Is this possible?

Hard to say, without seeing more about their proposal than is said in the proposal.

From the way that they say they're planning to "synthesize" an individual cell site around each phone by *using* interference, I infer that it might be some sort of phased-wave or "multiple-in, multiple-out" arrangement. 802.11n and similar systems have been doing MIMO for some time now.

Doing MIMO or phased transmission and reception using *multiple* access point radios, as they say they intend to do, would seem to me to be quite tricky. You could end up needing to have extremely precise frequency and time synchronization between the radios. Cell sites today need to have this - they've used rubidium time standards, GPS-disciplined oscillators, etc. in order to keep all of the sites in a given area well-synchronized. Doing this with a multiplicity of simple, inexpensive, ubiquitous access points would be interesting, to say the least.

...

Yeah, looking at some of the articles and commentary posted in the last day or so seems to say that this is what they're planning to do. It will require a large data center in each service area to do the "data crunching" needed to design the wireless data-signal streams for the devices using the system. Since each device will be moving around more or less constantly, the data patterning algorithms will have to change all the time.

In effect, it's sort of a "centralized MIMO". Instead of the antennas in a single access-point (or cell tower) being managed by a controller at that AP, you'd have a whole cluster of radios scattered around a metro area being used as "phased array antennas" by a data-center controller. Rather than transmitting to one cellphone per frequency at a time, you'd be transmitting to a whole bunch of cellphones simultaneously, from a whole bunch of scattered transmitters. The transmitted signals would combine differently at each cellphone (based on the distances from the phones to each transmitter).

According to one article I read, Perlman claims that it will work with existing LTE-capable cellphones and other LTE devices.

This will be interesting to watch, and see just where the actual bottlenecks turn out to be.

I don't think so. A bit more detail:

Methinks I see some problems:

1. It works in only one direction. Cancellation (interference) of the RF signal might possibly work for the signal from tower to handset, but the return path from the handset is not so well controlled. A 1 cm cell at 700 MHz requires a stability of: 1/43 * 360 = 8.4 degrees which methinks rather tight for fixed antenna, and probably impossible for a moving handset.

1. It does nothing for the basic channel loading and density problems. For LTE (or any scheme that evens out the peak power using spread spectrum), the base line noise level is determined by the number of channels transmitted and their power level. If this scheme can increase the number of handsets that operate in a given airspace, the base line noise level in that space will increase proportionally. Note that the power required needed to communicate a given distance at a given SNR (signal to noise ratio) is unchanged, so the noise goes up with more handsets.

2. The scheme seems to rely on RF cancellation (nulling). That's fine for a CW signal, but I doubt it can be done with a spread spectrum signal where every phase/amplitude slot in the constellation diagram is already occupied. It might be possible to achieve cancellation if there were a carrier, but there is not carrier. Even if it were possible to cancel some of the phase/amplitude points, the thruput would suffer.

1. Much of the data performance of LTE relies on the spatial multiplexing of MIMO. This requires that the signal path between orthogonal signals be different lengths in order to function. Handset have multiple antennas to do this, as do the base stations. With different path lengths, creating a null with a single antenna isn't going to work. The photos only show a single (gold plated?) base antenna. The phase cancellation would also need to be done for each MIMO stream separately. Four antennas?

2. The antenna looks very odd. I'm quite familiar with various antenna designs and I've never seen anything quite like that pancake. I'm at a loss as to how it works or at what frequency it's expected to operate.

1. For stadiums, cellular providers usually use lower power "small cell" system such as DAS (distributed antenna systems). This allows frequency reuse by simply limiting the size of a cell. I don't see the need for this scheme: Most of the larger DAS systems and vendor and modulation independent allowing multiple carriers to share the same DAS equipment. It seems that the Artemis scheme requires duplicated equipment for each cellular provider.

2. I think this is the applicable patent application: The application is HUGE and will take me some time to read. In addition, much of it refers to DIDO which is allegedly described in:

1. A media blitz is not a proper or common way to sell cell phone network equipment. From my perspective, it looks like he tried to sell DIDO to the various Wi-Fi vendors, with poor results. So, now he's switching to cellular providers by sell the idea to users, apparently without an operational demo system. Something is not quite right here.

The backhaul from the TX network nodes (assuming and it is a *VERY* big assumption that they can make a phased array of cheap emitters work on the scale necessary to do this). I suspect they might be able to make the signal better for a handful of lucky customers at a time but only if they keep their phones rock steady and nothing conductive moves about anywhere near them. A stadium full of people will brick it!

Sounds extremely like the sort of thing that would fall foul of advertising standards in the UK since the physics doesn't add up.