Cognitive Radio Communications for Vehicular Technology – Wavelet Applications

Wireless communications are nowadays a dominant part of our lives: from domotics, through industrial applications and up to infomobility services. The key to the co-existence of wireless systems operating in closely located or even overlapping areas, is sharing of the spectral resource. The optimization of this resource is the main driving force behind the emerging changes in the policies for radio resources allocation. The current approach in spectrum usage specifies fixed frequency bands and transmission power limits for each radio transmitting system. This approach leads to a very low medium utilization factor for some frequency bands, caused by inefficient service allocation over vast geographical areas (radiomobile, radio and TV broadcasting, WiMAX) and also by the usage of large guard bands, obsolete now due to technological progress. A more flexible use of the spectral resource implies that the radio transceivers have the ability to monitor their radio environment and to adapt at specific transmission conditions. If this concept is supplemented with learning and decision capabilities, we refer to the Cognitive Radio (CR) paradigm. Some of the characteristics of a CR include localization, monitoring of the spectrum usage, frequency changing, transmission power control and, finally, the capacity of dynamically altering all these parameters (Haykin, 2005). This new cognitive approach is expected to have an important impact on the future regulations and spectrum policies. The dynamic access at the spectral resource is of extreme interest both for the scientific community as, considering the continuous request for wideband services, for the development of wireless technologies. From this point of view, a fundamental role is played by the Institute of Electrical and Electronic Engineers (IEEE) which in 2007 formed the Standards Coordinating Committee (SCC) 41 on Dynamic Spectrum Access Networks (DySPAN) having as main objective a standard for dynamic access wireless networks. Still within the IEEE frame, the 802.22 initiative defines a new WRAN (Wireless Regional Area Network) interface for wideband access based on cognitive radio techniques in the TV guard bands (the so-called “white spaces”). Coupled with the advantages and flexibility of CR systems and technologies, there is an ever-growing interest around the world in exploiting CR-enabled communications in vehicular and transportation environments. The integration of CR devices and cognitive radio networks into vehicles and associated infrastructures can lead to intelligent interactions with the transportation system, among vehicles, and even among radios within vehicles. Thus, improvements can be achieved in radio resource management and energy efficiency, road traffic management, network management, vehicular diagnostics, road traffic awareness for applications such as route planning, mobile commerce, and much more. Still open within the framework of dynamic and distributed access to the radio resource are the methods for monitoring the radio environment (the so-called “spectrum sensing”) and the transceiver technology to be used on the radio channels. A CR system works on a opportunistic basis searching for unused frequency bands called “white spaces” within the radio frequency spectrum with the intent to operate invisibly and without disturbing the primary users (PU) holding a license for one or more frequency bands. Spectrum sensing, that is, the fast and reliable detection of the PU’s even in the presence of in-band noise, is still a very complex problem with a decisive impact on the functionalities and capabilities of the CRs. The spectrum sensing techniques can be classified in two types: local and cooperative (distributed). The local techniques are performed by single devices exploiting the spectrum occupancy information in their spatial neighbourhood and can be divided into three categories (Budiarjo et al., 2008): "matched filter" (detection of pilot signals, preambles, etc.), "energy detection” (signal strength analysis) and “feature detection" (classification of signals according to their characteristics). Also, a combination of local techniques in a multi-stage design can be used to improve the sensing accuracy (Maleki et al., 2010). Nevertheless, the above-mentioned techniques are mostly inefficient for signals with reduced power or affected by phenomena typical for vehicular technology applications, such as shadowing and multi-path fading. To overcome such problems, cooperatives techniques can be used. Cooperative sensing is based on the aggregation of the spectrum data detected by multiple nodes using cognitive convergence algorithms in order to avoid the channel impairment problems that can lead to false detections. (Sanna et al., 2009). Within the energy detection method, a particular attention needs to be paid to the properties of the packets wavelet transformation for subband analysis, which, according to the literature, seems to be a feasible alternative to the classical FFT-based energy detection. Vehicular applications are in most cases characterized by the need of coping with fast changes in the radio environment, which lead, in this specific case of cognitive communication, to constrains in terms of short execution time of the spectrum sensing operations. From this point of view, the computational complexity of the wavelet packets method is of the same order of the state-of-the-art FFT algorithms, but the number of mathematical operations is lower using IIR polyphase filters (Murroni et al., 2010). In our work we are investigating the use of the wavelet packets for energy detection spectrum sensing operations based on the consideration that they have a finite duration and are self- and mutually-orthogonal at integer multiples of dyadic intervals. Hence, they are suitable for subband division and analysis: a generic signal can be then decomposed on the wavelet packet basis and represented as a collection of coefficients belonging to orthogonal subbands. Therefore, the total power of the signal can be evaluated as sum of the contributions of each subband, which can be separately computed in the wavelet domain. Furthermore, the wavelet packets can be used also for the feature detection spectrum sensing, using statistical parameters such as moments and medians. We concentrate in our research on both applications of the wavelet packets to the spectrum sensing operations, investigating their efficiency in terms of reliability and execution time, applied specifically to the needs of vehicular technology and transportation environments. The other key issue for the development of the previously mentioned standard is the choice of an adaptive/multicarrier modulation as basic candidate for data transmission, having as the most known representative the Orthogonal Frequency Division Multiplexing (OFDM) modulation. OFDM-like schemes are mature enough to be chosen as a core technology for dynamic access wireless networks. At the same time, the potentialities in terms of optimization for this specific purpose are not yet thoroughly investigated. Particularly, the Wavelet Packet Division Multiplexing (WPDM) modulation method, already known for about ten years to the scientific community, is a suitable candidate to satisfy the requirements on physical level for a dynamic access network (Wong et al., 1997): WPDM has already proven to be able to overcome some of the OFDM limits (limited spectral efficiency, problems with temporal synchronization especially in channels affected by fading) and is at the same time based on use of the same wavelet packets employed for subband analysis used for spectrum sensing operations . Our research investigates the use of the WPDM for cognitive radio purposes, combined with the wavelet approach for spectrum sensing, for offering a complete, wavelet-based solution for cognitive application focused on the problematic of vehicular communication (channel impairments, high relative velocity of the communication peers etc.).