In a data communication network the routing of data items from a group of eight source processors to a group of eight destination processors is controlled by code values assigned to each of the data values. The data items to be routed are bit-serial binary signals. These are encoded by the source processor using a pair of code values, one for each value of a bit of the binary signal. The eight source processors each encode eight bit-serial signals using respectively different code tables. The code tables are generated as different permutations of a basic code set. Each code table has the property that any additive combination of encoded values for all of the data items in a group produces a unique sum. The eight encoded values produced by each processor are summed to produce a set of analog channel symbols that are transmitted over an analog channel to a central switch fabric. In the switch fabric, the channel symbols are partially decoded to recover the encoded data values. The data values are regrouped according to their destination processors and summed to produce further sets of channel symbols which are transmitted to the destination processors. At the destination processors, the data values are completely decoded and assigned to separate input terminals of the destination processors based on their source processor.

Base radio stations are spatially arranged in a radio transmission system in accordance with a cellular system, separating the message transmission channels from adjacent base stations being effected either by using the frequency-division multiplex method or by using the code-division multiplex method or by using a combination of these multiplex methods. In appropriately large spatial distances the same set of channels can be repeated in a further radio cell. If a mobile radio station moves during the conversation from one radio cell into another, then it is necessary to hand-over the then existing radio connection. To avoid the necessity of using additional receivers in each base station and to prevent a repeated hand-over in the event of high co-channel interferences, the measurements used for hand-over decision are effected in the mobile radio station. As part of the method the reception quality of the co-channel message transmission channels of remote base stations are additionally measured in the mobile station during the existence of a radio connection to a near base station.

A composite signal is formed by simultaneously transmitting multiple asynchronous data bit sequences, that are coded with respective spreading codes, in a single channel; and a circuit is provided which decodes any bit b(x) in that composite signal. This circuit includes a set of filters which are matched to all of the spreading codes and which obtain (a) a matched filter output signal y(x) for the x-th data bit b(x) and (b) matched filter output signals y(x+1) thru y(x+k-1) for the k-1 data bits that immediately follow data bit b(x); K is the number of bit sequences in the composite signal. An arithmetic unit combines the matched filter output signals via the expression: ##EQU1## where H(x,x.+-.i) is the cross correlation of the spreading codes for data bits b(x) and b(x.+-.i) over the time period that those data bits overlap (and thus add) in the composite signal, and ESTb(x-i) is an estimate of data bit b(x-i) which precedes bit b(x). An output unit decodes the data bit b(x) as a "1" if the above expression is positive, and as a "0" if the above expression is negative. In one embodiment, ESTb(x-i) is the SIGN of a matched filter output y(x-i) for data bit b(x-i). In another embodiment, ESTb(x-i) is the result from the output unit which decoded data bit b(x-i).

In a new receiver design, the specific transmitter unit is unchanged from a prior patent application but may also be substantially as described in the prior art. The transmitter is initially activated by the input of the pseudo-random generator code. This code may be any alpha-numeric series, but must be identical to that input to the receiver. When the code is input through the keypad, the transmitter uses internal algorithms to generate a unique shift-register feedback combination. This creates a pseudo-random (PN) code which repeats after a programmable number of bits, depending upon desired maximal signal correlation time. That is the longer the code the longer the correlation time. This PN code sequence is clocked at the desired clock (CHIP) rate inside the transmitter unit. The PN code sequence which is thus generated is then added in phase to the data which is a substantially lower rate than the CHIP rate. After mixing with the PN code and data, the RF frequency CW signal power is spread across a frequency band equivalent to twice the PN code clock rate. The signal power density per hertz is now well below an identical non-spread signal at the communications data rate and ideally below the systems noise level. At the receiver the same pseudo-random code which was used to transmit the information is entered into the receiver unit for activation. This is then used to configure three programmable PN sequence generators which output three identical PN code sequences which vary only in a one-half phase shift of the CHIP rate from each other. The signals received by the antenna are converted to an intermediate frequency and mixed with the locally generated PN code sequences. When code alignment occurs, threshold detector circuitry senses the presence of the correlated data signals. The relative level of the signals are then sensed by the microprocessor which varies the PN clock oscillator to maintain lock with the incoming signal. The data is now available as useable information at the data port.

Method and apparatus for optimally decoding messages from CDMA signal sent by a plurality of users and received asynchronously at one receiver, using a minimal number of computations and minimal memory and processing resources. It is based on a novel approach to signal correlation, as well as to subsequent decoding by decorrelation, that requires a correlation period of only one symbol length, by providing for each user a pair of partial signatures sequences, with which the signal is correlated, deccorelating the results with the inverse of the cross-correlation matrix of all partial sequences, and combining the resultant partial symbol estimates--to obtain final estimated symbol values. The partial sequences are formed by separating each original sequence at a point corresponding to the estimated symbol boundary time, relative to an arbitrary correlation window, of one symbol length. The method can be modified to also apply to the case that any of the signals is received over multiple paths, any path possibly undergoing Doppler shift.