22. January 2018 at 9:11 #9247NaserParticipant
In LTE_sim_main_launcher_antenna_separation, the data throughput increase with the spacing among antennas array elements is increased.
Does this take into consideration the increase in grating lobes ?
Naser22. January 2018 at 12:09 #9249
I guess you are referring to the LTE_sim_main_launcher_3D_model. The antenna spacing can be changed in both vertical and horizontal direction when using a planar antenna array. Spacing in horizontal direction refers to the distance between antenna ports, larger the distance less correlated are the channels, while spacing in vertical direction refers to the distance between antenna elements stacked to one antenna port; increasing the distance will cause a sharper beam but also granting lobes will start appearing and increase dependent on the number of antenna elements and distance. This is taken in consideration when generating the small scale fading in +channel_gain_wrappers/TR36873_Fading_3D_Channel.m, e.g. in the function calculate_channel_coefficient_OTOI() parameter element_weight changes according to the antenna element distance.
More granting lobes will increase the interference from neighboring base stations, however the overall gain/decrease in performance will depend on geometry, since when using sharps beams, if the desired base station is pointing towards the user the received signal will be higher, or when interfering BS is pointing toward the granting lobes, the interference will be high. Any combination of these situations will lead to higher/lower SINR. In general, more granting lobes, the throughput should decrease.
Fjolla22. January 2018 at 20:28 #9250NaserParticipant
Thanks for your feedback, really appreciated.
1- Does that also applied to the antenna separation with WINNER2 channel model, similar to the demo in reproduciblity model ?
2- And is the horizontal separation between antenna port produce NO grating lobes in both winner and 3D model?
3- Finally, could you please let me know if there is a limit to the antenna separation (upper bound), what factors control this separation ?, and weather if there is an optimum maximum separation so that the data throughput is maximized ?
Thanks for your help.
Naser23. January 2018 at 13:49 #9251
1. Yes, also applicable to the WINNER2 channel model, however note that in the demo only linear arrays are assumed, you can check in the code if WINNER supports the elevation dimension.
2. The horizontal separation only determines how correlated the channels are from different ports, since there is no connection between elements horizontally. However, it is possible to define antenna arrays where you connect several element horizontally in one port, but we don’t support that currently.
3. If the distance between the adjacent elements is equal or less than the wavelength, linear antenna array has only one beam peak within the visible observation angle region (-90° –90°). When this distance is larger than one wavelength, the unwanted beam peak (grating lobe) occurs in the real angle range of (-90° –90°). Therefore, observation angle range dictates the value of the maximum element spacing to avoid the occurrence of the grating lobe.
To understand how this works, I recommend the following material where this is explained: https://owncloud.nt.tuwien.ac.at/index.php/s/Zw1nol4pHCpOuKj
To access the link, use the password: lte
Fjolla24. January 2018 at 18:27 #9254
Regarding the section of the code below in the 3D model of V1.9 Q2-2016, +macroscropic_pathloss_models/generalPathlossModel.m,
if strcmp(LTE_config.antenna.antenna_gain_pattern, 'TR36.873 3D antenna') % Vertical angle grid % (0?--> zenith, 90?--> perpendicular to the antenna) degrees to cartesian distance_3D = sqrt(distance_matrix_2D(:,:,b_).^2 +(current_site_sectors(s_).tx_height - 1.5).^2); theta_arrival = acosd((current_site_sectors(s_).tx_height - LTE_config.rx_height)./distance_3D); vertical_angle_grid_el = 180 - theta_arrival;
Why is the UE_height = 1.5 and LTE_config.rx_height (which is also 1.5), in 4th and 5th line, respectively? Does this not mean the antenna would be radiating towards height 1.5 m from the ground always?
With the now varying UE_height within the building, one would expect it taken directly from networkMacroscopicPathlossMap.UE_height_map
distance_3D = sqrt(distance_matrix_2D(:,:,b_).^2 +(current_site_sectors(s_).tx_height - networkMacroscopicPathlossMap.UE_height_map).^2); theta_arrival = acosd((current_site_sectors(s_).tx_height - networkMacroscopicPathlossMap.UE_height_map)./distance_3D);
Sherif30. January 2018 at 14:19 #9289
Yes, as you have noticed the varying user height has to be considered in the +macroscropic_pathloss_models/generalPathlossModel.m and this is already considered in release v1.9 Q2-2016. Probably you have the first version of the simulator, in the meantime several bugs have been fixed, so I would suggest to download the latest version of system level simulator from our homepage.
Thanks for your feedback!
Fjolla30. January 2018 at 14:52 #9291
Thank you for your response. I was referring to the v1.9 Q2-2016 (obtained Jan. 2017).
The varying UE_height is actually taken care of in the pathloss calculation but not in the antenna section line 149-150 of +macroscropic_pathloss_models/generalPathlossModel.m (as shown in the code section I mentioned above). Perhaps, one can assume it is a bug.
Sherif30. January 2018 at 16:03 #9292
The bug you are referring to, no longer exists, therefore I suggest you to download the latest version of our simulator under the link:
If you have any troubles downloading, don’t hesitate to contact firstname.lastname@example.org.
Fjolla30. January 2018 at 17:43 #9293
Many thanks Fjolla.
- You must be logged in to reply to this topic.