Patent Application
20200028525
The present invention relates to wireless communication systems, and, more particularly, to decoders and decoding methods for wireless communication systems.
As low power consumption has been more and more important in Low-Power Wide-Area Network (LPWAN), various product makers have been attempting in recent years to achieve coverage enhancement under a limited system resource. Driven by the market, the concept of the Internet of Things (IoT) is becoming more mature. A large number of devices need to be connected to the network, and some of which require only a small amount of data transmission but a long waiting time. Because these devices are designed to possess low energy consumption, low complexity, low cost, high coverage, and other characteristics, and there is a good change that these devices operate in an environment in which the signal-to-noise ratio (SNR) is poor (e.g., operate in the cell edge or basement), the transmitting end (e.g., the base station) repeatedly transmits the signals in an effort to help the receiver to correctly decode the signals. In order to effectively improve the demodulation accuracy, the receiving end must spend a long time waiting and receive these repeatedly transmitted signals. However, this extended computing time increases power consumption.
Therefore, it becomes an important issue to improve the bit error rate (BER) performance at the receiving end to thereby reduce power consumption and costs and increase the service life of the battery.
In view of the issues of the prior art, an object of the present invention is to provide a convolutional code decoder and a convolutional code decoding method, so as to make an improvement to the prior art.
A convolutional code decoder is provided. The convolutional code decoder performs decoding to obtain a target data according to a received data and an auxiliary data. The convolutional code decoder includes a first error detection data generation circuit, a channel coding circuit, a first selection circuit, a first Viterbi decoding circuit, a second error detection data generation circuit, a comparison circuit, a second selection circuit, and a second Viterbi decoding circuit. The first error detection data generation circuit is configured to perform an error detection operation on the auxiliary data to obtain an error detection data. The channel coding circuit is coupled to the first error detection data generation circuit and configured to perform channel coding on the auxiliary data and the error detection data to obtain an intermediate data. The first selection circuit is coupled to the channel coding circuit and configured to generate a first to-be-decoded data according to the received data and the intermediate data. The first Viterbi decoding circuit is coupled to the first selection circuit and configured to decode the first to-be-decoded data to obtain an intermediate decoded data. The intermediate decoded data includes an estimated data and an intermediate error detection data. The second error detection data generation circuit is coupled to the first Viterbi decoding circuit and configured to perform an error detection operation on the estimated data to obtain a reference error detection data. The comparison circuit is coupled to the second error detection data generation circuit and the first Viterbi decoding circuit and configured to compare the intermediate error detection data and the reference error detection data. The second selection circuit is coupled to the channel coding circuit, the second error detection data generation circuit, and the comparison circuit and configured to generate a second to-be-decoded data according to the received data, the intermediate data, and one of the intermediate error detection data and the reference error detection data. The second Viterbi decoding circuit is coupled to the second selection circuit and configured to decode the second to-be-decoded data to obtain the target data.
A convolutional code decoding method is also provided. The convolutional code decoding method performs decoding to obtain a target data according to a received data and an auxiliary data. The convolutional code decoding method includes the following steps: performing an error detection operation on the auxiliary data to obtain an error detection data; performing channel coding on the auxiliary data and the error detection data to obtain an intermediate data; generating a first to-be-decoded data according to the received data and the intermediate data; decoding the first to-be-decoded data to obtain an intermediate decoded data, the intermediate decoded data including an estimated data and an intermediate error detection data; performing an error detection operation on the estimated data to obtain a reference error detection data; comparing the intermediate error detection data and the reference error detection data; generating a second to-be-decoded data according to the received data, the intermediate data, and one of the intermediate error detection data and the reference error detection data; and decoding the second to-be-decoded data to obtain the target data.
The convolutional code decoder and the convolutional code decoding method of the present invention use predicted data to decode. Compared with the conventional technology, the convolutional code decoder and the convolutional code decoding method of the present invention can demodulate/decode signals more quickly. One of advantages of early completion of the demodulation/decoding process is early termination of the operation state, which can save power.
These and other objectives of the present invention no doubt become obvious to those of ordinary skill in the art after reading the following detailed description of the preferred embodiments with reference to the various figures and drawings.
The following description is written by referring to terms of this technical field. If any term is defined in this specification, such term should be explained accordingly. In addition, the connection between objects or events in the below-described embodiments can be direct or indirect provided that these embodiments are practicable under such connection. Said “indirect” means that an intermediate object or a physical space exists between the objects, or an intermediate event or a time interval exists between the events.
The disclosure herein includes convolutional code decoders and convolutional code decoding methods. On account of that some or all elements of the convolutional code decoders could be known, the detail of such elements is omitted provided that such detail has little to do with the features of this disclosure and this omission nowhere dissatisfies the specification and enablement requirements. Some or all of the processes of the convolutional code decoding methods may be implemented by software and/or firmware, and can be performed by the convolutional code decoders or their equivalents. A person having ordinary skill in the art can choose components or steps equivalent to those described in this specification to carry out the present invention, which means that the scope of this invention is not limited to the embodiments in the specification.
The following description of the Viterbi architecture is based on (2,1,2) convolutional code encoding and hard decision. Those skilled in the art, however, may apply this invention to different types of convolutional coding and soft decision after having an understanding of the embodiments of the invention.
In this invention, the demodulation and/or decoding scheme of the receiving end is suitable for a communication system based on a general modulation technique, such as a communication system using a convolutional code encoder/decoder. Such communication systems can be, for example, the Internet of Things (IoT), machine-to-machine (M2M) systems, 802.11ah HaLow of Wireless Fidelity (Wi-Fi), etc. The following description is based on a narrow band IoT (NB-IoT) receiver, but the present invention is not limited to this communication system.
In network systems, during data transmission, needs for transmitting repeated data are common, or partially identical (thus predictable) and partially unknown data are transmitted, but all data must be processed together in the demodulation process, resulting in some unnecessary errors and the increase of unnecessary power consumption. The present invention therefore provides a scheme that utilizes known data to facilitate the demodulation and/or decoding process(s). These known data can be obtained by the receiving end's prediction for the value(s) of certain bit(s) of the current signal based on statistics and analysis of previously transmitted data over a relatively long period of time. The method for generating these known data can be referred, but not limited, to U.S. patent application Ser. No. 16/251,685.
Since the CRC-based error detection data B is associated with all bits of the original data A, when the original data A has one or several unknown bit(s) (marked in gray), the error detection data B (the last NB bits of the data. C) are considered unknown bits (also marked in gray). Next, unknown bits are increased in the data Z owing to the properties of the convolutional code; the number of unknown bits is related to the code rate and the constraint length of the convolutional code. Rate matching and scrambling do not change the ratio of the number of unknown bits to the number of known bits but affect the positions of the unknown bits in the data Y to be transmitted.
The Viterbi algorithm is extensively used in convolutional code decoding.
Because some or all of the bits of the to-be-decoded data E are known (in comparison, all bits of the received data {circumflex over (Z)} are unknown, that is, whether the bit values of the received data {circumflex over (Z)} are correct is undetermined), the Viterbi decoding circuit 340 can generate the target data Ĉ more accurately. Therefore, the convolutional code decoder 250 of the present invention has good performance. As a result, the demodulation and/or decoding time at the receiving end in the wireless communication system can be shortened, thereby reducing power consumption of the wireless devices.
m
accum,0
=m
prev,0
+m
cur,0 (1)
m
accum,1
=m
prev,1
+m
cur,1 (2)
where mcur,0 and mcur,1 are the current metrics corresponding to branch b0 and branch b1, respectively, and mprev,0 and mprev,1 are previously accumulated metrics corresponding to branch b0 and branch b1, respectively.
The default value mpreset is related to how the Viterbi decoding circuit 510 determines the branch (step S760). In step S760, the Viterbi decoding circuit 510 selects one of the two branches of the target state as the survivor path, records the source of the selected branch (e.g., using a bit 0 or bit 1 to indicate the source), and finally updates the previously accumulated metric of the next stage with the accumulated metric. If the default value mpreset is a maximum value (e.g., +2N−1, N being the number of bits of the default value mpreset), the Viterbi decoding circuit 510 selects in step S760 a branch having a smaller accumulated metric. Alternatively, if the default value mpreset is a minimum value (e.g., −2N−1), the Viterbi decoding circuit 510 selects in step S760 a branch having a greater accumulated metric.
Reference is made to
If step S710 is negative, the flow proceeds to step S720, in which whether the information bit of the auxiliary data A′ is known is determined. If, for example, the auxiliary data A′ has multiple information bits (A′0, A′1, A′2, . . . , A′k−1, A′k, A′k+1, . . . ), then whether branch b0 or branch b1 will be selected is highly relevant to the value of information bit A′k−1.
When the information bit A′k−1 is unknown (step S720 is negative), the Viterbi decoding circuit 510 calculates the current metric of each branch (step S740), calculates the accumulated metrics according to the equations (1) and (2) (S750), and then determines the branch (step S760). After step S760 is complete, the Viterbi decoding circuit 510 executes the flow of
When the information bit A′k−1 is known (step S720 is positive), the Viterbi decoding circuit 510 determines whether the target state is a candidate state (step S730). Assuming in
When the target state is the candidate state (step S730 is positive), the Viterbi decoding circuit 510 performs step S740 to S760. When the target state is not the candidate state (step S730 is negative), the Viterbi decoding circuit 510 performs step S715. More specifically, when determining that the target state will not be selected (i.e., the target state is not a candidate state), the Viterbi decoding circuit 510 sets the accumulated metrics corresponding to the target branches, which in this instance are all branches of the target state (i.e., branches b0 and b1), to the default value mpreset (step S715). Then in step S760 the Viterbi decoding circuit 510 can select either the branch b0 or branch b1 as the branch of the target state S01 of stage k. Regardless of whether branch b0 or b1 is selected by the Viterbi decoding circuit 510 in step S760, the Viterbi algorithm in the end will not select the survivor path containing the state S01 (that is, both branches b0 and b1 of this state S01 can be regarded as having been excluded or eliminated) because the accumulated metrics of all the branches of the target state S01 have been set to the default value mpreset.
In the embodiment of
Since in the first to-be-decoded data E all bits corresponding to the error detection data are unknown (unless all bits of the auxiliary data A′ are known) and continuously distributed, and the Viterbi algorithm performs poorly when addressing continuous errors, the error rate of the intermediate error detection data {circumflex over (B)}dec is higher than that of the estimated data Â. In other words, the high accuracy of the estimated data  is beneficial for the decoding of the error detection data. Thus, the error detection data generation circuit 920 then performs an error detection operation on the estimated data  to generate the reference error detection data {circumflex over (B)}reb (step S1060). The function of the error detection data generation circuit 920 is the same as that of the error detection data generation circuit 310, and the detail of the error detection data generation circuit 920 is omitted for brevity. Since the reference error detection data {circumflex over (B)}reb is rebuilt based on the estimated data Â, the accuracy of the reference error detection data {circumflex over (B)}reb is in general higher than that of the intermediate error detection data {circumflex over (B)}dec. The reference error detection data {circumflex over (B)}reb and the intermediate error detection data {circumflex over (B)}dec have the same number of bits.
The comparison circuit 950 compares the reference error detection data {circumflex over (B)}reb with the intermediate error detection data {circumflex over (B)}dec and generates a control signal Ctrl. In one embodiment, the number of bits of the control signal Ctrl is the same as the number of bits of the reference error detection data {circumflex over (B)}reb and the number of bits of the intermediate error detection data {circumflex over (B)}dec. Accordingly, in determining the content of the control signal Ctrl, the comparison circuit 950 can set the bit whose value in the reference error detection data {circumflex over (B)}reb is the same as that in the intermediate error detection data {circumflex over (B)}dec as a known information bit (e.g., by setting the corresponding bit in the control signal Ctrl to a logic 1), and set the bit whose value in the reference error detection data {circumflex over (B)}reb is different from that in the intermediate error detection data {circumflex over (B)}dec as an unknown information bit (e.g., by setting the corresponding bit in the control signal Ctrl to a logic 0) (step S1070). In a case where the number of bits whose values in the reference error detection data {circumflex over (B)}reb are different from those in the intermediate error detection data {circumflex over (B)}dec is smaller than a certain threshold (e.g., the number of logic 0s in the control signal Ctrl is smaller than the threshold), which implies that errors caused by noises may have occurred to the Viterbi decoding circuit 340 in decoding the intermediate error detection data {circumflex over (B)}dec, the selection circuit 930 refers to the control signal Ctrl and the control signal Prek to generate the second to-be-decoded data E′ based on the intermediate error detection data {circumflex over (B)}dec (or the reference error detection data {circumflex over (B)}reb), the received data {circumflex over (Z)}, and the intermediate data Z′ (step S1080). In an alternative case where the number of bits whose values in the reference error detection data {circumflex over (B)}reb are different from those in the intermediate error detection data {circumflex over (B)}dec is not smaller than the threshold (e.g., the number of logic 0s in the control signal Ctrl is not smaller than the threshold), the selection circuit 930 refers only to the received data {circumflex over (Z)} and the intermediate data Z′ to generate the second to-be-decoded data E′.
In some embodiments, the control signal Prek, the control signal Ctrl, the intermediate data Z′, and the received data {circumflex over (Z)} have the same number of bits. A logic 1 in the control signal Prek indicates that the corresponding bit in the intermediate data Z′ is known, whereas a logic 0 indicates that the corresponding bit in the intermediate data Z′ is unknown. A logic 1 in the control signal Ctrl indicates that the corresponding bits in the intermediate error detection data {circumflex over (B)}dec and the reference error detection data {circumflex over (B)}reb are known, whereas a logic 0 indicates that the corresponding bits in the intermediate error detection data {circumflex over (B)}dec and the reference error detection data {circumflex over (B)}reb are unknown. Therefore, based on the contents of the control signal Prek and the control signal Ctrl (e.g., based on the result of the bitwise OR operation of the two), the selection circuit 930 can determine the bit values of a second to-be-decoded data E′ in the following manner: for the bits that correspond to the error detection data, the selection circuit 930 uses the bit whose value in the intermediate error detection data {circumflex over (B)}dec is the same as its value in the reference error detection data {circumflex over (B)}reb to replace the corresponding bit in the received data {circumflex over (Z)}; for the bits that do not correspond to the error detection data, the selection circuit 930 uses the bit that is known in the intermediate data Z′ to replace the corresponding bit in the received data {circumflex over (Z)}. As a result, the selection circuit 930 generates the second to-be-decoded data E′. In other words, if the control signal Ctrl and/or the control signal Prek indicates that a certain bit of the received data {circumflex over (Z)} has a corresponding predicted value, the selection circuit 930 replaces the value of that bit with the corresponding predicted value. If, on the other hand, the control signal Ctrl and/or the control signal Prek indicates that a certain bit of the received data {circumflex over (Z)} is unknown (i.e., not predicted), the selection circuit 930 selects the value in the received data {circumflex over (Z)} for that bit.
Finally, the Viterbi decoding circuit 940 decodes the second to-be-decoded data E′ to obtain the target data Ĉ′ (step S1090). The selection circuit 930 can be implemented with a multiplexer. Note that in other embodiments some of the steps in
Since the rebuilt reference error detection data {circumflex over (B)}reb has high accuracy, which enables the embodiment of
The aforementioned circuits (i.e., the functional blocks of
Since a person having ordinary skill in the art can appreciate the implementation detail and the modification thereto of the present method invention through the disclosure of the device invention, repeated and redundant description is thus omitted. Please note that there is no step sequence limitation for the method inventions as long as the execution of each step is applicable. Furthermore, the shape, size, and ratio of any element and the step sequence of any flow chart in the disclosed figures are exemplary for understanding, not for limiting the scope of this invention. In addition, although the foregoing embodiments are illustrated based on the NB-IoT, those skilled in the art can appropriately apply the present invention to other types of communication systems in accordance with the disclosure of the present invention.
The aforementioned descriptions represent merely the preferred embodiments of the present invention, without any intention to limit the scope of the present invention thereto. Various equivalent changes, alterations, or modifications based on the claims of the present invention are all consequently viewed as being embraced by the scope of the present invention.
| Number | Date | Country | Kind |
|---|---|---|---|
| 107124893 | Jul 2018 | TW | national |
