In this series:
Within emerging wireless concepts, one often finds cognitive radio and massive multiple-input multiple-output (MIMO) technologies. Massive MIMO technology has been covered in previous blog posts, where we discussed the subject from its theory to real-world prototyping. There is more to come latter, but for this blog post, we will discuss concerns related to the physical design issues unique to cognitive radio as well as MAC protocols vulnerability.
It is important to stress that the ultimate objective of wireless system and network design is to increase spectrum efficiency while achieving fairness in terms of serving unfortunate cell-edge users. This objective can be meet using one or both of the following strategies:
- Efficiently utilizing the spectrum (even within the licensed band). Actual measurements taken in an urban setting reveal a typical utilization rate of 0.5% in the 3–4 GHz frequency band. It drops to just 0.3% in the 4–5 GHz band.
- Maximize throughput as a function of channel state (including exploiting spatial signatures). This strategy is accomplished by using high-order modulation, MIMO arrangements, and efficient re-transmission schemes based on channel state information. Orthogonal frequency-division multiplexing (OFDM) is currently adopted as the modulation technique and multiple access schemes (OFDMA) readily lend themselves to a MIMO scheme.
Cognitive radio devices consider the first approach while massive MIMO attempts to improve per user throughput by reducing multi-user (stream) interference through beamforming at transmission and multi-user detection techniques at the receiver. Massive MIMO can also adopt cognitive radio opportunistic approaches wherein key features such as spectrum sensing and access coordination are used.
Following from the fact that spectrum is usually under-utilized, cognitive radio is a promising technology for implementing opportunistic spectrum access. Cognitive radio devices are capable of sensing and coordinating access to idle portions of the spectrum while not interfering with the activity of primary users (PUs).
Physical design issues unique to cognitive radio
The implementation feasibility when using state-of-the-art wideband radios is determined according to two central facts:
- Spectrum sensing and PU activity detection pertaining to the physical layer receiver function. Avoiding harmful interference to the PU mostly belongs to transmitter’s ability to control its transmission power and spectral mask (lower adjacent channel leakage).
- Access coordination belongs to the MAC layer function, which in turn considers spectrum sensing output to advise if the channel is idle or not.
Unfortunately the physical layer design comes with unique issues; the sensing and detection algorithm must be reliable in at least two scenarios:
- The low signal-to-noise ratio (SNR) scenario (including deep fading case) where, for instance a -116 dBm digital TV signal shall be detected with an SNR of -21 dB and a probability of detection of 0.9 (a receiver noise figure (NF) of 11 dB is assumed, a typical noise figure of a low-cost wideband radios).
- Channel sensing in the presence of a strong adjacent PU or cognitive radio device. Given the state of current wideband receivers and ADC dynamic range, it is unlikely that this scenario will be supported. Unfortunately, a strong adjacent channel is not an isolated test scenario and must be taken into account.
All 3GPP and IEEE standards suggest these scenarios as part of radio performance compliance. One potential solution is the use of spatial filtering.
MAC protocols vulnerability
Channel access coordination pertains to MAC function. Cognitive radio MAC protocols can be classified into three categories: (i) split-phase, (ii) dedicated control channel, and (iii) frequency hopping. Regardless of the protocol choice, the vulnerability from selfish or malicious cognitive radio users who seek to gain an unfair share of the available spectrum is a serious problem. A few countermeasures have already been suggested. These include threshold voting and locating a misbehaving cognitive radio device, for instance. A misbehaving device’s radio frequency signal signature can be exploited, which guarantees a reliable countermeasure, assuming that a misbehaving device is not close to well-behaved cognitive radio devices or a real PU.