Ai for Cognitive Radios

Only available on StudyMode
  • Download(s) : 6
  • Published : April 25, 2013
Open Document
Text Preview

AI FOR COGNINITIVE RADIO

Abstract: Cognitive radio (CR) is an enabling technology for numerous new capabilities such as spectrum access, spectrum sensing, spectrum decision, spectrum sharing and self organizing networks. This paper reviews several AI techniques used in cognitive radios such as : artificial neural network (ANN), hidden Markov models (HMMs), metahueristic algorithms, these techniques are proposed to provide the cognition capability in a cognitive engine.

The modern software defined radio is the heart of a cognitive radio. The applications executing on the radio distinguish a cognitive radio from a software-defined radio. Additional hardware in the form of sensors and actuators enables more cognitive radio applications. Various artificial intelligence approaches to machine learning and decision making may be applied to the cognitive radio system. In this paper, a survey of related cognitive radio topics is reviewed.

Keywords: software defined radio, cognitive radio, cognition cycle, artificial intelligence techniques for CR, characteristics of CR, types of CR .

1. INTRODUCTION

Software Defined Radio (SDR) and Artificial Intelligence (AI) technology enable the new field of Cognitive Radio (CR). An overview of these two areas is presented in this section :

A. Software Defined Radio :
A software-defined radio has several major sub-systems, but
is essentially a distributed computing system. The major sub-systems are :

• RF Front End :
A SDR requires a general purpose RF front end. This typically requires a wide tuning range, ideally it tunes from DC to light notwithstanding the technical problems. A pre-selector / power amplifier allows the selection of a subset of the spectrum, and either tunable filters or fixed filters are used for this purpose. A Rx or Tx chain processes the radio signal in the analog domain. These chains may be combined into a single, bi-directional path at design time. Finally, a synthesizer is required to generate the local oscillators for processing (Figure 2). The analog processing is software configurable through the setting of switches to select filters and through the setting of registers to control mixer frequencies. [pic]

Figure 1 – General SDR RF Frontend. The RF Frontend
has the following sections: Pre-selector / Power Amplifier, Receive Chain, Transmit Chain, and Synthesizer .

• MODEM :
The MODEM subsystem accomplishes the modulation and
demodulation of the data stream using digital methods.
.
1. General Purpose Processing Capabilities :
FPGA and DSP processing resources are not optimized for protocol applications or for AI applications. High performance SDRs typically have General Purpose Processing resources that are programmed in high-level languages.

2. Cryptographic Capabilities :
Military radios usually require Cryptographic security functions. For High Assurance systems, this involves dedicated hardware. There are commercially available, software programmable cryptographic processors for these requirements.

• Software Architecture :

One of the most critical architectural features in a SDR is the software structure or software architecture. Usually a standard Real Time Operating System with a standard interface such as POSIX is considered advantageous. Figure 2 shows a solid software architecture for a SDR. Cognitive Radio Applications execute in the top layers where waveforms execute.

[pic]

Figure 2 – SDR Software Architecture. A layered SW
Architecture supports SW portability and enables CR
application to be added easily.

B. Cognitive Radio :

A Cognitive Radio is an extension of modern Software Defined Radio. This extension creates new capabilities for users. An “aware radio” has sensors and is aware of the environment (or at least a subset of the environment). An “adaptive radio” is aware of its...
tracking img