Design and simulation of a DS-TH-UWB system using Simulink/MATLAB for first arriving rays in a non line of sight wireless scenario
An Ultra Wideband system was developed using Simulink in previous work. Using this developed model for Ultra Wideband (UWB) Pulse Position Modulation (PPM) within Simulink, we apply Direct Sequence Walsh codes across Time Hopping Patterns to smooth out the spectral characteristics with Pulse Position Modulation. A non-line of sight Saleh-Valenzuela (SV) model is used to characterize the wireless channel. The system is described and results are presented for the scenario where the first arriving rays are used to detect the transmitted message. For this non-optimum technique (in that for non line of sight the strongest signal rays are available after the first ray arrives) we find that we can to some extent successfully receive a digital bit stream with better than a Bit Error Rate of 0.5.
A simulator was developed to simulate an Ultra Wide Band (UWB) Pulse Position Modulation scheme in MATLAB/Simulink. The channel used in this study is a SV channel as described in the wireless literature and this is again used here in this simulation study. A Gaussian pulse is transmitted across the channel and a Maximal Ratio receiver is used to collect different numbers of arriving rays (using a Rake receiver with L fingers, where L=1 or 4). The number of rays that can be used in the receiver can be selected using the Simulink Model properties. It is also assumed that all the channel gain co-efficients are known perfectly at the receiver (in practice pilot signals would be used to estimate these coefficients but in simulation pilot signals are not required). Around this channel we apply the Direct Sequence Walsh code to spread the signal across the spectrum so that it appears as background noise to other telecommunications systems with no knowledge of the used spreading sequence (which are ideally orthogonal to other such sequences). Inside the simulator we enforce a set of eight time slots by eight time slots where each time slot is one hundred nanoseconds in time. A Time Hopping pattern is then implemented at the transmitter. This required the development of a Time Hopping Latch (whose construction will be described) which allowed the signal to only be monitored when an active time slot was experienced. That is, a time slot which contained a Gaussian pulse indicating that a one or zero pulse had been transmitted with a time shift of either zero or one nanosecond.
The measured results from the simulator show that the measured BER (Bit Error Rate) drops from 0.5 down to approximately 0.2 when using the first arriving ray only (L=1) as the measured Energy per symbol to noise ratio (Eb/No) is varied from -20dB to 10dB in steps of 2dB. When the first four arriving rays are used over the same range of Eb/No the measured BER drops from 0.5 to approximately 0.02. Showing that even the least complex algorithm used to select the transmitted binary symbol stream, given perfect knowledge of the channel gain coefficients, can result in useful information transfer across the wireless medium.