Abstract: A system for locating an agricultural vehicle in a field using two position sensors. A first position sensor is a GPS sensor using satellite signals to locate the vehicle. The second position sensor is a crop edge sensor for locating the vehicle relative to a crop being harvested. A controller receives the first and second position signals and calculates the position of the vehicle based on these two position signals. Both position signals contain signal quality information that is evaluated by the controller and weights the position calculation accordingly. The system can be used in conjunction with a steering controller to automatically guide the vehicle through a field.

Abstract: An automatic engine stop/restart-type vehicle and control method determine whether a vehicle is in a non-running range. The vehicle comprises: an engine; a transmission, an oil pump and a controller. The transmission has a coupling element that transmits an output torque of the engine. The oil pump supplies a hydraulic pressure to the transmission. The controller controls an output from the engine; controls the hydraulic pressure supplied to the transmission; and determines whether the engine is in an automatic stop state when a manually set mode of the transmission is shifted from a non-running range to a running range. A hydraulic pressure compensation procedure is performed to conform friction coupling characteristics of the coupling element and input characteristics of the transmission with each other if the controller determines that the engine is in the automatic stop state.

Abstract: A control method and system according to the present invention for a displacement on demand engine estimates cylinder air charge and/or predicts cylinder air charge for future cylinder interrupts. A model is provided that estimates cylinder air charge and/or predicts cylinder air charge for future cylinder interrupts. The model includes a history vector of inputs and states. The history vector of inputs and states are updated when a cylinder firing interrupt occurs. An operating mode of the engine is determined. Based on the operating mode, model parameters and model inputs are selected. The cylinder air charge is estimated and predicted for future cylinder interrupts.

Abstract: A fuel injector (5) injects fuel to generate an air-fuel mixture that is burnt in the engine (1), and a sensor (16) detects an air-fuel ratio of the air-fuel mixture from an exhaust gas composition of the engine (1). A controller (11) calculates a dead time representing a lag between the air-fuel ratio variation of the air-fuel mixture and the air-fuel ratio variation detected by the sensor (16), and calculates an estimated state quantity by applying a dead time compensation according to Smith method and a disturbance compensation to the air-fuel ratio detected by the sensor (16). By determining an air-fuel ratio feedback correction amount based on the estimated state quantity applying a sliding mode control process, the response and stability of air-fuel ratio control is improved.

Abstract: An exhaust emission control system for an internal combustion engine having an exhaust system is disclosed. The control system may include an exhaust gas purifying device provided in the exhaust system and an oxygen concentration sensor provided downstream of the exhaust gas purifying device. The exhaust gas purifying device may include at least an oxygen storing ability or a nitrogen oxide storing ability. An air-fuel ratio of an air-fuel mixture supplied to the engine may be enriched with respect to the stoichiometric air-fuel ratio, so as to reduce the oxygen or nitrogen oxides stored in the exhaust gas purifying device. A predicted value of the output from the oxygen concentration sensor may be calculated using a predictor based on the fuzzy logic reasoning. The completion of the reduction of the oxygen or nitrogen oxides stored in said exhaust gas purifying device may be determined according to the predicted value.

Abstract: A microprocessor based interengageable engine generator set control device that provides bi-directional communication and control via a common J1939 protocol of electronic engine control units. Such engine control units provide critical engine information and control using manufacturers proprietarycodes that are readable by the interface engine controller. Engine commands are issued by programmable software and operational input in response to information received and analyzed thereby.

Abstract: A navigation apparatus wherein “latitude” and “longitude” are used as coordinate information on a set location.

Abstract: A method for operating a navigation device comprising a central processing unit (100) and a memory medium (130) with a digitized map is disclosed. The method comprises the following steps: inputting a route with start and target coordinates that can be selected on the map; receiving traffic jam data via an external source (140); displaying a calculated route and the respective traffic jam data; and editing said traffic jam data. In the case of acceptance of a traffic jam a possible bypass of the corresponding route section is calculated and displayed, and in the case of non-acceptance the calculated route section is maintained.

Abstract: A route guiding method for an in-vehicle navigation device. To guide a vehicle to a destination by an optimum route, it is determined whether the vehicle is on track within a predetermined intersection range based on route guidance data received from a traffic information center via a mobile communication network. If the vehicle is on track within the intersection range, it is determined whether the vehicle is in a free run state according to predetermined free run conditions to prevent a preliminary route guidance message that is confusing to a driver in the free run state. According to the determination result, the preliminary route guidance message is selectively provided.

Abstract: Global positioning systems (GPSs) estimate positions of vehicles based on signals from earth-orbiting satellite transmitters. For accuracy and reliability reasons, these systems have traditionally not be used for critical phases of aircraft navigation and guidance, such as aircraft landings. However, recent years have seen the development of ground-augmented GPS systems for use in automatic landing systems. These augmented systems rely on broadcast correction data to correct positions estimates, or solutions, and thus provide more accurate position solutions. Unfortunately, the conventional methods of measuring accuracy in these augmented systems cannot adequately cope with loss of correction data or satellite signals and thus lead to more aborted landings than acceptable. Accordingly, the inventor devised a ground-augmented GPS system that incorporates a better method for determining the accuracy of its position solution.

Abstract: This invention is drawn to an autonomous navigation system using Global Positioning System (GPS) and magnetometers for low Earth orbit satellites. As a magnetometer is reliable and always provides information on spacecraft attitude, rate, and orbit, the magnetometer-GPS configuration solves GPS initialization problem, decreasing the convergence time for navigation estimate and improving the overall accuracy. Eventually the magnetometer-GPS configuration enables the system to avoid costly and inherently less reliable gyro for rate estimation. Being autonomous, this invention would provide for black-box spacecraft navigation, producing attitude, orbit, and rate estimates without any ground input with high accuracy and reliability.

Abstract: A graphical user interface of a well logging display program for editing data records included in well logging data includes multiple windows and controls for user interaction. The well logging data includes measurements of logging variables collected at successive depths to define the data records. The multiple windows include a first window for displaying depth versus time measurements, a second window for defining a logging variable, a third window for selecting and displaying a first set of instances of the logging variable, and a fourth window for selecting and displaying a second set of instances of the logging variable. The multiple controls include a bar line for selecting a splice point, and an interface button for instructing the program to edit the data records. The edited data records include measurements corresponding to the first and second set of instances being spliced in accordance with a predefined function.

Abstract: A method for determining the anisotropic resistivity properties of a subterranean formation traversed by a borehole utilizing a multi-component induction logging tool. The method utilizes the measured phase and attenuation signals induced by eddy currents in the formation to determine the azimuthal angle and to create a second set of signals based on the measured signals and the azimuthal angle. The anisotropic resistive properties, as well as the dip angle, are then simultaneously derived from the second set of signals by means of minimizing error functions within an inversion model.

Abstract: A method for generating P-S or S—S seismic data from P—P AVO data in P—P travel time for a geographic area includes obtaining a plurality of amplitude and angle of incidence pairs of P—P AVO data. The plurality of amplitude and angle of incidence pairs of P—P AVO data are inverted. Lastly, estimates of the rock property contrasts are produced from the inverted plurality of amplitude and angle of incidence pairs of P—P AVO data, the rock property contrasts including at least one of &Dgr;Vp/Vp, &Dgr;Vs/Vs and &Dgr;&rgr;/&rgr;.

Abstract: In order to align DNA sequence data traces, an experimental data trace representing the positions of a first species of base within a target polynucleotide and a reference data trace representing the positions of a second species of base (which may be the same as or different from the first species) within a reference polynucleotide are obtained by separating appropriate sequencing fragments generated from the target and reference polynucleotides on an electrophoresis gel. For each reference data trace, a plurality of peaks corresponding to fragments having a size in the range of 40 to 1200 bases are selected. A base number is assigned to each of the selected peaks in the reference data trace, and a numerical “peak file” is created with information about the peak number and migration time (or distance).

Abstract: A method of reducing loss in an interior permanent magnet drive system is provided. A current input DC link power value is calculated. The current input DC link power value is then compared with a previous input DC link power value. A change in input DC link power value is determined from this comparison. A flux decrement value is calculated. The flux decrement value is based on the change in input DC link power value. A stator current value is generated. The stator current value is based on the flux decrement value. Finally, the stator flux current is reduced, based on the stator current value.

Abstract: The invention describes a method, system, device, and computer-readable medium having computer-executable instructions for classifying faults on an electrical power line. In particular, the invention permits the classification of crossover faults using a local measurement technique. The inventive method includes providing a first electrical power transmission line and a second electrical power transmission line, and monitoring the first electrical power transmission line to identify a crossover fault occurring between the first and the second electrical power transmission lines.

Abstract: A method and implementation disclosed for detecting interference. A state machine controller is provided for establishing an interference detection cycle including a power sample period and a periodic sampling interval. A receiver component, responsive to the state machine controller, performs an energy measurement at a predetermined wireless band during the power sample period. A threshold comparator determines whether the energy measurement exceeds a predetermined threshold. A processing implementation processes the energy measurement to determine whether it corresponds to interference on the predetermined wireless band, if the measurement exceeds the predetermined threshold. A deactivating implementation is used to instruct the state machine controller to await the next power sample period, if the energy measurement does not exceed the predetermined threshold.

Abstract: Battery backed memory for use in an industrial controller allows volatile memory to be retained in instances of line power loss. A method and apparatus are disclosed for providing assurance that sufficient battery voltage was present during the entire power down period to maintain data integrity.

Abstract: Sampling techniques and circuits for a waveform measuring instrument. The sampling techniques and circuits process a series of digital signal samples through a set of sample extractors and subsequently process the extracted digital signal samples. The set of sample extractors include a uniform decimator, a low frequency dither decimator, and a digital peak detector. The uniform decimator extracts a uniform decimated sample value for each decimated sample interval in a series of decimated sample intervals. The low frequency dither decimator extracts a low frequency dither decimated (random) sample value for each decimated sample interval in a series of decimated sample intervals. The digital peak detector extracts a maximum sample value and a minimum sample value for each decimated sample interval in the series of decimated sample intervals.

Abstract: A spectrum analyzer that may be implemented by a simple microcontroller that does not have a hardware multiply function is disclosed. The spectrum analyzer of the present invention utilizes at least five frequency bins. The input signal is sampled at four times the bin frequency. The input signal is sampled at twice the Nyquist rate, which results in symmetries in the sin(wn) and cos(wn) functions. These symmetries allow the in-phase and quadrature components of the input signal to be calculated by add, ignore or subtract operations instead of the more complex multiplication and integration operations. Accordingly, the energy for each bin may be calculated with a minimum number of multiply operations. Because the number of multiply operations have been significantly reduced, these multiply operations may be performed by software instead of hardware. As a result, the spectrum analyzer may be implemented with a simple processor that does not have a hardware multiply.

Abstract: A controllable pulse generator apparatus and method are provided according to the invention. The pulse generator generates an adjustable, high current, high voltage pulse train output that may be transmitted to an external device under operation. The pulse generator includes a controller device and a pulse generator device. The controller device accepts one or more pulse train requests, may accept one or more external signals, and may accept feedback from the device under operation. The controller device generates one or more pulse train signals in response to the pulse train requests, the external signals, and the feedback. The controller device transmits the one or more pulse train signals to the pulse generator device. The pulse generator device receives the one or more pulse train signals and in response generates a pulse train output of a predetermined current level and of a predetermined voltage level according to the one or more pulse train signals.

Abstract: Once an eye diagram measurement is begun and there is an eye diagram displayed, different on-screen measurement tools may be used singly, or in combination. Each measurement involves indicating with cursors and line segments regions of the eye diagram that are of interest, and a parameter or parameters associated with each measurement tool in use is reported in a (usually) separate area of the display. An Eye Limits measurement allows the specification of a point within an eye diagram, whereupon it finds and reports the eye diagram coordinates first encountered along horizontal and vertical lines extended from the selected point (i.e., “eye opening” size). The coordinates of the point itself are also reported. A Four Point Box measurement allows the construction on the eye diagram of a rectangle having sides parallel to the coordinate axes of the eye diagram.

Abstract: The objective of the present invention is to achieve a measured data synchronizing system and a measured data synchronizing method which can determine the measured data whose synchronization is secured among measuring units without being restricted by the number of measuring units in the measuring part. The present invention is characterized by preparing a communication line, a plurality of measuring units which receive a reference time from the above communication line as an input and output tuple data in which at least the above reference time and the measured data are contained and arranged as a tuple, and a data processing unit which outputs the reference time to the above communication line, receives tuple-data from each of the above plurality of measuring units as inputs and secures synchronization of the measured data among the measuring units based on the reference time of the tuple-data.

Abstract: A method and system for continuously and automatically calibrating the settings of an electronic compass, as needed, while a vehicle moves from place to place is shown. The calibration technique of the present invention measures the ocal magnetic field while the vehicle moves from place to place and determines points on a reference Cartesian coordinate system. Each such point defines the endpoint for a vector corresponding with the measured field intensity and orientation of the local magnetic field. Any three such points are used to calculate the center of a circle that extends through all of those points. A vector that begins at the origin of the reference Cartesian coordinate system and terminates at the calculated center of the circle is thereafter subtracted from subsequent field measurement vectors that correspond to the field intensity and direction of the local magnetic field.

Abstract: The present invention is directed to a technique for performing calibration of an automatic sampler device. According to an aspect, the automatic sampler device includes a cell with a sample platform and a reference platform; a sample arm; a sample tray, and a platen. The sample tray includes wells into which pans are inserted. The platen may include conductive and/or reflective areas for calibration. The sample arm has an electronic sensor and an optical sensor. The electrical sensor and the optical sensor are used to calibrate the positions of one or more of: the sample platform, the reference platform, and a well. According to another aspect, autocalibration is optimized by adjusting autocalibration results with a set of stored offset coefficients. The offset coefficients are generated by performing a manual calibration. The difference between the results of the manual calibration and an autocalibration are stored as offset coefficients.

Abstract: Apparatus for testing functionality of a printed circuit board used to control operation of a printing press, including a plurality of input switches for providing to the circuit board test input signals simulating function control signals from a printing press control input, the function control signals controlling a plurality of functions of the printing press; a plurality of test loads simulating printing press loads for receiving from the circuit board output drive signals developed in response to the test input signals; and a processor for controlling operation of the input switches and for monitoring responses of the test loads in response to the output drive signals, whereby proper functionality of the printed circuit board in response to each of the plurality of function control signals is analyzed under simulated conditions of an actual operating printing press.

Abstract: A stream of pulses related to rotation rate drives an interrupt to store first timer values of a first timer and to set a second timer to run from zero. At predetermined fixed time intervals, second timer values of the second timer are stored. The stored first timer values are sampled to determine a first rate related to the input stream. The stored second timer values are sampled to determine a second rate. The first and second rates are selectively used as a final rotation rate related to the input stream.

Abstract: A position measuring device that includes a scanning device and a measurement representation that is scanned by the scanning device. The measurement representation includes a code of a succession of code elements, which form at least a first codeword and a second codeword that define an absolute position, the first codeword and the second codeword are arranged interlaced with each other in that code elements associated with the second codeword are inserted into a succession of code elements associated with the first codeword. The code elements associated with the first codeword have inverse properties with respect to the code elements associated with the second code codeword.

Abstract: A method for evaluating the performance of an instrument, particularly a hematology analyzer, by comparing its operating data with data compiled from a group of like instruments is provided. The method contemplates collecting historical data which is deposited in a central database where it may be retrieved by any participating laboratory. The participating laboratory performs an analysis and then accesses the database to retrieve data obtained from comparable instruments performing like analyses. The comparative analysis is then conducted from which a determination may be made with regard to a particular instrument.

Abstract: Disclosed is a method and mechanism for sizing a hardware system for a software workload. An aspect is directed to modeling a workload into a set of generic system activities, in which the generic system activities are not directly tied to a specific hardware platform. Another aspect is directed to profiling hardware systems or hardware components into the generic system activities. Suitable hardware systems or components are selected by analyzing the workload and hardware profiles in terms of the generic system activities.

Abstract: A simplified method of matching a golfer to a golf club and a golf ball by measuring the golfer's clubhead speed and comparing that measured value to recorded sets of data which correlates a few key variables that can accurately match the golfer with the most suitable golf club and golf ball designed to achieve optimum driving performance.

Abstract: A device for easily calculating a value remaining in an electrical apparatus and decreasing with use includes a sensor for detecting a state of use of the electrical apparatus, and issuing an electrical signal corresponding to the detected state, a determining portion for determining a history of use of the electrical apparatus based on the electrical signal sent from the sensor, a communication portion for receiving a standard residual value of a member of the electrical apparatus, a storing portion connected to the communication portion for storing the received residual value, a calculating portion connected to the history determining portion and the storing portion for calculating the value remaining in the electrical apparatus based on the detected history and the stored residual value, and a monitor connected to the calculating portion for displaying the residual value calculated by the calculating portion.

Abstract: A sequence of events detection system for use in a process control system uses sequence of events detection cards to detect and store indications of events and the times at which these events take place within the process control network. The sequence of events reporting system includes a stable master time source that is used to periodically time synchronize secondary clocks within each of the nodes of the process control system. Free running counters are located within each of the sequence of events detection cards at each node and these free running counters are used to mark each event with a counter value when the sequence of events detection card first detects the event.

Abstract: An appliance maintenance apparatus for contactlessly detecting the operation states of appliances and operating the appliances in response to the operation states of the appliances, and an appliance remote maintenance system which comprises a plurality of appliance maintenance apparatuses, and a remote diagnosis control apparatus which is communicatable with the appliance maintenance apparatus through a public network to ensure the remote diagnosis control of the appliance maintenance apparatus from a remote place.

Abstract: The system and method of the invention provide for remotely monitoring the operation of at least one turbine, the turbine being disposed in an operating location. The method comprises inputting input data from the turbine; and processing the input data to generate verified performance data, the processing including determining validity of the input data and performing at least one calculation on the input data, which is valid, to generate the verified performance data. The method further includes generating a collection of stored data in a database based on at least one of the input data and the verified performance data.

Abstract: A CPU of an effector calculates envelopes of an impulse response waveform from sample data of impulse response waveform data supplied from a microphone via an A/D converter circuit. Next, CPU detects a section during which a slope of the detected envelopes takes a value near “0” during a predetermined time or longer. CPU calculates an average value of sample data during the detected section to correct a DC offset. Thereafter, CPU acquires a maximum value of absolute values among the sample data during the detected section, and determines this maximum value as a background noise component value. The background noise components are removed from each sample data constituting the impulse response waveform data.

Abstract: A strength distribution diagram 33 of selected building structural components, which are selected from a virtual three-dimensional model of building structural components of a building structure created by a CAD program, is constructed on a plane. Stress values of the building structural components arranged in the strength distribution diagram 33 are computed based on stress computational data of the building structural components consistent with attributes of the building structural components. A plurality of strength level display modes is provided for each building structural component, and strength of each building structural component is indicated with a predetermined color based on the computed stress value.

Abstract: The algorithm used for training a state machine has been disclosed in another application. The objective of this application is to reduce this theoretical framework to a more concrete structure. The structure disclosed involves the use of nodes to perform the function of state equations and calculating the value of particular state variables. After the nodes where classified as either Lead-Type Nodes or Non-Lead-Type Nodes, this classification was used to define a structure of nodes to build a State Machine Block. This State Machine Block model also restricts the location of system inputs and system outputs. The internal structure of a node is discussed with the components being a Function Block and a Complex Impedance Network. The Function Block is typically a multivariable power series and the Complex Impedance Network is a linear circuit of resistors and capacitors. Such a Complex Impedance Network is referred to as an Electrical Component Model.

Abstract: A method of integrating computer visualization for the design of a vehicle includes the steps of determining a low-level geometric model of the vehicle and determining a modifiable parameter to modify the model of the vehicle. The method also includes the steps of morphing the model of the vehicle into a morphed model of the vehicle including the modifiable parameter using a computer visualization, and analyzing the morphed model of the vehicle using a computer aided engineering (CAE) analysis. The method further includes the steps of determining if the CAE analysis of the morphed model of the vehicle meets a predetermined criteria and using the morphed vehicle model in the design of the vehicle.

Abstract: Systems and corresponding methods that facilitate the automatic organization and collection of electronic information accessed or communicated by a computer user are provided. One embodiment of the system automatically monitors the type and content of electronic files, such as web pages, emails, and other sources for content that are accessed, received, or transmitted by a user. The system then analyzes this information and determines keywords or categories of information included in each content source. The content sources are then classified and referenced in one or more databases for future rapid access.

Abstract: An automated natural language translation system takes source natural language text (preferably in Japanese) and translates them into a target natural language (preferably English). The system also allows an operator to re-translate automatically selected portions of the source text. The system includes an improvement directed to transforming kanas in the source text into alphabetic letters of the target language which allows the presence of a word or phrase boundary to be recognized in the middle of a kana. The system also includes an improvement involving performing concurrently on the source text both a morphological analysis and a syntactic analysis.

Abstract: Voice recording is initiated immediately upon pressing a record button on a digital computing device such as a palm computer. To enable an immediate response to the button being pressed without a delay that would be normally be incurred to load a user interface and a recording application, a stub program is continuously executed on the device, even when a main user interface program is not running. When the record button is pressed, the stub program immediately begins recording a digital signal corresponding to the sound waves incident on a microphone of the device in a compressed format. If the record button is depressed for more than a second, the recorded digital signal is stored in a file so that it can subsequently be played back, and if the record button was pressed for less than one second, the file is not retained. Sound recording can be initiated in an “eyes free” mode, e.g., while using the device under low ambient light conditions, or while engaged in another activity.

Abstract: Described herein is a system that enables service provider's to integrate speech functionality into their applications. A service provider maintains a set of application servers. To provide a particular speech service to a client of the application server, the application server causes the client to request the speech service from another set of servers. This set of servers is responsible for providing this speech service as well as others. Such speech services include recording digital speech data at the client, and storing the recordings. Later, the application servers may retrieve the recordings, and even more, retrieve data derived from the recordings, such as data generated through speech recognition processes.

Abstract: A speech coding system includes an adaptive codebook containing excitation vector data associated with corresponding adaptive codebook indices (e.g., pitch lags). Different excitation vectors in the adaptive codebook have distinct corresponding resolution levels. The resolution levels include a first resolution range of continuously variable or finely variable resolution levels. A gain adjuster scales a selected excitation vector data or preferential excitation vector data from the adaptive codebook. A synthesis filter synthesizes a synthesized speech signal in response to an input of the scaled excitation vector data. The speech coding system may be applied to an encoder, a decoder, or both.

Abstract: A method and apparatus for performing automatic speech recognition (ASR) in a distributed ASR system for use over a wireless channel takes advantage of probabilistic information concerning the likelihood that a given, portion of the data has been accurately decoded to a particular value. The probability of error in each feature in a transmitted feature set is employed to improve speech recognition performance under adverse channel conditions. Bit error probabilities for each of the bits which are used to encode a given ASR feature are used to compute the confidence level that the system may have in the decoded value of that feature. Features that have been corrupted with high probability are advantageously either not used or are weighted less in the acoustic distance computation performed by the speech recognizer.

Abstract: A method of proofreading and correcting dictated text contained in an electronic document can include selecting proofreading criteria for identifying textual errors contained in the electronic document; playing back each word contained in the electronic document; and, marking as a textual error each played back word in nonconformity with at least one of the proofreading criteria.

Abstract: The voice print system of the present invention is a subword-based, text-dependent automatic speaker verification system that embodies the capability of user-selectable passwords with no constraints on the choice of vocabulary words or the language. An estimate of the enrollement channel and of the test channel is developed for inverse filtering of the enrollment or the test speech, respectively. Automatic blind speech segmentation allows speech to be segmented into subword units without any linguistic knowledge of the password. Subword modeling is performed using a multiple classifiers. The system also takes advantage of such concepts as multiple classifier fusion and data resampling to successfully boost the performance. Key word/key phrase spotting is used to optimally locate the password phrase. Numerous adaptation techniques increase the flexibility of the base system, and include: channel adaptation, fusion adaptation, model adaptation and threshold adaptation.

Abstract: A method for generating candidate word strings in speech recognition is provided, which is based on the nodes in the word lattice to search candidate word strings. The associated maximum string score for each node is first determined. Next, all nodes are sorted based on the associated maximum string score to group the nodes with the same string score into the same node set. Then, the node sets with relative high string scores are selected to connect the nodes by their starting time frame and ending time frame, thereby generating the candidate word strings.

Abstract: A speech synthesis method that generates a speech pitch wave from a reference speech signal by subjecting the reference speech signal to one of Fourier transform and Fourier series expansion to produce a discrete spectrum, that interpolates the discrete spectrum to generate a consecutive spectrum, and that subjects the consecutive spectrum to inverse Fourier transform. A linear prediction coefficient is generated by subjecting the reference speech signal to a linear prediction analysis. The speech pitch wave is subjected to inverse-filtering based on the linear prediction coefficient to produce a residual pitch wave. Information regarding the residual pitch wave is stored as information of a speech synthesis unit in a voice period. A speech is then synthesized using the information of the speech synthesis unit.