Improvements to apparatus for decoding error-correcting codes

Claim 3

A method according to claim 1 wherein the receiver:- (a) stores a metric for each end state, and (b) updates each metric for each time slot in accordance with signals received in said time slot, wherein said updating comprises (b)(i) accessing metric storage means providing a storage location for each end state metric with a plurality of addresses for each end state, said addresses constituting the identities of all the legitimate predecessor states of each state whereby a plurality of predecessor metrics are retrieved; (b)(ii) adding to each retrieved metric a signal metric measuring the correlation between the received signal and the legitimate signal corresponding to the transition from the predecessor state to the state whereby a plurality if potential metrics is generated for each state; and (b)(iii) selecting the highest of said potential .metrics as the updated metric for the end state and the state from which it was derived as the assigned predecessor state; whereby performing steps (b)(i), (b)(ii) and (b)(iii) for each end state generates an updated end metric for each state and the decision code is derived from the assigned predecessor state.

Claim 4

A method according to claim 2 wherein each decision code consists of a single bit which takes the value "O" when the assigned predecessor is the 0-predecessor and the value "1" when the assigned predecessor is the 1-predecessor.

Claim 7

A method according to claim 5 where the constraint length is 6 with 64 end states and the retrograde search consists of 32 to 2048 steps.

Claim 8

Receiver adapted to perform a method according to claim 1 to recover output from convolutionally encoded signals which receiver comprises:- (A) a decision code generator for generating a decision code for each end state in each time slot; (B) a decision storage means providing storage locations, addressable by identity of time slot and end state, for path decision codes said code storage means being operatively connected to said generator so as to store each code in the location addressed by the code's end state and the current time slot.

Claim 9

Receiver according to claim 7 which also comprises:- (C) first address register operatively connected to the decision storage means, said first register being adapted to retain a time slot identity as address; (D) second address register operatively connected to the decision storage means, said second register being adapted to retain a state as address; (E) means for updating the content of the first address register to the identity of the preceding time slot; and (F) means for updating the second address register to the state defined by the retrieved decision code and the content of the second register.

Claim 10

Receiver according to claim 8, wherein the second address register is a shift register connected to receive as input the code retrieved from the decision storage means.

Claim 11

Receiver according to any one of claims 7 to 9, which also comprises:- (G) Metric storage means including a storage location for a metric for each end state; (H) Correlating means adapted to produce signal metrics measuring the correlation between a received signal and a legitimate signal; (K) Updating means operatively connected to the correlating means and the metric storage means, said updating means being adapted to update the stored metrics in accordance with the signal metrics, said updating means comprising:- (K)(I) means for accessing the metric storage means with addresses constituting the identities of all the legitimate predecessor states of each end state to retrieve predecessor metrics; (K)(II) adder for adding to each retrieved metric the signal metric corresponding to the transition from the predecessor state to the end state; (K)(III) selection means for selecting the highest sum and its predecessor state corresponding to each end state; (K)(IV) means for transferring said highest sum into the storage location of the metric storage means which location is addressed by the end state; and (L) means for deriving the path decision code from the precessor selected in step (K)(III).