The author describes the construction of a universal code for minimizing L.D. Davisson's minimax redundancy in a range where the true model and stochastic parameters are unknown. Minimax redundancy is defined as t...
详细信息
The author describes the construction of a universal code for minimizing L.D. Davisson's minimax redundancy in a range where the true model and stochastic parameters are unknown. Minimax redundancy is defined as the maximum difference between the expected per-symbol code length and the per-symbol source entropy in the source range. A universal coding scheme is formulated in terms of the weight function, i.e., a method is presented for determining a weight function which minimizes the minimax redundancy even when the true model is unknown. It is subsequently shown that the minimax redundancy achieved through the presented coding method is upper-bounded by the minimax redundancy of J. Rissanen's semi-predictive coding method.< >
This article addresses the problem of realizing the second textual substitution compressor in hardware and evaluate the combined effects of dictionary size and of various dictionary management mechanisms on the perfor...
详细信息
This article addresses the problem of realizing the second textual substitution compressor in hardware and evaluate the combined effects of dictionary size and of various dictionary management mechanisms on the performance of a Ziv-Lempel encoder, and then present a new dictionary manager which resolves the time-space trade-offs of its predecessors. The second data compression scheme repeatedly matches the input stream to words contained in a dictionary, and returns pointers to locations in the dictionary of the longest matches. Dictionary management scheme employs a trie data structure with uniquely fit tagged nodes whose tags satisfy a modified heap property. Dictionary access is done by using hashing technique and associative memory approaches. Performance of deletion heuristics, Ziv-Lempel compression on binary files is done FREEZE, FLUSH, SWAP are tested and the compression ratio is also plotted. This tagged decompressor architecture consists of three coupled parallel processes.
The authors describe a model for, and an implementation of, production-based synthesis of hardware description language (HDL) code in which the overall structure of the resultant machine is derived from a hierarchy of...
详细信息
The authors describe a model for, and an implementation of, production-based synthesis of hardware description language (HDL) code in which the overall structure of the resultant machine is derived from a hierarchy of sub-machine descriptions, each represented by a production. The production-based specification (PBS) consists of productions annotated with HDL action code, and forms the input to a design tool which outputs procedural HDL tailored for hardware synthesis. Due to the concise nature of this form of specification, the technique can save enormous labor in the construction of procedural specifications for these machines. Novel aspects of this research include the compilation of a PBS with HDL action clauses into synthesizable procedural HDL and the approach to specification of machine behavior in the event of exceptional conditions.< >
Efficient stop-and-wait, go-back-N and selective-repeat hybrid automatic-repeat-request (ARQ) protocols with adaptive forward error correction using punctured and repetition convolutional coding are proposed and analy...
详细信息
Efficient stop-and-wait, go-back-N and selective-repeat hybrid automatic-repeat-request (ARQ) protocols with adaptive forward error correction using punctured and repetition convolutional coding are proposed and analyzed. The basic idea is to vary the coding rate for error correction according to system parameters, such as the signal-to-noise ratio, the round trip delay and the buffer size at the receiver, so as to maximize the throughput efficiency. The authors refer to these schemes as adaptive coding rate (ACR) ARQ protocols. The performance of the proposed ACR ARQ protocols is evaluated for an additive white Gaussian noise channel.< >
A frequency-hopped spread-spectrum multiple-access (FH-SSMA) communication system with a rate-adaptive coding scheme is considered. The scheme uses a variable-rate Reed-Solomon code, with a single decoder. Two cases a...
详细信息
A frequency-hopped spread-spectrum multiple-access (FH-SSMA) communication system with a rate-adaptive coding scheme is considered. The scheme uses a variable-rate Reed-Solomon code, with a single decoder. Two cases are considered: N-ary frequency-shift keying (FSK) without erasure and N-ary FSK with perfect side information as to whether a channel symbol is hit or not. It is shown that FH-SSMA with this rate-adaptive coding scheme can serve a higher number of users than with fixed-rate coding for the above mentioned cases. This increase in the number of users is due to the higher throughput which results in less interference for the other users. It is shown that perfect side information increases the capacity of the system. The number of users can further increase if more time delay is allowed in the system.< >
One of the important issue in speech interpolation is that of speech freeze-out which becomes a very serious consideration with multihop networks. The paper discusses the implementation of two such techniques namely s...
详细信息
One of the important issue in speech interpolation is that of speech freeze-out which becomes a very serious consideration with multihop networks. The paper discusses the implementation of two such techniques namely speech storage and variable bitrate strategies with their independent and combined effects on the performance of multihop, DSI networks. The trade-off between the effective increased channel capacity and the degradation of the speech quality has also been pointed out. The basic architectural configuration of the system using multimicroprocessor system has been discussed and the potentialities of the method in accomodating newer versions of coding algorithms at desired speed of response have been included.
A high performance data communication system is designed using a 2.5 GHz radio with a data transmission rate of 6.312 Mbit/s and a 800 MHz radio with a data transmission rate of 4800 bit/s. To countermeasure both rand...
详细信息
A high performance data communication system is designed using a 2.5 GHz radio with a data transmission rate of 6.312 Mbit/s and a 800 MHz radio with a data transmission rate of 4800 bit/s. To countermeasure both random bit and burst errors, Reed-Solomon (RS) codes are considered for the forward and reverse channels. The system reliability is improved by using an adaptive forward error correction (FEC) scheme in the forward channel and a conventional FEC scheme in the reverse channel. As a result, the stop-and-wait protocol is used to simplify the circuitry and avoid large buffers, and yet the system requirement can still be met. The proposed coding and retransmission strategy seems suitable for use in a severe multipath fading.< >
Since the local characteristics of an image usually change from region to region, adaptive coding techniques are necessary to achieve high compression rates. An effective coding scheme is based on the principle of var...
详细信息
Since the local characteristics of an image usually change from region to region, adaptive coding techniques are necessary to achieve high compression rates. An effective coding scheme is based on the principle of variable resolution image representation, where compression is generally achieved by adaptive sampling or subsampling. Various adaptive sampling schemes that are effective for image compression are developed and evaluated. The main focus is the design of the nonuniform sampling patterns associated with the sample selection process. Two different design procedures, based on nonrecursive and recursive image decomposition, are investigated. Simulation results indicate that some of the sampling schemes proposed here attain higher compression rates than other reported coders of this nature.< >
Three commercially available data compression ASICs are compared: the Hewlett-Packard MYSTIC, the Infochip IC-1053, and the Stac 9703. All three are suitable for lossless compression of general data including text, ob...
详细信息
Three commercially available data compression ASICs are compared: the Hewlett-Packard MYSTIC, the Infochip IC-1053, and the Stac 9703. All three are suitable for lossless compression of general data including text, object code, databases, etc. All use a form of Ziv-Lempel adaptive coding and can be integrated into a standard microprocessor system. Compression performances are compared and found to be similar, in the general vicinity of 2 to 1.< >
暂无评论