Waveform reconstruction from ontological description
Analog Integrated Circuits and Signal Processing
| Hi-index | 0.00 |
Software Defined Radio (SDR) is a radio in which some or all of the physical layer functions are software defined, i.e. radio's operating functions, such as modulation scheme or frequency, are implemented with the use of software processing, as opposed to being implemented in hardware. Existing SDR architectures differ in the scope and the partition of the offered functionality between the reconfigurable software and the hardware components. Thus, each SDR offers a different set of adjustable and observable parameters (also called knobs and meters). Cognitive Radio (CR) is a term to describe a radio communication paradigm which takes advantage of the SDR architecture and allows for dynamic changes of radio's operational behavior in order to achieve a variety of goals, including interoperability, performance optimization, opportunistic use of resources and others. Although CR could be implemented in many ways, employing different artificial intelligence and optimization algorithms, in this work we focus solely on the knowledge-based CR architecture which utilizes an inference engine to provide the cognitive capability. This approach separates the knowledge about the radio domain from the cognitive algorithms, allowing domain experts to express their knowledge in an implementation-independent fashion by specifying ontologies and rules. Existing CR engines interface SDR platforms relying on APIs that are radio-specific and sometimes platform-dependent. So far a standard SDR API has not been established. As a consequence, the CRs are bound to a specific type of SDR, they are not reusable and are unable to adapt to changes in the SDR functionality without recoding. The lack of a standard CR architecture and a standard API for interfacing reasoners with the SDR's parameters (knobs and meters) is the main problem of implementing CR according to the knowledge-based approach. We propose a Cognitive Radio Framework (CRF) architecture, which aims to meet the following requirements: 1) support reusability of expert knowledge, 2) be SDR-independent, 3) be platform-independent, and 4) be able to adapt to changes in the domain without recoding. Instead of using a domain-specific API to interface SDRs, the CRF architecture allows the reasoner to interface an SDR using a thin and generic API. This architecture is capable of maintaining the interface with heterogenous SDRs that change their APIs without the need of recoding the reasoner API due to the fact that the interface information is provided dynamically. Moreover, since the CR can be viewed as a subclass of self-controlling software, the domain-adaptable CRF architecture can be utilized in other domains where the self-controlling software paradigm could apply, for instance, in radar tracking.