Published: 12.03.08
New approaches for wireless communication

Distinctly higher transmission rates

In theory, wireless networks can transfer only a limited amount of information. ETH Zurich researchers have now shown how the limits of conventional W-LAN networks can be significantly overcome.

Felix Würsten
The new test bed’s MIMO transceiver. The four antennas, which can transmit or receive on the same frequency simultaneously, are clearly recognisable. The chip in the centre is used for signal processing and control.
The new test bed’s MIMO transceiver. The four antennas, which can transmit or receive on the same frequency simultaneously, are clearly recognisable. The chip in the centre is used for signal processing and control. (large view)

It is a battle with the boundaries of physics: ever since information transmission theory was developed, it has been known that only a limited amount of data can be transferred by wireless communication within a defined frequency band. Since that time, attempts have been made to exploit as fully as possible this quantity defined by physics. In view of the considerable amounts of data transmitted nowadays via mobile phone networks and W-LAN links, there is no doubt that the previous efforts were very successful. Nonetheless, scientists are searching for new ways to transfer even more data – because transmission capacities are a scarce and thus also a valuable commodity that must be used efficiently.

Messages from babbling voices

Two professors from Stanford had already shown in the nineties that theoretically far more information can be transferred in wireless networks than was originally assumed. They suggested the MIMO technique as a new approach. MIMO stands for “Multiple Input Multiple Output”, and in concrete terms it means that communication is no longer between only one transmitter and one receiver. On the contrary, several antennas can broadcast several messages simultaneously in the same frequency channel. The receiver equipment also has several antennas and can use them to retrieve the original information from the overlapping signals by sophisticated signal processing.

Helmut Bölcskei, Professor at the Communication Technology Laboratory of ETH Zurich, explains that the principle is “like a conversation with several people speaking to several other people at the same time. At first glance it seems that each listener perceives just an incomprehensible babel of voices. However, each listener hears a slightly different noise because the sound waves are scattered in the room, which creates interference effects. If the listeners combine what they hear skilfully, they can filter the original message from the babble. This enables significantly more information to be communicated than is possible with conventional technology.”

Proven feasibility

Even though the theory of the MIMO technique clearly makes sense, experts debated for a long time as to whether the approach is at all feasible in actual practice. Four years ago the ETH Zurich researchers had already used an experimental system to prove that the MIMO technique works – although with only one user at that time. In the context of the European research project “Mascot”, in which the group led by Wolfgang Fichtner, Professor at the ETH Zurich Integrated Systems Laboratory, is also participating, the plan is now to show that the capacity gains achieved can be implemented in complex networks with several users. In this project the ETH Zurich researchers have successfully built a compact lab-scale system with four antennas. The scientists presented this system, known as the MIMO test bed, to interested experts at a recent workshop.

The crux of their work was to develop a way of decoding the signals as efficiently as possible, so as to unravel the tangle of signals in the receiver. The researchers realised they faced conflicting challenges in this task: the more antennas the system has, the more data can in principle be communicated, but decoding also becomes all the more laborious as a result. Because the receiver antennas are to be installed in devices manufactured as economically as possible, the signals must be decoded using the smallest possible chip. However, the smaller the chip, the less computing power is available.

Bölcskei says “We were already able to show four years ago that a MIMO system functions with four transmitters. We have gained a considerably better understanding of the basic theoretical principles since then, which has allowed us to develop powerful and efficient decoding algorithms. This paved the way for the construction of compact receivers.” The receivers developed at ETH Zurich are so efficient that they can immediately be installed in ordinary commercial laptops and W-LAN stations. According to Bölcskei, “This would have been impossible without close collaboration between communications theory and chip design (VLSI design).” The plan is now to strengthen interdisciplinary collaboration further in the next few years, through the appointment of Andreas Burg as holder of an SNF Professorship.

A unique device

Bölcskei explains that “In the shape of our MIMO test bed we possess a device that is unique anywhere in the world, and which we can now use to try out new ideas and development approaches.” With their experimental system, the researchers achieve a spectrum utilisation up to four times higher compared to present-day W-LAN networks.

Bölcskei is convinced that MIMO is the technology of the future, because the industry is currently also engaged in defining standards for such networks. On the other hand it is slightly surprising that apparently only a few companies are active in this field. The ETH Zurich researcher thinks that “Many major companies have missed the boat in this very promising field. This is why we founded the spin-off company Celestrius, which now plans to market our development.” He says this will involve a reciprocal exchange between his laboratory and the young company. “Many current research questions arise from implementing technology. Developing a system in the laboratory in one thing, but incorporating the technology into marketable products is quite another. It suddenly raises issues that we never thought of – and these are precisely the aspects that are immensely exciting for researchers like us.”

Reader comments: