Chapter 3
On Building Testbeds
Setting up a testbed for wireless multi-antenna measurements can be considered a straightforward task. Several companies promise off-the-shelf working baseband hardware and software, whilst other companies offer the high-frequency hardware also required.1 To put the elements together, “nothing more” than a skilled engineer is required.
An error-free testbed that is able to “reliably” carry out novel “outdoor” wireless measurements 24 hours a day, seven days a week, is another story. It is a story of ongoing redesign, spending five times as much on accessories than on the actual testbed hardware, always searching for bugs and faulty hardware, and never having enough computing, let alone manpower.
This chapter will first introduce our testbed design before reporting on possible pitfalls regarding testbed design in general.
Figure 3.1 Development of our testbed. It took about 4 years from the first publication in 2004 to actually carry out realistic real-world outdoor-to-indoor throughput measurements at the end of 2007.
3.1 Basic Idea
To carry out measurements as described in all the following chapters we designed a 4 × 4 Multiple-Input Multiple-Output (MIMO) testbed that consists of the following key parts (see Figure 3.2):
TX unit: The TX unit is capable of transmitting pre-generated, complex-valued baseband data samples on four antennas at a center ...
