Simulation of Third Generation CDMA Systems - Virginia Tech

3.6 Code OptimizationThe simulator was implemented in MATLAB to take advantage of the built-infunctions that perform various signal processing application like filtering and Fast FourierTransformation (FFT) and the superior array handling capability. To make the simulator runwithin a reasonable simulation time, we had to optimize the MATLAB code. We generatedthe complex scrambling codes separately and stored them in binary data files. During thesimulation run, the simulator reads the desired scrambling codes form these data files. Thiswas necessary since the generation of long scrambling code takes long time and it would notbe practical to generate the scrambling codes during running the simulator.Figure 3.23 shows the MATLAB profile of the uplink simulator. The simulator is runfor Vehicular A Outdoor Channel with one interfering user. The Rayleigh waveform isgenerated at the chip rate. As we notice the simulator spends significant amount of time inlines 184 and 173. They correspond to the functions implementing MAI and time varyingchannel respectively. The generation of Rayleigh waveform takes up most of the simulationtime in both the functions. However generating Rayleigh waveform at symbol rate ratherthan at chip rate can significantly reduce the simulation time. Another way to reduce thesimulation time is to demodulate the signals from the all the interfering users and count theirerrors towards the BER calculation. However this will increase the complexity of the receiverimplementation and we will not have the option of assigning different spreading factors fordifferent users.55