Carlos,
I'm one of the authors of that VRD. Your question isn't coming through clearly.
The design we are recommending should not cost less than hallway APs, it should cost more.
APs seeing each other in the hallway is certainly an issue with hallway placements. However, the real problem is propagation losses through walls, doors and furniture.
The core point of the VRD is that trying to get away "on the cheap" with a few APs to cover a whole floor from a hallway is incompatible with today's user expectations in the rooms themselves. The only way to meet the need is to go denser and into the rooms.
The level of density depends on the construction type. In the paper - for simplicity - we divide the world into 2 kinds of buildings - high attenuation (HA) and low attenuation (LA). Meaning 10-20dB per wall for the former, and 3-6dB per wall for the latter.
In an HA building, you must put APs in each and every room. We've tested this extensively in many real customers. We find over and over again that in HA, the rooms to the left and right of a room with an AP have very poor throughput (even if they are showing 4-5 bars). Remember that "bars" are just beacon strength and beacons are transmitted at lowest OFDM rate and therefore have highest power. Actual higher level modulations like MCS6/7 or MCS14/15 are sent with significantly LESS transmit power due to backoff requirements in the radio power amplifer design to maintain signal linearity.
In an LA building, our position is you can cover 1 room on either side of the room with the AP. In other words, 1 AP per 3 rooms.
You actually do not need hallway APs in this design in most cases.
So in your example of 24 rooms, in an LA building you need 8 APs which is twice the number you said you had now. In an HA building you need 24 APs.
In LA you run hotter TX power, in HA you run very low power because you don't need to go past the first wall and we want to limit CCI.
-cl