[in Japanese]

View full abstract

We first examined the applicability of optical wireless transmission systems to a Field Pick-up Unit (FPU) for TV relay. Assuming that the transmission was digital, the characteristics of light wave transmission through the atmosphere under various weather conditions and then investigated the code error characteristics of the transmission path. Using the results of these investigations, we developed an optical FPU and confirmed it's suitability for practical use by testing it on a variety of programs. We also made an experimental optical FPU for high-definition television and demonstrated the possibility.

View full abstract

A phase-diversity technique is a viable technique for transmitting wideband TV programs with closely-arranged optical signals in broadcasting stations. This paper describes DQPSK coherent detection using phase diversity. An excellent BER of 10^-11 was achieved with a received optical power of -27 dBm at 2.97 Gb/s. 2.97 Gb/s corresponds to the transmission rate of two channels of the uncompressed digitalized HDTV signals. Optical intensity modulation with a microwave subcarrier is a key technology that can achieve QPSK coherent detection more easily than the direct optical phase modulation technology proposed so far.

View full abstract

We have studied equalization techniques using pilot carriers for an OFDM system. The signal to interference ratios (SIRs) of the equalizer output signal over a time-invariant multipath channel (D/U =10 dB) are calculated for several interpolation schemes : interpolation using FIR filters with the sinc functional impulse response, constant interpolation, linear interpolation, and simplified linear interpolation (which can be achieved without multipliers). The results show that an SIR of 27 dB (input CNR=30 dB, delay time of interference signal=11 μs) is obtained for simplified linear interpolation which can reduce the hardware complexity compared to other interpolation schemes. Also, we have confirmed that the tap numbers of the FIR filter can be reduced without degradation of the SIR by using a window function. Furthermore, for the FIR filters and the simplified linear interpolations, the SIRs are calculated in case of the timing phase offset for the FFT window position. The results show that the degradation of the SIR is very little even when the offset value becomes up to 20 samples for both cases.

View full abstract

This paper describes a new equalizer for Orthogonal Frequency Division Multiplexing (OFDM) using scattered pilot symbols. The channel transfer function is modeled as a combination of several propagation paths whose complex attenuation factors and delay times are estimated so as to minimize the differences between the received pilot symbols and the pilot symbols generated by this model. The received data symbols are equalized by using this model. Computer simulations are used to compare the bit error rate characteristics of the new equalizer with those of the conventional equalizer that esitimates channel transfer functions by interpolation. Under severe channel conditions such as fading channels, the new equalizer gives better bit error rate characteristics. When the carrier-to-noise ratio is low, the convensional method is a little better. This paper also describes a modification of the new equalizer which improves the bit error rate characteristics without much increasing number of calculations.

View full abstract

Orthogonal Frequency Division Multiplexing (OFDM) is a prime candidate for use as a modulation method for digital terrestrial broadcasting, because it is robust against multipath interference and can be demodulated correctly under Single Frequency Network circumstances Mobile reception or HDTV broadcasting can be achieved by using DQPSK or 64 QAM modulation, respectively, as the carrier modulation method for OFDM When a receiver demodulates 64 QAM, pilot symbols are necessary as amplitude and phase references This paper describes a study on how transmission performance depends on symbol duration for mobile reception, and on the frequency, time intervals, and level of pilot symbols The results show that a 500 μs OFDM symbol duration can support mobile reception in the case of DQPSK OFDM It is suitable for the pilot symbols to have a frequency interval of 3, time interval of less than 5, and a data symbol level of 1 2 to 1 4 times that of the average level of data symbols

View full abstract

In Japan, the Digital Terrestrial Television Broadcasting (DTTB) system is expected to have mobile reception capability. It is well known that the effect of multipath interference in the fixed condition can be reduced by using the Orthogonal Frequency Division Multiplex (OFDM) method. However, the signal quality degrades seriously due to frequency selective fading and Doppler shifts of the mobile environment. This is because the orthogonality between the carriers has been lost due to this interference. In this paper, we discuss about multipath channel modeling for mobile reception with an OFDM. And we propose a symbol equalization method for coherent demodulation. We also show that the BER characteristics have been improved, and that mobile reception using the 16QAM OFDM system is possible.

View full abstract

This paper proposes a new frequency offset compensation scheme for OPFDM (Orthogonal Polarization and Frequency Division Multiplexing) signals OPFDM is a class of OFDM (Orthogonal Frequency Division Multiplexing) whose sub-channels are divided into even and odd sub-channels, and these sub-channels' signals are transmitted using different polarization in order to eliminate the performance degradation caused by frequency offset. In this scheme, the frequency offset can be continuously compensated for without a known training sequence, pilot signal or guard time symbols The proposed scheme can further compensate for the frequency offset by making use of the cyclostationarity of the OPFDM signals Computer simulation results confirm that the proposed scheme can efficiently compensate for the frequency offset, and improve the bit error rate performance

View full abstract

In a hybrid analog (AM VSB) /digital (QAM) sub carrier multiplexed (SCM) optical cable TV system, clipping induced impulsive distortion occurs when the SCM signal's amplitude drops below the laser diode (LD) threshold This distortion may induce bursts of bit errors in digital channels We have investigated the duration of clipping induced distortion and its effect on the BER performance of QAM signals As a result, we found that the frequency accuracy of the analog sub carriers is one of the main factors that characterize the clipping induced impulsive distortion which affects the performance of the digital channel The more accurately the frequencies of the analog sub carriers are set, the longer the duration of the clipping induced distortions becomes If the duration exceeds the ability of the forward error correction (FEC) system, the FEC system will not work well because of the error bursts This factor has not been considered in previous approaches

View full abstract

The BST-OFDM (Band Segmented Transmission-OFDM) method is considered for use in a transmission system in Japan, taking into account requirements for terrestrial digital broadcasting, such as providing HDTV services and multimedia services for mobile reception This method makes hierarchical transmission and partial reception possible This paper describes a transport signal which is compliant with MPEG-2 Systems and which is applicable to hierarchical transmission It proposes the following First, the introduction of a frame structure to both an OFDM signal and a transport signal, to assign TS packets to the proper transmission layer Second, the insertion of a number of null packets, to interface a transport signal between a multiplexer and modulator with a single constant clock Third, the introduction of an algorithm for specifying the allocation of the TS packets in a multiplex frame, to regenerate the correct transport stream on the receiver side easily

View full abstract

[in Japanese]

View full abstract

In this paper, we propose a method for the detection of actual moving objects in a moving image obtained by a moving camera. First, we construct a basic method for the detection of actual moving objects by utilizing motion signals of a camera. In this method, actual moving objects are detected based on the difference between the actual image flow observed and the predicted image flow computed by utilizing motion signals of a camera. The basic method has some requirements which are difficult to satisfy, such as exact computation of depth values in the surroundings. Therefore, we improve this method to perform the correct detection of actual moving objects without depth signals. In the improved method, the detection is performed based on whether the actual image flow is involved in a subregion of the image flow which is computed as that of stationary objects or not. In this method, the exact computaion of depth values in the surroundings and the computation of a speed component for the translational motion of a camera are not required.

View full abstract

This work shows that the application of a MgO and MgF₂ co-evaporated protecting layer to a combination of transmissive and reflective phosphor type color PDPs improves the luminance and the luminous efficiency and the optimum composition of the co-evaporated layers is 24% MgO and 76% MgF₂. This work also confirms that the sustaining voltage and life of reflective phosphor type color PDPs with the co-evaporated protecting layer are almost the same as those with a MgO single protecting layer. The sustaining voltage of combination of transmissive and reflective phosphor type color PDPs, however, is higher than that of reflective phosphor type color PDPs. It is probable that sustaining voltage can be decreased by thinning the thickness of the phosphors.

View full abstract

This paper discusses a neural network for a rotation invariant pattern recognition system. The proposed network can normalize a rotated random pattern into the unrotated standard pattern. The normalization is achieved by cascading a preceding rotation-angle extracting network and the succeedingnormalizing network. This basic structure is the same as the previously reported shifted-pattern normalizing network. The network is trained to normalize input random pattern direction into its center of gravity downwardsor in some other predetermined direction. The weight distribution viewed from the input layer in each hidden unit reveals a Fourier transform like in the radius and angle directions.

View full abstract

A display system requiring observers to have no special glasses is useful for 3-D images. The authors have researched real-time 4-view and 32-view holographic stereogram displays using a holographic screen. These simple display systems present stereoscopic 3-D images, but the quality of the images is poor because of the lack of usable pixels. This paper shows that this problem can be solved by using multiple LCD panels and describes the results of a fundamental study on a stereo display system using multiple panels. In addition, we propose a stereogram synthesizer using this newly developed 3-D display system.

View full abstract

Three experiments were conducted. In each experiment, a fixation point (FP) simultaneously appeared in the center of a monitor with the saccade target (ST), which was presented randomly either 4 deg or 8 deg to the right of the FP. Four subjects were instructed to fixate on the FP for one second (as judged by each subject) after the appearance of the ST and to then make a single saccade to the ST. The ST Diameter, ST luminance, and FP diameter were changed randomly in each experiment, independently. The results were as follows. (1) The larger or brighter ST significantly decreased the normalized saccadic peak velocity (NSPV) as well as the saccadic amplitude. (2) The larger FP significantly decreased the saccadic amplitude and the NSPV. Some recent physiological findings suggested that the strength of the fixation before saccade generation had produced these results.

View full abstract

Visual stimuli were presented on a monitor in complete darkness. The head of each subject was immobilized. The following two tasks were carried out randomly in different trials : the disappearance of the central fixation point (FP, 0 deg) preceded the appearance of the peripheral saccade target by 200 ms (gap task) and the FP remained after the appearance of the target (overlap task). The target was randomly presented 4 and 8 deg to the right or left of the FP (1 : 3=left : right). The diameter of the FP was randomly set at 0.2, 0.8, or 3.2 deg. The subject was asked to make a saccade to the target after the target appeared. Only the saccades elicited to the right targets were analyzed. The results showed that the normalized saccade peak velocity and amplitude for the 3.2 deg FP were less than those for the 0.2 and 0.8 deg FP. These effects may be caused by the suppression of the saccade because the larger FP would stimulate the fixation cells on the superior colliculus more strongly. However, the larger FP decreased the saccadic reaction time. We explained this contradicting phenomenon by modifying Fischer's hypothesis of the attention dynamics during the saccade.

View full abstract

It is possible to evaluate viewer's interest in visual information using pupillary changes caused by mental activity. This paper examines the possibility of estimating viewer's interest and enjoyment of T.V. programs by measuring dynamic pupillary changes, blinking, and subjective interest responses. The primary factors for pupillary changes are derived, for evaluating images of T.V. programs from questionnaires which are based on the SD method. The results showed, in dynamic evaluation, that there is a relationship between pupillary changes and subjective interest response. In evaluating images from T.V. programs, there is a significant difference in pupil size between the higher and the lower group for the factor score of 'Potency', and a difference in blinking between the higher and the lower group for the factor score of 'Activity'. Also there is a significant correlation between subjective interest response and all three factors.

View full abstract