SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Several methods for realizing erasure declaring Viterbi decoders for the (7,1/2) NASA code are discussed. Only bit oriented algorithms are considered. When such decoders are used in a concatenated system with a (255,223) Reed-Solomon decoder, improvements on the probability of word error of at most 0.1 dB were obtained.
This research project is intended to develop an efficient system for reliably transmitting image and video data over low bit rate noisy channels. The basic ideas behind the proposed approach are the following: employ statistical-based image modeling to facilitate pre- and post-processing and error detection, use spare redundancy that the source compression did not remove to add robustness, and implement coded modulation to improve bandwidth efficiency and noise rejection. Over the last six months, progress has been made on various aspects of the project. Through our studies of the integrated system, a list-based iterative Trellis decoder has been developed. The decoder accepts feedback from a post-processor which can detect channel errors in the reconstructed image. The error detection is based on the Huber Markov random field image model for the compressed image. The compression scheme used here is that of JPEG (Joint Photographic Experts Group). Experiments were performed and the results are quite encouraging. The principal ideas here are extendable to other compression techniques. In addition, research was also performed on unequal error protection channel coding, subband vector quantization as a means of source coding, and post processing for reducing coding artifacts. Our studies on unequal error protection (UEP) coding for image transmission focused on examining the properties of the UEP capabilities of convolutional codes. The investigation of subband vector quantization employed a wavelet transform with special emphasis on exploiting interband redundancy. The outcome of this investigation included the development of three algorithms for subband vector quantization. The reduction of transform coding artifacts was studied with the aid of a non-Gaussian Markov random field model. This results in improved image decompression. These studies are summarized and the technical papers included in the appendices.
Statistical encoding techniques enable the reduction of the number of bits required to encode a set of symbols, and are derived from their probabilities. Huffman encoding is an example of statistical encoding that has been used for error-free data compression. The degree of compression given by Huffman encoding in this application can be improved by the use of prediction methods. These replace the set of elevations by a set of corrections that have a more advantageous probability distribution. In particular, the method of Lagrange Multipliers for minimization of the mean square error has been applied to local geometrical predictors. Using this technique, an 8-point predictor achieved about a 7 percent improvement over an existing simple triangular predictor.
In a simple multiple-access communication system, two geographically separated users attempt to communicate binary data to two data sinks over a common channel called a multiple access channel. User one sends codewords from a block code C sub one, while user two sends codewords from a block code C sub two. The two users occupy the same frequency slot, transmit at the same time, and use the same type of modulation. Block codes which are uniquely decodable and capable of correcting errors are constructed for two multiple-access channel models. The first model is referred to as a noiseless multiple-access binary erasure channel. If the two transmitted bits from the two users are zeros, a zero is transmitted over the channel to the receiver; if the two transmitted bits are ones, a one is transmitted to the receiver; if the two bits are different, an erasure symbol is transmitted to the receiver. The second model is also a multiple-access binary erasure channel but with noise introduced.
Symmetry-breaking transitions are a well-understood phenomenon of closed quantum systems in quantum optics, condensed matter, and high energy physics. However, symmetry breaking in open systems is less thoroughly understood, in part due to the richer steady-state and symmetry structure that such systems possess. For the prototypical open system—a Lindbladian—a unitary symmetry can be imposed in a “weak” or a “strong” way. We characterize the possible $\mathbb{Z}_n$ symmetry-breaking transitions for both cases. In the case of $\mathbb{Z}_2$, a weak-symmetry-broken phase guarantees at most a classical bit steady-state structure, while a strong-symmetry-broken phase admits a partially protected steady-state qubit. Viewing photonic cat qubits through the lens of strong-symmetry breaking, we show how to dynamically recover the logical information after any gap-preserving strong-symmetric error; such recovery becomes perfect exponentially quickly in the number of photons. Further, our study forges a connection between driven-dissipative phase transitions and error correction.
The primary accomplishments have been in the analysis and simulation of receivers and bit synchronizers. It has been discovered that tracking rate effects play, a rather fundamental role in both receiver and synchronizer performance, but that data relating to recorder time-base-error, for the proper characterization of this phenomenon, is in rather short supply. It is possible to obtain operationally useful tape recorder time-base-error data from high signal-to-noise ratio tapes using synchronizers with relatively wideband tracking loops. Low signal-to-noise ratio tapes examined in the same way would not be synchronizable. Additional areas of interest covered are receiver false lock, cycle slipping, and other unusual phenomena, which have been described to some extent in this and earlier reports and simulated during the study.
One- and two-dimensional adaptive delta modulator (ADM) algorithms are discussed and compared. Results are shown for bit rates of two bits/pixel, one bit/pixel and 0.5 bits/pixel. Pictures showing the difference between the encoded-decoded pictures and the original pictures are presented. The effect of channel errors on the reconstructed picture is illustrated. A two-dimensional ADM using interframe encoding is also presented. This system operates at the rate of two bits/pixel and produces excellent quality pictures when there is little motion. The effect of large amounts of motion on the reconstructed picture is described.
The cognitive behavior of pilots was examined in an experiment involving mock ups of an eight display electronic attitude direction indicator for an airborne integrated display. Displays were presented in digital, analog digital, and analog format to experienced pilots. Two tests were run, one involving the speed of memorization in a single exposure and the other comprising two five second exposures spaced 30 sec apart. Errors increased with the speed of memorization. Generally, the analog information was assimilated faster than the digital data, with regard to the response speed. Information processing was quantified as 25 bits for the first five second exposure and 15 bits during the second.
The optimum decoding of component codes in block coded modulation (BCM) schemes requires the use of the log-likelihood ratio (LLR) as the signal metric. An approximation to the LLR for the least reliable bit (LRB) in an 8-PSK modulation based on planar equations with fixed point arithmetic is developed that is both accurate and easily realizable for practical BCM schemes. Through an error power analysis and an example simulation it is shown that the approximation results in 0.06 dB in degradation over the exact expression at an E(sub s)/N(sub o) of 10 dB. It is also shown that the approximation can be realized in combinatorial logic using roughly 7300 transistors. This compares favorably to a look up table approach in typical systems.
The design of a reliable satellite communication link involving the data transfer from a small, low-orbit satellite to a ground station, but through a geostationary satellite, was examined. In such a scenario, the received signal power to noise density ratio increases as the transmitting low-orbit satellite comes into view, and then decreases as it then departs, resulting in a short-duration, time-varying communication link. The optimal values of the small satellite antenna beamwidth, signaling rate, modulation scheme and the theoretical link throughput (in bits per day) have been determined. The goal of this thesis is to choose a practical coding scheme which maximizes the daily link throughput while satisfying a prescribed probability of error requirement. We examine the throughput of both fixed rate and variable rate concatenated forward error correction (FEC) coding schemes for the additive white Gaussian noise (AWGN) channel, and then examine the effect of radio frequency interference (RFI) on the best coding scheme among them. Interleaving is used to mitigate degradation due to RFI. It was found that the variable rate concatenated coding scheme could achieve 74 percent of the theoretical throughput, equivalent to 1.11 Gbits/day based on the cutoff rate R(sub 0). For comparison, 87 percent is achievable for AWGN-only case.
An architecture, and a method of utilizing the architecture, have been proposed to enable error-free operation of a data bus that includes, and is connected to, commercial off-the-shelf (COTS) circuits and components that are inherently susceptible to singleevent upsets [SEUs (bit flips caused by impinging high-energy particles and photons)]. The architecture and method are applicable, more specifically, to data-bus circuitry based on the Institute for Electrical and Electronics Engineers (IEEE) 1394b standard for a high-speed serial bus.
We investigated the single-event effect (SEE) susceptibility of the Micron 16 nm NAND flash, and found that the single-event upset (SEU) cross section varied inversely with cumulative fluence. We attribute the effect to the variable upset sensitivities of the memory cells. Furthermore, the effect impacts only single cell upsets in general. The rate of multiple-bit upsets remained relatively constant with fluence. The current test standards and procedures assume that SEU follow a Poisson process and do not take into account the variability in the error rate with fluence. Therefore, traditional SEE testing techniques may underestimate the on-orbit event rate for a device with variable upset sensitivity.
Video transmission will become an important part of future multimedia communication because of dramatically increasing user demand for video, and rapid evolution of coding algorithm and VLSI technology. Video transmission will be part of the broadband-integrated services digital network (B-ISDN). Asynchronous transfer mode (ATM) is a viable candidate for implementation of B-ISDN due to its inherent flexibility, service independency, and high performance. According to the characteristics of ATM, the information has to be coded into discrete cells which travel independently in the packet switching network. A practical realization of an ATM video codec called Mixture Block Coding with Progressive Transmission (MBCPT) is presented. This variable bit rate coding algorithm shows how a constant quality performance can be obtained according to user demand. Interactions between codec and network are emphasized including packetization, service synchronization, flow control, and error recovery. Finally, some simulation results based on MBCPT coding with error recovery are presented.
Plasma measurements in space are becoming increasingly faster, higher resolution, and distributed over multiple instruments. As raw data generation rates can exceed available data transfer bandwidth, data compression is becoming a critical design component. Data compression has been a staple of imaging instruments for years, but only recently have plasma measurement designers become interested in high performance data compression. Missions will often use a simple lossless compression technique yielding compression ratios of approximately 2:1, however future missions may require compression ratios upwards of 10:1. This study aims to explore how a Discrete Wavelet Transform combined with a Bit Plane Encoder (DWT/BPE), implemented via a CCSDS standard, can be used effectively to compress count information common to plasma measurements to high compression ratios while maintaining little or no compression error. The compression ASIC used for the Fast Plasma Investigation (FPI) on board the Magnetospheric Multiscale mission (MMS) is used for this study. Plasma count data from multiple sources is examined: resampled data from previous missions, randomly generated data from distribution functions, and simulations of expected regimes. These are run through the compression routines with various parameters to yield the greatest possible compression ratio while maintaining little or no error, the latter indicates that fully lossless compression is obtained. Finally, recommendations are made for future missions as to what can be achieved when compressing plasma count data and how best to do so.
The reliability of many industrial processes depends on the sensor system. However, these sensors can be affected by noise, perturbations and failures. Hence, sensor monitoring and diagnosis are fundamental to guarantee the quality of an industrial process. Nowadays, artificial neural networks (ANN) are widely used in sensor signal processing and diagnosis. However, those ANNs usually require many artificial neurons, being difficult to implement in software and hardware due to their high computational costs. This paper presents an optimized implementation of artificial neurons in ANNs for sensor data analysis using a Genetic Algorithm (GA). The objective of GA is to find an adequate segmentation to reduce the activation function approximation error. One of the advantages of the proposed approach is that the cost function used in GA considers the effect of factors such as the ANN architecture or the number of bits used in arithmetic operations. The proposed ANN implementation technique aims to get the best possible approximation for a specific ANN architecture, making easier its implementation in software and hardware. Simulation and experimental results using FPGA (Field Programmable Gate Array) prove the advantages of the proposed approach for implementing sensor data analysis systems based on ANNs.
Pegasus-III is an ultralow aspect ratio spherical tokamak providing a dedicated US experiment for comparative solenoid-free startup studies. A new magnetic diagnostic suite for equilibrium and low frequency (<200 kHz) magnetohydrodynamic mode analysis has been installed. These new diagnostics address the significant challenges of measuring magnetic field in a high noise environment with the majority constrained to fit in an 8 mm diagnostic gap on the high field side. Electrostatic switching noise generated by the 16 independent current feedback-controlled power supplies produces dV cm /dt ~ 1 kV/μs and volt level common mode noise on the magnetics. Immunity to this switching noise is accomplished through differential signal runs and signal processing, along with end-to-end electromagnetic interference shielding. The magnetic measurements are simultaneously digitized at 1 MHz and conditioned by precision 8 pole Butterworth filters with a corner frequency of 200 kHz to prevent aliasing down to the 16-bit level over the full passband. Ex-vessel calibrations of the B p coils were completed with a typical uncertainty of <0.5%. Stray toroidal field pickup from coil misalignment or positioning errors is corrected using a physics-based model. Comparisons of the corrected measurements to modeling agree to within 1.3% on average. This is within the 1.5% measurement uncertainty that a sensitivity analysis determined is needed for accurate fast boundary and equilibrium reconstruction.
The D Machine is described as a small user microprogrammable computer designed to be a versatile building block for such diverse functions as: disk file controllers, I/O controllers, and emulators. TRANSLANG is an ALGOL-like language, which allows D Machine users to write microprograms in an English-like format as opposed to creating binary bit pattern maps. The TRANSLANG translator parses TRANSLANG programs into D Machine microinstruction bit patterns which can be executed on the D Machine simulator. In addition to simulation and translation, the two programs also offer several debugging tools, such as: a full set of diagnostic error messages, register dumps, simulated memory dumps, traces on instructions and groups of instructions, and breakpoints.
A maximum aposteriori probability (MAP) approach to joint source/channel coder design is presented in this paper. This method attempts to explore a technique for designing joint source/channel codes, rather than ways of distributing bits between source coders and channel coders. For a nonideal source coder, MAP arguments are used to design a decoder which takes advantage of redundancy in the source coder output to perform error correction. Once the decoder is obtained, it is analyzed with the purpose of obtaining 'desirable properties' of the channel input sequence for improving overall system performance. Finally, an encoder design which incorporates these properties is proposed.