Читать книгу Dynamic Spectrum Access Decisions - George F. Elmasry - Страница 46

There is a significant difference between energy and power when analyzing a signal. A signal can be categorized in different ways, as shown in Table 2A.1, including an energy/power signal category. Notice that in Table 2A.1 the signal can be either category A or category B.

Table 2A.1 Different categories of signals

Signal category	Category A	Category B
1	Continuous time	Discrete time
2	Deterministic	Random
3	Periodic	Aperiodic
4	Even	Odd19
5	Energy	Power

A signal is time varying and can convey information or not convey information. Noise is a type of signal that does not convey information. Some jammers can also be a type of signal that does not convey information. A signal can be a function of time and a function of other independent variables.

A signal is categorized as an energy signal if it has finite energy, that is, 0 < E < ∞. When the signal is analyzed over a given time duration and the signal energy can be measured, it is an energy signal. An example of that signal is a pulse signal where the pulse is either positive or negative. A decaying pulse signal is also an energy signal because its energy can be measured over a given time duration. On the other hand, a sinusoidal signal energy cannot be measured in time domain, and hence it cannot be categorized¹⁹ as an energy signal. The sinusoidal signal is a power signal. Moving from time domain to frequency domain, the sinusoidal signal power spectral density can be measured. The sinusoidal signal is a power signal over infinite time. A signal can be categorized as a power signal when it has finite power without time limitation.

Notice that with spectrum sensing, we may sense a modulated signal over a sinusoidal wave (carrier). We look at frequency bands of carrier frequencies and hence we measure the signal power. The term “energy detection” is used loosely with spectrum sensing and it means integrating the measured signal power over a limited time period (time of sensing or dwell time). With spectrum sensing, the correct term for energy detection should be power integration over a finite time. The term “energy detection” is widely used because spectrum sensors integrate the sensed power spectral density over the sensing time period and the process of integration over time leads to using the term energy detection. In spectrum sensing references, spectrum sensing is a multidimensional process that considers time, frequency, and power. Power here can be power spectral density measured over limited time. The simplest way of spectrum sensing is known as energy detection, which integrates the power spectral density measured by the spectrum sensor, in frequency domain, over a given sensing period.