Cognitive radio characterizes an ambient aware and software-defined radio system that is capable to autonomously and intelligently adjust the system parameters. It has the potential to optimize the usage of all relevant resources for service- driven wireless communications. We identify four types of resources; these are the radio, the computing, and the radio and user application resources. This paper presents their modeling, including several models for each resource type. The modeling serves as the basis for a joint resource management in a cognitive radio system that may eventually trade-off one resource (type) against the other(s). Two approaches for such a management are presented: a distributed-cooperative and a centralized- integrated. We finally discuss the applicability and utility of the proposed models, and the potentials of a joint resource management in cognitive radio in terms of user satisfaction and revenues.

Software-defined radio (SDR) is an emerging concept that leverages the design of software-defined and hardware-independent signal processing chains for radio communication. It introduces flexibility to wireless systems, facilitating the dynamic switch from one radio access technology to another or, in other words, the de- and reallocation of computing resources from one SDR application to another. This paper introduces an SDR computing resource management framework. It accounts for several SDR system characteristics, including real-time computing requirements, limited computing resources, and the use of heterogeneous multiprocessor plat-forms, including multiprocessor systems-on-chip. The framework features the tw-mapping, a dy-namic mapping algorithm that is apt for many cost functions and, thus, adaptable to any radio scenario. The cost function proposal dynamically manages the available computing resources to satisfy the given SDR computing constraints. Two relevant SDR scenarios and the corresponding simulations, based on representative SDR platforms and processing chains, demonstrate the framework's importance and suitability for software-defined radios.

Cognitive radio is generally understood as an intelligent wireless communication system aiming at the efficient utilization of radio resources. We argue for the extension of its scope to also address the management of computing resources, including processing and bandwidth capacities of SDR platforms, and present a cooperative resource management framework: The joint radio resource management (JRRM) and the computing resources management (CRM) entities interchange information to cooperatively decide if and what kind of terminal reconfiguration would be the most appropriate in each situation. Therefore, the cognitive radio system continuously observes the radio and the computing environments. We discuss a realistic case study and present a simple CRM algorithm. Simulation results show that such a cooperative resource management approach can achieve important improvements over the JRRM by itself. In particular, our proposal considerably reduces the number of lost user sessions due to inappropriate reconfiguration decisions and, thus, serves more wireless users.

The division in several layers of the implementation of systems is a solution adopted to avoid complexity, provide flexibility and improve portability and code reusability through different hardware. Middleware (intermediate layer between two other layers) implementations are based on the use of increasingly high-level languages and application programming interfaces (API). The field programmable gate arrays (FPGA) world can also apply this approach to produce building blocks independent from hardware platforms and devices. This paper presents details of the implementation of a middleware, called platform and hardware abstraction layer (P-HAL) when applied to FPGA devices. It was specially designed for radio applications and allows designing specific functions independently of the hardware context where they are applied, thus providing flexibility to the so-called software radios employing FPGA devices.

Software Defined Radio (SDR) is an emerging technology that is based on the software implementation of the signal processing blocks found in a radio transceiver. The switch between radio access technologies may then be as easy as changing the software running on a future SDR terminal. SDR terminals refer to mobile equipment and base stations. These terminals will comprise general purpose processors, digital signal processors and/or reconfigurable logic devices. As a result, typical heterogeneous computing problems may appear in the SDR context. This article focuses on the mapping issue, discusses its relevance in software defined radio, and introduces an adequate mapping algorithm. The algorithm efficiently tackles the problem of mapping SDR function chains, i.e. signal processing blocks of a SDR transceiver, to heterogeneous processing platforms. We expose our approach, discuss its performance, present extensive simulation results and derive conclusions.

Prioritisation based on traffic class is widely used within radio resource control strategies to provide quality of service (QoS) assurances in 3G wireless systems. Under high load conditions, traffic prioritisation can lead to situations where, momentarily, most of the radio resources are assigned to delay-constrained connections (e.g. conversational users) and, consequently, the available bandwidth for less priority users (e.g. interactive traffic) is significantly reduced. In This work we analyse the impact of this bandwidth oscillation phenomenon on TCP connections over a downlink UMTS channel. Experimental results have been obtained using legacy TCP connections over a real-time UMTS testbed.

Software radio applications require a framework to develop and deploy applications, especially those related to radio infrastructure. It is interesting that a given application may be executed on any software radio. But, since hardware platforms used in this context will have multiple architectures and devices, a software layer to make applications independent from hardware is mandatory. Ad-hoc software for a given hardware platform may produce the best software performance. Conversely, when software is not targeted to any concrete platform the lost of performance may be excessive and the overhead introduced by any platform-dependent library could become intolerable. In this paper the resource utilization of a software radio application using a simple hardware abstraction layer is studied and compared to an ad-hoc implementation to make an assessment of the introduced overhead. The particularity of the hardware abstraction layer is that it runs on a platform which only processors are FPGA devices.

Software radio will have an important role in future terminal and base station definition. Although the ideal software radio architecture is highly restricted by technology, at present it is possible to build complex systems completely customised and reconfigurable using FPGA devices. They provide the designer with the possibility of modifying the hardware functional qualities without giving up high performance or using resources poorly. To check how important is this affirmation considering future wideband systems, a representative DS-CDMA subsystem has been built over an FPGA platform that allows fast reconfiguration of internal system blocks.