In the beginning, mobile voice was deployed in a typical city with one big tower. With a range of over 200 km, this system supported one conversation at a time for a given radio resource*.

1

With the advent of cellular, our example city was covered with 200 sites, each with a range more like 2 km. Because of interference between users in different sites, the network could only support about 30 conversations at one time per resource. [Note that different colors denote different radio resources in our simplified schematics here.]

30

To bring us up to the present, our example adds sectorization, virtually ubiquitous in today's cellular networks. Cell radius stays roughly the same, but each site is subdivided into typically three sectors. This enables roughly 90 simultaneous conversations per resource, since the sectors inherently limit the spread of interference to some degree.

90

Introducing MAS into the picture dramatically improves the efficiency with which radio resources are used. In a system where multiple antennas are installed on the base station, the antennas create a focus of radio energy around each subscriber and create the absence of energy in the direction of other subscribers trying to use the same resource in neighboring sectors and cells. Our example 200-site network can now support 600 simultaneous connections with a single resource. In a system with multiple antennas only on the client devices (which necessarily would have a smaller number of antennas per device), you might still see the connection count rise by 2x — or the data rate for each connection rise by 2x.

600

We have proven, through extensive commercial operation of PHS and now HC‑SDMA broadband wireless base stations, that it is possible to take this a few steps further. With the right enablers in the radio access protocol, we can control the distribution of radio energy so well we can reuse the same resource multiple times within the same sector - in fact for users standing right next to each other. Commercial HC‑SDMA systems have shown real-world performance of on average 2.5 of these 'spatial channels' per sector, performance that would enable up to 1,500 simultaneous connections in our example network here.

If we put multiple antennas on the client device as well (in a so-called MIMO architecture), then these spatial channels can be used to multiply the data rate to individual users. The total number of connections in the network stays the same, but subscribers would be able to use more than one of them at a time to get higher access speeds.

1,500

As you will see in our MAS Principles section, there are many different flavors of MAS. Current system architectures treat these different flavors as separate processing modes, and the ability of a subscriber device or network element to choose the "best tool for the job" in a given moment's radio environment and user behavior profile is limited. Future systems will have MAS and resource scheduling software more tightly integrated, and a more seamless interplay within the same system of the various MAS tools. It is difficult to forecast the efficiency gains these developments will yield, but we have every reason to believe they will be substantial.

???