Multiplex communications – Communication over free space – Combining or distributing information via time channels
Reexamination Certificate
1999-03-25
2003-08-12
Yao, Kwang Bin (Department: 2664)
Multiplex communications
Communication over free space
Combining or distributing information via time channels
C370S521000
Reexamination Certificate
active
06606312
ABSTRACT:
BACKGROUND OF THE INVENTION
Human verbal communication has evolved into a very effective method of carrying information via sound waves, despite the distortion introduced in the acoustic medium. Although the radio medium is similar to the acoustic medium and presents similar challenges to effective communication, current radio systems do not exploit the natural abilities of the human aural system to deal with these challenges.
Both sound and radio waves propagate not just in a direct path from the transmitter to the receiver, but also by reflections off objects in the environment. This is known as a multipath channel. Reflected signals must travel a further distance than direct signals, therefore they arrive at the receiver later in time. The composite of all the signals from the different paths, each with a different amplitude and delay, make up the multipath channel. For a detailed discussion of the properties of multipath channels see Proakis [1].
The behavior of a signal in a multipath channel depends on whether it is wideband or narrowband. Narrowband signals experience little or no distortion as they pass through the channel other than additive white noise. The received power level does, however, fluctuate drastically due to a process called flat fading, which causes the signal to be lost entirely at times. In contrast, the overall power level of wideband signals is relatively stable. A process called frequency selective fading distorts the wideband signal in the time domain to cause intersymbol interference, and distorts the frequency domain with narrow regions of frequencies that are severely attenuated. Wideband radio systems are desirable because they avoid the problem of narrowband flat fading. However, to use a wideband system the receiver must be able to deal with the distortion of frequency selective fading.
The main difference between the radio and acoustic multipath channels is the delay spread. Delay spread is the difference in time that it takes the signal to pass through the shortest path versus the time through the longest significant path. The inverse of delay spread is roughly the coherence bandwidth of the channel, which is the benchmark for defining the type of the signal. The signal is narrowband if the signal bandwidth is much less than the coherence bandwidth. It's a wideband signal if its bandwidth is much greater than the coherence bandwidth. In a small room, the acoustic delay spread may be 50 ms giving a coherence bandwidth of 20 Hz (the actual value varies considerably). The human voice uses a bandwidth of about 3 kHz, which is much greater than the coherence bandwidth of the acoustic channel and therefore voice is wideband in its natural environment. Radio waves propagate at a much greater speed than sound so even in a large area like a cellular radio cell, the maximum delay spread may be only 50 &mgr;s (again the actual value varies considerably). The coherence bandwidth of the radio channel, 20 kHz in this case, is greater than the bandwidth of the voice signal, therefore voice transmitted through the radio channel behaves like a narrowband signal. This is the reason that voice carried through radio experiences flat fading, and is subject to occasional signal loss, but voice carried through the acoustic medium does not.
One way of combating the effects of multipath is with spread spectrum signals. A spread spectrum signal is, by definition, a signal that occupies a much greater bandwidth than the signaling rate requires. Spread spectrum signals are most useful when the bandwidth of the signal is wide enough to avoid flat fading, while at the same time the signaling rate is low enough to avoid intersymbol interference. Modern direct sequence spread spectrum (DSSS) radios are one example of systems with this property. The human voice is another.
In direct sequence spread spectrum radios, the wideband signal is created by modulating a spreading code. The code is chosen to distribute the energy of the signal in a roughly uniform pattern across the entire frequency band so that there are no critical frequencies in the signal that could be attenuated by a frequency selective fade. As long as enough signal power falls outside the fades, the signal will get through.
The human voice is very similar to a DSSS radio signal. The bandwidth of the voice, about 3 kHz, is much greater than the signaling rate of two to five syllables per second, and therefore the voice is a spread spectrum signal. The voice bandwidth is much greater than the coherence bandwidth of the acoustic channel to avoid flat fading, and the syllables are longer than the delay spread of a typical acoustic channel to avoid intersymbol interference. The three types of sounds that make up a voice signal (voiced sounds, fricative sounds, and plosive sounds [2]) are all inherently resistant to frequency selective fading. The energy of all the sounds is distributed across the voice band to make them wideband, and none of the sounds contain any critical tones that may be lost to a multipath fade.
SUMMARY OF THE INVENTION
When translated to radio, the human voice does not make a good spread spectrum signal because its bandwidth is well below the coherence bandwidth of most radio channels. The purpose of this invention is to coerce the radio medium to behave like the acoustic medium when carrying voice. It does this by artificially increasing the bandwidth of the voice signal through time compression. After decompression, the signal from the radio channel sounds like a natural acoustic signal so there is no need for complex digital signal processing at the receiver to correct for multipath distortion. The signal processing is actually performed by the listener's ear and brain. Since the signal is wideband, it resists the problems of flat fading associated with narrowband radios.
The invention operates as follows. A segment of speech is sampled and stored within the radio. When storage is complete, the voice segment is replayed at a much higher sampling rate. This compresses the signal in time and expands it in frequency. In the example above, the coherence bandwidth of the radio channel is a thousand times that of the acoustic channel, so to achieve the required bandwidth expansion, the signal is played back at a sample rate a thousand times faster. The wideband voice signal is then modulated to radio frequencies using a single sideband (SSB) modulator, amplified, and broadcast though the antenna.
The receiver expands the signal in time to restore it to its original narrow bandwidth. After detecting the radio signal with a SSB demodulator, it is sampled at the high sample rate, stored, and played back at the low sample rate. Not only does this restore the signal bandwidth so the listener can understand the speaker, it also expands the effective impulse response of the radio channel. If for example the radio channel has a delay spread of 50 &mgr;s, the effective delay spread appears to be 50 ms after the signal is expanded which makes it sound like an acoustic channel.
The invention inherits many of the good properties of acoustic voice signals including resistance to flat fading. This allows the average power requirement for this invention to be substantially lower than an equivalent narrowband system. Extra power is normally added to narrowband signals to allow them to pass through all but the deepest flat fades. This fading margin may add 20 dB or more to the output power at the transmitter. The wideband signals of this invention, like acoustic signals, are not as susceptible to flat fading and the fading margin can be virtually eliminated.
Compressing signals in time is known, generally in radio communications, as time compression multiplexing (TCM). TCM was applied to telegraph signals in 1867, to voice telephony as early as 1943, and to radio systems by 1958 [3]. Previous applications of TCM, for example the system described by Jacob and Mattern [4], have been limited to multiplexing two or more signals into a common channel. This invention
Christensen O'Connor Johnson & Kindness PLLC
Telecommunications Research Laboratories
Yao Kwang Bin
LandOfFree
Analog radio system with acoustic transmission properties does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Analog radio system with acoustic transmission properties, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Analog radio system with acoustic transmission properties will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3109259