Reduced dimension space-time processing for multi-antenna wireless systems

  • Authors:
  • J. Jelitto;G. Fettweis

  • Affiliations:
  • IBM Res. Lab., Zurich, Switzerland;-

  • Venue:
  • IEEE Wireless Communications
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

The need for wireless communication systems has grown rapidly during the last few years. Moreover, there is a steady growth in the required data rates due to the fact that more and more users request high-bit-rate services. To meet those requirements, current and next-generation wireless systems and networks such as wireless LANs (e.g., IEEE 802.11a) will support much higher data rates compared with established standards. This is basically done by applying advanced transmission schemes and usage of bandwidth resources. Another very promising approach is the introduction of multiple antennas at one or both ends of a link to exploit the spatial dimension of signal transmission for improved link quality and enhanced system capacity. Smart antenna concepts are extensively discussed in this context. The application of concepts with multiple antennas necessitates the introduction of more advanced and computational expensive transmitter and receiver structures, where space-time (ST) processing techniques are required to carry out spatial and temporal information processing jointly. This article introduces a new ST processing concept to enable reduced dimension ST receiver signal processing. The signal dimension can be considerably reduced compared to the number of antennas by exploiting spatial correlation properties of the received antenna signals. The associated signal transformation applies the concept of the Karhunen-Loeve transformation (KLT). A great advantage of the proposed ST processing concept over traditional multiple antenna approaches is the insensitivity of the algorithms to the antenna characteristics and antenna spacing, which allows the use of low-cost antennas. Another significant advantage of the proposed concept is more robust channel estimation due to spatial dimension reduction and the resulting limitation of estimation parameters.