Superimposed training for single carrier transmission in future mobile communications
Levanen, Toni (2014)
Levanen, Toni
Tampere University of Technology
2014
Teknis-taloudellinen tiedekunta - Faculty of Business and Technology Management
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:ISBN:978-952-15-3345-7
https://urn.fi/URN:ISBN:978-952-15-3345-7
Tiivistelmä
The amount of wireless devices and wireless traffic has been increasing exponentially for the last ten years. It is forecasted that the exponential growth will continue without saturation till 2020 and probably further. So far, network vendors and operators have tackled the problem by introducing new evolutions of cellular macro networks, where each evolution has increased the physical layer spectral efficiency. Unfortunately, the spectral efficiency of the physical layer is achieving the Shannon-Hartley limit and does not provide much room for improvement anymore.
However, considering the overhead due to synchronization and channel estimation reference symbols in the context of physical layer spectral efficiency, we believe that there is room for improvement. In this thesis, we will study the potentiality of superimposed training methods, especially data-dependent superimposed training, to boost the spectral efficiency of wideband single carrier communications even further.
The main idea is that with superimposed training we can transmit more data symbols in the same time duration as compared to traditional time domain multiplexed training. In theory, more data symbols means more data bits which indicates higher throughput for the end user. In practice, nothing is free. With superimposed training we encounter self-interference between the training signal and the data signal. Therefore, we have to look for iterative receiver structures to separate these two or to estimate both, the desired data signal and the interfering component.
In this thesis, we initiate the studies to find out if we truly can improve the existing systems by introducing the superimposed training scheme. We show that in certain scenarios we can achieve higher spectral efficiency, which maps directly to higher user throughput, but with the cost of higher signal processing burden in the receiver. In addition, we provide analytical tools for estimating the symbol or bit error ratio in the receiver with a given parametrization.
The discussion leads us to the conclusion that there still remains several open topics for further study when looking for new ways of optimizing the overhead of reference symbols in wireless communications. Superimposed training with data-dependent components may prove to provide extra throughput gain. Furthermore, the superimposed component may be used for, e.g., improved synchronization, low bit-rate signaling or continuous tracking of neighbor cells. We believe that the current systems could be improved by using the superimposed training collectively with time domain multiplexed training.
However, considering the overhead due to synchronization and channel estimation reference symbols in the context of physical layer spectral efficiency, we believe that there is room for improvement. In this thesis, we will study the potentiality of superimposed training methods, especially data-dependent superimposed training, to boost the spectral efficiency of wideband single carrier communications even further.
The main idea is that with superimposed training we can transmit more data symbols in the same time duration as compared to traditional time domain multiplexed training. In theory, more data symbols means more data bits which indicates higher throughput for the end user. In practice, nothing is free. With superimposed training we encounter self-interference between the training signal and the data signal. Therefore, we have to look for iterative receiver structures to separate these two or to estimate both, the desired data signal and the interfering component.
In this thesis, we initiate the studies to find out if we truly can improve the existing systems by introducing the superimposed training scheme. We show that in certain scenarios we can achieve higher spectral efficiency, which maps directly to higher user throughput, but with the cost of higher signal processing burden in the receiver. In addition, we provide analytical tools for estimating the symbol or bit error ratio in the receiver with a given parametrization.
The discussion leads us to the conclusion that there still remains several open topics for further study when looking for new ways of optimizing the overhead of reference symbols in wireless communications. Superimposed training with data-dependent components may prove to provide extra throughput gain. Furthermore, the superimposed component may be used for, e.g., improved synchronization, low bit-rate signaling or continuous tracking of neighbor cells. We believe that the current systems could be improved by using the superimposed training collectively with time domain multiplexed training.
Kokoelmat
- Väitöskirjat [4862]