Pulse or digital communications – Synchronizers – Frequency or phase control using synchronizing signal
Reexamination Certificate
1998-11-12
2002-11-12
Chin, Stephen (Department: 2634)
Pulse or digital communications
Synchronizers
Frequency or phase control using synchronizing signal
C370S514000
Reexamination Certificate
active
06480559
ABSTRACT:
BACKGROUND AND SUMMARY OF THE INVENTION
The present application relates to frame synchronization architectures and techniques, particularly in time-division-multiple-access (“TDMA”) mobile phone systems.
Background: Mobile Communication
Mobile communications are a very important area of the electronics industry, and have become an important part of the lives of most citizens in developed countries. One common mobile communications architecture is cellular phone systems. In such systems a mobile phone will establish communication with a nearby stationary transceiver when it is turned on, and the stationary transceiver station provides a connection into the telephone system. The mobile phone will then switch its communication to other stationary transceivers as a user drives around, so that it is always in touch with some transceiver. Thus each transceiver defines one “cell” of wireless interface. Each stationary transceiver can interface to multiple mobile umts at once, and each mobile unit can transfer to a new transceiver when it gets too far from its previous transceiver.
More recently satellite telephone systems have been proposed. In such systems the mobile units can interface to communications satellites rather than to stationary ground units. In some systems low-Earth-orbit satellites are used, since these are many times closer to the surface than are geosynchronous satellites. However, low-Earth-orbit satellites move around the Earth fairly quickly, so a mobile station must be able to transfer from one satellite to another. Here too one satellite transceiver must be able to interface to many mobile units.
Thus each base transceiver must be able to talk to multiple mobile units. One of the basic techniques is TDMA, as described below. For example, TDMA techniques are used in the Digital European Cordless Telephone (“DECT”) system (which is used for mobile telephony in Europe), the Japanese Personal Handy-Phone System (“PHS”), and the MIL-STD-88-183 for UHF satellite systems.
Background: TDMA
One general problem in telecommunications is how to carry multiple signals on one communications channel. TDMA is one of the basic techniques for doing this. (Other quite different ways to do this are frequency division multiple access, or “FDMA,” and code division multiple access, or “CDMA”). In TDMA, a transmitter sends out a series of “frames.” A frame is a sequence of symbols which contains data for several different destinations, strung together in a definite format. Each receiver can (ideally) pick out just the data it needs, from just one part of the frame.
Background: Frame Synchronization
One requirement of TDMA is that each receiver must know where the frame starts. That is, each receiver must become synchronized to the transmitter. This can be difficult.
More precisely, a receiver must synchronize in three ways:
it must synchronize to the transmitter's carrier wave (in an RF system);
it must synchronize to the transmitter's symbol rate and phase, so that the receiver knows where one symbol starts and another ends; and
it must synchronize to the transmitter's frames, so that the receiver knows where one frame starts and another ends.
Each of these synchronizations presents its own difficulties. The present application is particularly directed to the problems of frame synchronization.
FIG. 10
shows a typical TDMA transmission bit stream. The preamble is comprised of a known series of binary bits, called the bit_sync pattern. These bits permit the receiver to acquire the symbol synchronization, and are followed by a unique word which signals the start of a frame. The unique word is the frame_sync pattern (bits u
i
).
Note also that, although the bit_sync pattern is fixed, there is an unknown timing offset between the bit_sync pattern and the start of the unique word. Thus, no simple concatenation of the bit_sync pattern with the unique word will accurately represent the incoming signal stream in all cases.
If frame synchronization errors occur, the receiver will not be able to pick out the correct bits in the stream of incoming bits (most of which are meant for other receivers). However, frame synchronization presents some difficulties. The start of a frame is indicated by a special sequence of bits (called the “unique word”), but it is still necessary for a receiver to recognize this unique word in a continuous stream of bits.
To detect the unique word when it occurs, the receiver typically uses a binary finite-impulse-response (“FIR”) filter structure. Conventionally this is operated as a simple correlation detector. As shown in
FIG. 12
, the filter coefficients are simply the bits u
i
of the unique word U itself, which are multiplied with the bits r
i
of the data stream. A binary register
100
′ holds the bits u
i
of the unique word, and provides them as outputs on 1-bit lines
102
′ to the combinatorial stages
104
′ (which are multipliers in this example, but could also be XOR gates). A series of delay stages
108
′ clock the input bits r
i
through sequentially to the combinatorial stages
104
′. The outputs of the combinatorial stages
104
′ are summed equally by a summer
110
, to produce a raw correlation output &ggr; (which can optionally be normalized by a stage
1212
, to provide a numerically accurate measure of correlation). The output will reach its highest value when each 1 in the coefficient is multiplied with a 1 in the data stream. In this architecture the receiver concludes that frame synchronization has been achieved when the correlator output crosses a threshold. Alternatively, if a bit rotation (shift) of the unique word is used instead, the correlator would operate with the same performance—no better and no worse—except that a different phase shift would be present at the time when the unique word was detected.
Background: Erroneous Sidelobe Acquisition
FIG. 3
shows a plot of the correlation values for a conventional correlation receiver for a unique word of length
24
, assuming that a maximum of 8 bits precede the frame sync (L
b
=8). The variation in correlations as the different phases of the correlation vector are tried can produce “sidelobes,” i.e. spurious local maxima in the correlator output. Such a sidelobe occurs, in this example, at a shift of five units. In conventional systems, such a sidelobe can cause a false detection of frame synchronization, and this can lead to an attempt to lock onto an incorrect frame synchronization value. (The high value at shift
9
shows the correct acquisition of frame synchronization.)
In terms of the system context, a false acquisition of a sidelobe peak can mean complete loss of synchronization between the mobile unit and the system. This will require reacquisition of correct timing from scratch. This can take seconds in some systems. Thus, in a cell phone context, this may result in a noise burst, or in complete loss of a call.
Frame Synchronization with Unique-Word-Dependent Coefficients
The present application discloses that unique word detection (and hence frame synchronization) can be improved by using a correlation coefficient vector which is not equal to the unique word (nor to any shift or multiple of it), but which is dependent on the unique word. In some embodiments, the correlation coefficients are dependent both on the unique word and also on the bit_sync pattern. Preferably the coefficient vector is optimized by picking it to be “near” a set of error offset vectors (using a special minimax metric), to maximally distinguish the unique word from the erroneous candidates. The error offset vectors are derived from a set of erroneous candidate vectors which are all of the same length as the unique word, and which are found within the concatenation of the unique word with the bit_sync pattern.
The disclosed optimal linear receiver provides better sidelobe suppression than the conventional correlation operation. The advantageous result is to decrease the potential for a false detection of the sidelobe thereby increasing the potential fo
Brady III Wade James
Chin Stephen
Kim Kevin
Neerings Ronald O.
Telecky , Jr. Frederick J.
LandOfFree
Frame synchronization with unique-word dependent filter... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Frame synchronization with unique-word dependent filter..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Frame synchronization with unique-word dependent filter... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2955028