Chapter 4
Quasi-Real-Time Testbedding
When it comes to evaluating the performance of mobile radio communication systems, the most common method by far is computer simulations. Fortunately, much of the backbreaking work has been eliminated by the incredible amount of computing power and predefined toolboxes available nowadays. Although certainly convenient, such simplicity is bewitching.
As we will show in this chapter, measuring the performance of the physical layer of a mobile communication system can be remarkably similar to simulating it. The main difference is that a measurement has to obey the laws of nature. For example, one cannot just assume perfect channel knowledge, perfect frequency and timing synchronization, known noise variance, double-precision feedback, and so on. One also cannot measure a million independent channel realizations within a small-scale fading scenario. Compared with simulations, all this may seem troublesome; on the other hand, reality is like this.
In other words, simulations reflect the simulation environment (and may reflect reality), whereas measurements do reflect reality.
4.1 Basic Idea
“Simulation is always a form of sampling experiment whenever the model contains one or more stochastic variables (although it is a very special type of sampling experiment since simulations are performed on abstract models instead of real-life objects)” [17, p. xi]. Figure 4.1 shows such an abstract model, which will henceforth be used to obtain the average ...
