outputs an estimate of the throughput at the nearest primary link from using as inputs the estimated modulation order in the same link and the transmit power for a sequence of probe messages. Moreover, leveraging the use of adaptive modulation in the PN also allows the CR to establish the PN interference threshold. This is because the use in a radio link of adaptive modulation allows for the background noise to increase up to a certain level before the average throughput in the network starts to decrease. In the context of underlay DSA where the PN does not exchange any information with the SN, the interference imposed on the PN by an underlay-transmitting CR can be seen as a background noise that can be increased up to a level which does not affect the average throughput in the PN. In the technique to be presented, a CR couples these observations with the estimation of the full AMC mode in the nearest PN link provided by the cognitive engine to decide on its transmit power and other waveform settings. Simulation results will show that the presented technique is able to limit the reduction in PN relative average throughput within a prescribed target maximum value, while at the same time finding transmit settings for the SUs that will result in as large throughput in the SN as could be allowed by the PN interference limit. As such, while succeeding in its main goal of autonomously and distributedly determining the transmit power of the SUs such that the interference they create remains below the PN limit, the presented technique is also able to manage the tradeoff between the effect of the SN on the PN and the achievable throughput at the SN. Specifically, simulation results will show that for the proposed system with a target primary network maximum relative average throughput reduction of 2%, the achieved relative change is less than 2.5%, while at the same time achieving useful average throughput values in the secondary network between 132kbps and 50kbps. Moreover, it will be seen that the enabling factor in the operation of the proposed technique is the ability of the NARX neural network cognitive engine to accurately predict the modulation scheme and channel coding rate used in a primary link without the need to exchange information between the primary network and the secondary (e.g. access to feedback channels). This ability also results in a significant increase in the transmission opportunities in the SN compared to schemes with the same basic approach but able to only use modulation classification information from the primary link, instead of the full AMC mode.

Cognitive Anti-jamming Satellite-to-Ground Communications on NASA's SCaN Testbed
Dale Mortensen, P. E. (NASA Glenn Research Center, USA); Sudharman Jayaweera (Bluecom Systems and Consulting, LLC, USA)

Machine learning aided cognitive anti-jamming communications is designed, developed and demonstrated on a live satellite-to-ground link. A wideband autonomous cognitive radio (WACR) is designed and implemented as a hardware-inthe- loop (HITL) prototype. The cognitive engine (CE) of the WACR is implemented on a PC while the software-defined radio (SDR) platform utilized two different radios for spectrum sensing and actual communications. The cognitive engine performs spectrum knowledge acquisition over the complete spectrum range available for the SATCOM system operation and learns an antijamming communications protocol to avoid both intentional jammers and inadvertent interferers using reinforcement learning. When the current satellite-to-ground link is jammed, the cognitive engine of the ground receiver directs the satellite transmitter to switch to a new channel that is predicted to be jammer-free for the longest possible duration. The end-to-end, closed-loop system was tested on the NASA's Space Communications and Networking (SCaN) testbed on the International Space Station (ISS). The experimental results demonstrated the feasibility of satellite-to-ground cognitive anti-jamming communications along with excellent anti-jamming capability of machine-learning aided cognitive protocols against several different types of jammers. Index Terms—Cognitive anti-jamming communications, cognitive radios, machine learning, Q-learning, reinforcement learning, satellite communications, wideband autonomous cognitive radios.

Boosting Artificial Intelligence in Software Defined Systems with Open Infrastructure
Li Zhou (National University of Defense Technology, P.R. China); Qi Tang (NUDT, P.R. China); Jun Xiong, Haitao Zhao, Shengchun Huang and Jibo Wei (National University of Defense Technology, P.R. China)

Software infrastructure is playing an increasing role in both terminals and clouds in recent years, such as software communication architecture (SCA) for terminals and OpenStack for clouds. The common benefit is that they both enable a system to decouple applications from the platform with a standardized application programming interface (API). As a component or a higher-level application in SCA systems, artificial intelligence (AI) is significant to meet various challenges, which potentially include wireless network planning, resource optimization, complicated decision making and in-depth knowledge discovery. Meanwhile, open clouds are evolving into intelligent clouds, which allow users to use AI functions to perform cognitive tasks such as analysis, gaining insights from data or transfer learning, which is also known as AI-as-a-Service (AIaaS). In this presentation, we would like to talk about how to take advantages of AlaaS in the open infrastructure, to boost AI and innovations in software defined systems.

Towards an Ontology for SCA APIs
Durga Suresh and Mieczyslaw Kokar (Northeastern University, USA)

The Software Communication Architecture (SCA) is an implementation-independent architectural framework that specifies a standardized infrastructure for a software defined radio. The intent behind the standardization of the SCA was to support waveform and component portability, interchangeability of implementations, interoperability of software and hardware components, software reuse, and architecture scalability for communications platforms. To the best our knowledge, there are no open-source implementations of the SCA. SCA has many APIs. All SCA APIs and the SCA itself are specified in UML. While UML tools provide some methods for syntactically constraining the development of a specification of a system, they don't support the capability of verifying or enforcing the semantic constraints. In this paper, we are presenting our work on creating a SCA Ontology (SCA4.1) that can be used as a base ontology in the radio domain to help with prototyping the SCA APIs. This ontology is based on Nuvio (Northeastern and VIStology) ontology, a new foundational ontology developed by our team, inspired by the original Cognitive Radio Ontology (CRO), Quantity, Units, Dimensions and Types Ontology (QUDT), DOLCE ultralight ontology (DUL) and Situation Theory ontology (STO-L). The top-level structure, properties and the relationships between the classes and objects will be presented to show the components and interfaces of the SCA. We will then discuss the evaluation of this ontology based on 1) coverage of knowledge, 2) inference capability, 3) precision in defining classes and 4) extendibility.

The middleware is among the three infrastructural system software. Embedded software defined systems (SDS), e.g., the SDR system, also deeply depend on the middleware. These systems are generally component-based, constructed using the model driven architecture and enabled by PIM, PSM and the transformation technique. Inter-component interaction and reconfiguration/reconstruction in either component level or application level on heterogeneous platforms are basic operations in the SDS, which are enabled by current middleware techniques, including CORBA, ICE, COM+, RMI, RPC, etc. Whereas, most available middlewares focus on large-scale distributed network computing systems, of which the size and weight are not well adapted to embedded systems. Especially, the SCA-based SDR system shifts its middleware from CORBA to transfer mechanism demonstrates the above issues exist in the off-the-shelf middleware. For this reason, we initiated a program to craft a function-complete but small and efficient middleware, named RPCExpress, for the embedded SDS. Initial efforts are demonstrated to be quite fruitful. The first edition of RPCExpress enables component-based software development, supports a set of semantics of IDL and provides an IDL to C++ transformation tool. Component-based evaluation is carried out on a practical embedded platform, and the integration with SCA is on the way. The experimental result demonstrates that the RPCExpress outperforms the off-the-shelf middleware in efficiency and footprint.

As a classical transfer mechanism, CORBA has been widely used in SCA-based software radio systems. However, the communication efficiency of the middleware based on CORBA, such as TAO and omniORB, is not very satisfactory, and this has become one of the challenges that constrain SCA technology to be further widely applied. In SCA 4.1 specifications, the transfer mechanism becomes more flexible. We try to improve the efficiency of the transfer mechanism from two aspects: inter-chip and intra-chip, we call them remote-call and local-call separately. Considering the remote-call delay is not low, improving the local-call efficiency becomes the premier task. This achievement also provides more possibilities and options for component granularity within GPPs. On this basis, we are working on optimizing the efficiency of inter-chip communications.

In order to facilitate interoperability and portability of software defined radio components, the conformance to SDR standards (including APIs and behavior specifications) is really important. Due to the huge number of requirements and to ensure reproducibility of the conformance assessment, a testing methodology is necessary to show SDR requirements coverage and test result status as well as an automatic test bench to ease tests execution. This presentation will show how to use a structured approach to develop and maintain a test bench for the assessment to SDR standards conformance. The presentation will be made by Alain Ribault, CTO of KEREVAL. KEREVAL is a French testing Lab. Draft plan: 1) Needs for the assessment of SDR conformance 2) Testing methodology a. From the SDR requirements to the tests b. Test design process c. Compliance checkpoints definition d. Modeling e. Testing generation 3) The test bench 4) Test of SDR components: an example 5) Conclusion / Q&A

Using SCAv4.1 Tools to Develop Applications and Platforms
Juan Pablo Zamora Zapata and Steve Bernier (NordiaSoft, Canada)

Software Defined Systems is the industry's response to the ever-increasing complexity and flexibility requirements of today's electronic systems. The Software Communications Architecture (SCA) enables the fulfillment of such requirements for complex Heterogeneous Embedded Distributed Systems (HEDS). In this tutorial, NordiaSoft will demonstrate how SCA software can be developed using tools that combine model-driven design (MDD) and rapid application development (RAD). Using Zero Merge code generation technology, the business source code for an SCA component (modulator, encoder, etc.) is kept separate from the infrastructure code that deals with deploying, instantiating, configuring, connecting, inter-process communications, etc. This tutorial will show how the proposed approach provides the greatest potential of reuse for intellectual property. This SCAv4.1 tutorial will also illustrate concepts like application creation, packaging, installation, deployment and execution using an SCA FM Waveform Application.

Mission Adaptable Ground Station for Small Satellites
Howen Fernando and Marcus Matsumura (SSC - Pacific, USA)

The technological growth for Small Satellites (i.e. CubeSats, etc.), and the increasing collaboration in this domain, prompted the development of a Software Defined Radio ground station to address scalability, interoperability, and adaptability using the Software Communications Architecture 4.1 Standard. SCA enables the use of one ground station platform for many different small satellites, allowing the SCA-based ground station to switch from one satellite system to another as they become visible.

Invited Presentations