Method and apparatus for decoding multiple bit sequences that are transmitted simultaneously in a single channel

• US 4,908,836 A
• Filed: 10/11/1988
• Issued: 03/13/1990
• Est. Priority Date: 10/11/1988
• Status: Expired due to Term

• Current Assignee: Unisys Corporation Blue Bell PA
• Original Assignee: Unisys Corporation
• Inventors: Short, Robert T., Xie, Zhenhua, Rushforth, Craig K.
• Primary Examiner(s): Safourek, Benedict V.
• Assistant Examiner(s): Edwards, Christopher O.
• Application Number:
  US07/255,809
• Time in Patent Office: 518 Days
• Field of Search: 375/1, 375/96, 375/99, 375/101, 375/103, 364/724.11, 364/728.01, 364/724.12, 364/728.05, 370/95, 370/3, 370/19, 370/18
• US Class Current: 375/152
• CPC Class Codes: H04B 1/7103 the interference being multiple access interference

Abstract

Data bits are decoded from a composite signal that is formed by coding multiple bit sequences with respective spreading codes, and transmitting the coded bit sequences simultaneously and asynchronously over a single channel in which the bit sequences are added. This decoding involves a metric in combination with a repetitive decision process which is only linearly dependent on the number of bit sequences in the composite signal.

Claims

1. A method of decoding superimposed data bits including the steps of:

receiving a composite signal that is formed by coding multiple bit sequences with respective spreading codes and transmitting the coded bit sequences simultaneously and asynchronously over a single channel;

comparing the content of several registers, each of which contains an estimate of a particular bit sequence in said composite signal and a metric for that sequence, to determine the largest metric M_max and the corresponding bit sequence b_s ;

evaluating a new pair of metrics for the very next bit b(x) in said composite signal which follows said bit sequence b_s by using estimated "0" and "1" values for bit b(x), bits from said sequence b_s which immediately precede bit b(x), and matched filter outputs for bits in said composite signal which immediately follow bit b(x);

replacing the register which contains said bit sequence b_s with two registers, one of which contains the bit sequence b_s b(x)=0 and the metric M_max plus the above evaluated metric for b(x)=0, and the other of which contains the bit sequence b_s b(x)=1 and the metric M_max plus the above evaluated metric for b(x)=1;

repeating said comparing, evaluating, and replacing steps multiple times; and

thereafter,

decoding said composite signal as the bit sequence in the register which contains the largest metric.

2. A method according to claim 1 wherein each metric of said pair that is evaluated for bit b(x) includes an estimated value for bit b(x) times a matched filter output for bit b(x), minus an interference term I_f which contains the matched filter outputs of the K-1 bits that immediately follow bit b(x), and minus an interference term I_p which contains K-1 bits from the sequence b_s, where K is the number of bit sequences in said composite signal.

3. A method according to claim 2 wherein said term I_f is of the form ##EQU1## where y(i) is the matched filter output for bit b(i), and H(x,i) is the cross correlation between the spreading codes for bits b(x) and b(i) over the period that bits b(x) and b(i) overlap in said composite signal.

4. A method according to claim 3 wherein said term I_p is of the form ##EQU2## where H(x,i) is the cross correlation between the spreading codes for bits b(x) and b(i) over the period that bits b(x) and b(i) overlap in said composite signal.

5. A method according to claim 4 wherein said metric for bit b(x) includes the terms 2b(x)y(x)-b² (x) W(x)-2b(x)I_p (x)-2b(x)I_f (x)-N_o 1n2 where W(x) is the power level of bit b(x) and N_o is the power level of any noise in the composite signal.

6. A method according to claim 5 wherein said composite signal is transmitted over said channel via radio waves.

7. A method according to claim 5 wherein said composite signal is transmitted over said channel via an electrical conductor.

8. A decoder for decoding superimposed data bits comprising:

a plurality of filters, which are coupled to receive and filter a composite signal that is formed by coding multiple bit sequences with respective spreading codes and transmitting the coded bit sequences simultaneously and asynchronously over a single channel, with each filter being matched to a respective one of said spreading codes;

a computing means for evaluating a metric for any bit b(x) in said composite signal which includes an estimated value for bit b(x), an estimate of a sequence of bits b_s which immediately precede bit b(x), and matched filter outputs for several bits in said composite signal which immediately follow bit b(x);

a plurality of register means, each of which is for holding a respective metric and a corresponding bit sequence; and

a control means for repeatedly (a) examining said registers to determine its largest metric M_max and corresponding bit sequence b_s, (b) directing said computing means to compute a new pair of metrics for the bit sequences b_s b(x)=0 and b_s b(x)=1, and (c) replacing the register which contains said bit sequence b_s with two registers, one of which contains the bit sequence b_s b(x)=0 and the metric M_max plus the above evaluated metric for b(x)=0, and the other of which contains the bit sequence b_s b(x)=1 and the metric M_max plus the above evaluated metric for b(x)=1.

9. A decoder according to claim 8 wherein said computing means evaluates a metric for bit b(x) which includes an estimated value for bit b(x) times a matched filter output for bit b(x), minus an interference term I_f which contains the matched filter outputs of the K-1 bits that immediately follow bit b(x), and minus an interference term I_p which contains K-1 bits from the sequence b_s, where K is the number of bit sequences in said composite signal.

10. A decoder according to claim 9 wherein said terms I_f and I_p are of the form ##EQU3## and ##EQU4## where y(i) is the matched filter output for bit b(i), and H(x,i) is the cross correlation between the spreading codes for bits b(x) and b(i) over the period that bits b(x) and b(i) overlap in said composite signal.