Browse Prior Art Database

Original Publication Date: 2001-Apr-18
Included in the Prior Art Database: 2001-Apr-18
Document File: 2 page(s) / 7K

Publishing Venue


Related People

Ed Callaway: AUTHOR [+3]



This text was extracted from a Rich Text Format document.
This is the abbreviated version, containing approximately 69% of the total text.


Ed Callaway, Claire Jackoski and Clinton Powell

Bluetooth is emerging as the de facto standard for personal area networking (PAN); however, it is limited to a maximum one-way data throughput of 723.2 kb/s. It would be desirable to increase this, especially for certain point-to-point wireless applications such as transferring video from a consumer camcorder to a monitor.

Further, one desires not to employ an "exotic" system useable only for this purpose; it would be desirable to use Bluetooth in some way due to its expected market ubiquity and low cost, making only modest changes to software and hardware. In these applications a single, high-speed, asymmetric link is required, and no other devices are active in the piconet.

To meet these requirements, multiple Bluetooth piconets may be used simultaneously, achieving an aggregate data transmission rate higher than that possible in a single piconet. In most applications it is envisioned that each piconet has only two devices in it, a master and a single slave. (Other slaves could, of course, be parked or otherwise inactive in the piconets.)

Without loss of generality, and for the purposes of exposition, assume that the master in each piconet is transmitting data to the slave, which is merely acknowledging correctly received packets. A plurality of fully functioning Bluetooth modules would be in each consumer electronic device and, although all Bluetooth modules are coupled to the same frequency reference, each has its own Bluetooth address.

In this way, the frequency hopping sequence of each module will be independent but, because of the common frequency reference, they will be time-synchronized. Further, due to the time synchronization, all masters will transmit simultaneously and receive simultaneously; this eliminates the problem of one master trying to receive a distant slave while a collocated master is transmitting, and means that the slaves receive signals of similar strength (neglecting multipath effects, which should not be significant due to the 40 dB in-band Bluetooth selectivity specification).

In the simplest embodiment, the transmitted data stream is split into n streams, with each of the n Bluetooth modules transmitting one stream. (The strawman use case would be the transmission of RGB video on three different piconets, one red, one green, and one blue; see Figure 1.) At the receiving device, the received data from the n Bluetooth modules present are then combined to recover the original transmitted data stream.

If one of the slaves had a packet error, for example in the green piconet, a nack is sent to all masters, and all packets on all piconets are resent, to maintain synchronicity. (At the cost of a little more complexity, only the errored piconet could retransmit, with the others silent; this requires the use of an additional flow control command, and complicates the data rate analysis to follow, but in most cases...