Reliability vs latency

Geometry

Import the required city geometry into AEDT. Choose base station and user device locations. Select the antenna type from the parametric antennas dictionary or design your own antennas/arrays in HFSS. In this example workflow, we import a part of downtown Los Angeles from open street maps. Both transmitting and receiving terminals are equipped with one Hertzian dipole antenna.

Set-up

Configure one terminal as transmitter and another one as receiver. Create a frequency sweep set-up with a specified center frequency and bandwidth. For this example, we simulate at a frequency of 5 [GHz] with 100 [MHz] of channel bandwidth.

Post-Processing

Assign transmit power to the transmitting terminal and noise power to the receiving terminal. In this example, we use 1 [W] of transmit power per subcarrier and -90 [dBm] of noise power. Calculate the signal-to-noise ratio (SNR) per subcarrier with perfect channel state information (CSI) as well as with channel error. Calculate the asymptotic finite blocklength bound on block error probability to find the target packet length.

Signal-to-Noise Ratio (SNR)

In this example, we plot the SNR vs frequency with perfect CSI as well as using a scalar LMMSE estimation algorithm to estimate the channel from pilot training symbols. The channel error is treated as a separate noise term. As can be seen, with only a single pilot used for channel estimation, the SNR is significantly lower than what could be achieved with perfect channel knowledge. As the number of pilots grow the SNR quickly approaches the perfect CSI case.

Block Error Probability

In this plot, we use a 5 [GHz] carrier and plot the block error probability vs blocklength using the Polyanskiy finite blocklength limit. The blocklength spans from 100 to 200 bits. We plot the block error probability for the perfect CSI scenario as well as with using an LMMSE channel estimate obtained with 10 pilot symbols. As can be observed, with channel error, a significantly longer code is required to achieve the same error probability as the perfect CSI case.

A Holistic Approach to Wireless Design Through Physics Based Simulations