Multi-antenna-based multi-input multi-output (MIMO) communications first burst onto the scene in the mid-1990s when researchers at Bell Labs and Stanford were looking for ways to increase system throughput without increasing bandwidth. In the decade since, thousands of research papers have been written on the topic dealing with both physical layer (PHY) and network layer ramifications of the technology. MIMO has gone through the adoption curve for commercial wireless systems to the point that today, all high throughput commercial standards (i.e., WiMax, Wi-Fi, cellular, etc.) have adopted MIMO as part of the optional, if not mandatory, portions of their standards. The adoption of MIMO into military wireless communications systems has to some extent lagged its adoption in the commercial arena. To date, the largest DoD-funded effort with a uniquely MIMO-centric focus is the Defense Advanced Research Projects Agency (DARPA) Mobile Networked MIMO (MNM) program. This program is a multiyear, multimillion-dollar effort that looks to exploit MIMO techniques to (a) provide reliable communications in urban canyons; (b) significantly extend the reach of conventional single-antenna wireless systems; (c) improve reliability of individual links: and (d) enhance mobile ad-hoc network (MANET) throughput rate by 10x or more compared to current SISO-based radios. Lucent Technologies was the performer on the first phase of the program, whereas Silvus Technologies was chosen as the performer on the second and third phases of the program.

This article will provide a technical overview of MIMO and its different variants, as well as quantify some of its benefits in relevant military tactical communications. Finally, we will identify key capabilities that efficient MIMO development/evaluation platforms must offer to the marketplace.

The pioneering work by Telatar, Foschini and Gans at Bell Labs demonstrated that MIMO in a wireless communication system can greatly improve performance as much as one order of magnitude or more, without requiring any additional bandwidth.

A MIMO wireless system consists of N transmit antennas and M receive antennas. However, unlike phased array systems where a single information stream, say x(t), is transmitted on all transmitters and then received at the receiver antennas, MIMO systems transmit different information streams, say x(t), y(t), z(t), on each transmit antenna. These are independent information streams being sent simultaneously and in the same frequency band. At first glance, one might say that the transmitted signals interfere with one another. In reality, however, the signal arriving at each receiver antenna will be a linear combination of the N transmitted signals.

Figure 1 shows a MIMO system with three transmit and three receive antennas. The received signals r₁(t), r₂(t), r₃(t) at each of the three received antennas are a linear combination of x(t), y(t), z(t). The coefficients {a_ij} represent the channel weights corresponding to the attenuation seen between each transmit-receive antenna pair. The affect is that we have a system of three equations and three unknowns as shown below.

In general, the matrix, A, of channel co-efficients {a_ij} must be invertible for MIMO systems to live up to their promise. It has been proven that the likelihood for A to be invertible increases as the number of multipaths and reflections in the vicinity of the transmitter or receiver increases. The impact of this is that in a Rayleigh fading environment with spatial independence, there are essentially NM levels of diversity available and there are min(N,M) independent parallel channels that can be established. Increases in the diversity order results in significant reductions in the total transmit power for the same level of performance. On the other hand, an increase in the number of parallel channels translates into an increase in the achievable data rate within the same bandwidth.

Let us now quantify the benefits of MIMO-based systems operating in a typical Rayliegh fading wireless channel. Figure 2 compares the achievable 95-percentile capacity (minimum capacity achieved over 95% of wireless channels encountered, or in other words, given a channel, there is a 95% chance that the capacity of that channel is higher than the capacity shown in the plot) for single antenna systems (yellow dot), for a phased array multi-antenna system (blue curve), and for MIMO systems (red curve). As shown, the capacity of the phased array system grows logarithmically with increasing antenna array size, whereas the capacity of the MIMO system grows linearly. With four antennas, the phase array system provides a capacity of 8 bps/Hz, whereas the MIMO system provides a capacity of 19 bps/Hz. It is also worth noting that in a phased array system, the array coefficients must be calculated to point the beam in the “best direction.” This is quite difficult to do when there are scattering elements within the environment. MIMO systems do not suffer from this problem as the geometry of the environment and position of the reflectors are automatically taken into account during the decoding of the MIMO signal.

The benefits of MIMO will now be considered in a different light. Assume that there is a fixed capacity that is desired, say 1 bps/Hz, and ask the question, “How much total transmit power is needed to achieve a 95-percentile capacity of 1 bps/Hz?” The results are summarized in Table 1. As is seen from the table, as the number of antennas increase in a MIMO system, less and less receive power is needed to achieve the same data throughput rate. So if a conventional single antenna system required 1 Watt of transmit power to achieve a certain throughput, then an 8 × 8 MIMO system would require only 6 mW of power to achieve the same performance.

1 2 Next