Voice over an OFDM based wireless LAN

This paper addresses the problem of efficiently integrating speech communication devices, such as cordless telephones, with an “orthogonal frequency division multiplexing” (OFDM) based wireless LAN. It is argued that traditional “voice over IP” (VoIP) approaches are inefficient for wireless LANs in terms of system overheads, and that more recent proposals, such as the “5-UP” protocol, offer greater promise with the co-existence of OFDM, for high speech packetised data, and simpler (e.g. single carrier) modulation schemes for lower data-rate devices with real time quality of service requirements. A new way of supporting packetised voice is suggested here which is compatible with the IEEE 802.11a and 802.11g standards in “DCF” mode and the “EDCF” features being proposed for 802.11e. The intention is to avoid the “5-up” MAC approach which requires the use of the “PCF” beacon to time-share the different protocols and to preserve full OFDM functionality for high-speed data. The new approach will operate alongside 802.11a or 802.11g systems, being compatible with both the physical and MAC layers. At the physical layer, the packetised speech communication will be based on single carrier (SC) modulation. The speech devices must be compatible with the CSMA collision avoidance protocol, to be capable of detecting whether other OFDM or SC devices are active, but need not be capable of demodulating or transmitting OFDM. For speech communication the new system´s efficiency will be much higher than VOIP because the speech channels will be frequency division multiplexed. The power efficiency of the speech units can also be much higher than if they used full OFDM. Some of the advantages and problems remaining with this approach will be discussed.