WO2003017018A1

WO2003017018A1 - Satellite navigation system and attitude determination method of object using the same

Info

Publication number: WO2003017018A1
Application number: PCT/KR2002/001285
Authority: WO
Inventors: Sangjeong Lee; Chansik Park; Seokbo Son; Byungyeun Kim
Original assignee: Navicom Co., Ltd.
Priority date: 2001-07-06
Filing date: 2002-07-06
Publication date: 2003-02-27
Also published as: KR100387136B1; KR20030004643A

Abstract

The present invention provides a satellite navigation system and method for providing fast and precise determination of a 2 or 3-dimensional attitude of a vehicle by using functionally-divided three processes. The three processes comprises: a first process operated by using an interrupt signal with 1 mili-second generated at a correlator for performing a reading of correlation value, a synchronization of data bit and frame, and a code and a carrier tracking; a second process allocating satellites to channels of the correlator, acquisiting a code from the correlator, a phase and the number of a cycle of the carrier wave at a periodic interval of about 100 mili-seconds; and a third process for performing an acquisition of the satellite information, a communication with another device and a determination of the vehicle-attitude by calculating navigation solution and determining an ambiguity interger at a periodic interval of about 1 second.

Description

Satellite Navigation System and Attitude Determination Method of Object Using The Same

Field of the Invention

The present invention relates to a satellite navigation system using satellite vehicles and an attitude determination method of a vehicle by using the satellite navigation system, more particularly, to a satellite navigation system and a method for performing a fast and precise determination of 2 or 3-dimensional attitude of a vehicle by performing of three-stage processes.

Description of the Prior Art

Inertial measurement systems such as a gyroscope or an inertial measurement unit (IMU) have been widely utilized for determining an attitude of a vehicle. The inertial measurement system, however, bears certain disadvantages that the system is vulnerable to trouble or been breakdown due to the complexity of the system. Further, analysis and processing of data are very complicated since a drift-rate error is prone to be accumulated as time passes, and finally the system is too expensive. Researches concerning a global positioning system

(GPS) applied to an attitude determination technology have been actively performed to overcome the disadvantages of the inertial measurement system.

The GPS has been widely used in determining a position of a vehicle on the ground thanks to precision^' and convenience. .The GPS is used generally in the area that absolute position is sought by using a coarse acquisition code (C/A code) or a precision code (P code) in condition that one GPS receiver involves one antenna. Positioning results by means of the C/A code may have error of about 100 meters (2dRMS: Root Mean Square) , affected by a selective_. availability. As an alternative, a differential global positioning system (DGPS) has been proposed with a view, to improving a positioning precision, wherein a base station, whose position is already known, estimates errors included in measurements for each satellite and transmits information on the errors to other neighboring GPS receivers, and thus the GPS receiver can eliminates errors with regard to the selective availability, an ionospheric time delay, a tropospheric time delay and a satellite orbit and finally determines the position of a vehicle with a precision of below meter order.

In a geodesy field, the GPS system utilizes a phase signal of a carrier rather than a code signal and determines the position with high precision. Positioning result using the C/A code has a resolution of about 3 meters when the GPS receiver processes GPS signals with 1 percent of resolution since bit interval i.e., a chip of the C/A code has wavelength of 300 meters. However, the wavelength of the carrier wave is 19 centimeters, positioning result using the carrier has a precision of 1.9 millimeter by means of signal process^' with 1 percent resolution, and thus more precise position determination can be achieved. Information on attitude of a vehicle can be obtained by using carrier phase measurement for carrier wave of the GPS signals. In the situation that two antennas are disposed toward a moving direction of the vehicle, a 2-dimensional attitude of the vehicle can be determined as a pitch (or roll) and a yaw. Similarly, a 3-dimensional attitude of the vehicle can be determined with three antennas deployed on the vehicle.

In order to determine an attitude of a vehicle by means of the DGPS or an attitude determination GPS with more than two antennas, precise carrier phase measurements of carrier obtained by a plurality of the antennas are required in addition to pseudorange and satellite position information. An integer ambiguity included in a double- differenced carrier wave should first be resolved to determine the attitude of the vehicle by using the carrier phase measurements obtained. The attitude of the vehicle can be determined by obtaining a baseline vector by means of the carrier phase measurements with resolved integer ambiguity, and by transforming the baseline vector into a navigation coordinate system and comparing the baseline vector with a body coordinate system. The baseline vector is a relative position vector directing from a reference antenna to a target antenna or an auxiliary antenna.

The carrier phase measurements received from at least^' three antennas are utilized^' to determine 3-dimensional attitude of the vehicle by using the GPS. Antennas comprise one reference antenna (Antenna REF) and a few auxiliary antennas (Antenna USER) . Measurements from the reference antenna may be used for obtaining a user position and a user attitude, and measurements from the auxiliary antennas are used for determining the baseline vector. Satellite tracking process should be independently performed for each antenna since each antenna should track identical satellite. Furthermore, a high-performance carrier tracking loop should be configured to acquire precise carrier phase measurements, thereby acquiring precise 3- dimensional attitude.

An integer ambiguity resolution process in the precise carrier phase measurement should be implemented in real time. A baseline vector determination process for determining the baseline vectors between the reference antenna and the auxiliary antennas and a coordinate transformation process for transforming the baseline vector into the navigation coordinate system are also required.

In accordance with the present invention, software programmed in a GPS receiver is classified into and separately implemented by three processes or routines. The precise carrier phase measurements are obtained from four antennas, and an ambiguity resolution with constraint equation (ARCE) method is used as an integer ambiguity resolution algorithm which may easily be applied to a real time implementation involving less calculation volume and employing smaller memory.

Summary of the Invention

An object of the present invention is to design a vehicle-attitude determination system and method by classifying the GPS receiver software into three parts to the processing time and function.

Another object of the present invention is to provide a combined frequency lock loop(FLL) with a phase lock loop(PLL) so that the carrier tracking performance is enhanced for determining the attitude of vehicle.

Another object of the present invention is to provide a method capable of assigning a specific satellite to the corresponding channel with the corresponding number at each DSP. Another object of the present invention is to provide a fast and precise integer ambiguity resolution method by using a method for validating an integer ambiguity by using only independent integer ambiguities, a method for validating the integer ambiguity by using angle information between baseline vectors, a method for validating the integer ambiguity by reducing a search range for the integer ambiguity by using velocity information and/or an inertial measurement unit .

Another object of the present invention is to provide a method for ^' validating a vehicle-attitude by means of vehicle velocity information and time variation, information of a vehicle pitch (elevation) and/or yaw (azimuth) .

A satellite navigation system in accordance with the present invention comprises: a satellite signal receiving antenna section including a reference antenna and 2 or more auxiliary . antennas; 3 or more . RF/IF (radio frequency/intermediate frequency) sections for converting satellite signals that are respectively received at the antennas of the receiving antenna section into IF signals and digitizing IF signals; 3 or more correlator sections, each including 5 or more tracking modules, for generating correlation values at the tracking modules by using the digitized IF signals and for tracking the satellite signals; a central processing unit for obtaining navigation solution and determining attitude of a navigating vehicle by using code and carrier phase measurements obtained at the correlator sections; and an input/output section for data communication between external devices and the central processing unit.

The attitude determining method using the satellite navigation system comprises: a first process performed by using interrupt signals within 1 ms generated at the correlator sections for reading and storing of the correlation values generated at the correlator sections, synchronizing data bit and frame of the satellite signals, and tracking the code and the carrier; a second process performed in a period longer than that of the first process •with a priority lower than that of -the first process for allocating satellites to channels of the respective correlator sections and acquiring the code and carrier phase information and the number of carrier cycles generated at the correlator sections; and a third process performed in a .period longer than tha.t of the second process with a priority lower than that of the second process for acquiring satellite information, communicating with the external devices, calculating navigation solution, and determining vehicle attitude through integer ambiguity resolution.

The execution period of the second and the third process are preferably 100 milliseconds and 1 second respectively.

Code and carrier tracking in .the first process is preferably performed by correcting error calculated at the central processing unit by using the correlation values obtained from the correlator sections, and a carrier tracking loop adopts a mixed structure of a FLL and a PLL; the PLL being operated only if frequency error is less than a threshold during FLL operation, the FLL being operated again if the frequency error is greater than the threshold. Characteristic of a tracking loop is determined depending on the order of the loop filter, signal integration time, and bandwidth of the filter. The carrier tracking loop used in the present invention employs a second-order loop filter as the FLL and a third-order loop filter as the PLL. The signal integration time and the loop filter bandwidth are adjusted according to experiment. Satellite allocation in the second process includes the steps of:

(2-1) determining whether all of visible satellites are allocated to the channels of the respective correlator sections at every TIC; (2-2) allocating remaining satellites to idle channels of the correlator sections when all of the visible satellites are allocated;

(2-3) if there is any visible satellite which is not allocated, allocating it to a idle channel, and if there is no idle channel, a channel to which an unpredicted satellite is allocated is disregarded so as to allocate the visible satellite and then proceeding to a next TIC.

Thus, the respective antennas can track the same satellite group since satellite allocation process is made at the respective antennas, independently. A satellite allocation process in accordance with the present invention will be described in detail referring to Fig. 4 .

Vehicle attitude determination in the third process includes the steps of: (3-1) determining an independent integer ambiguity search range for a first baseline between a reference antenna and a first auxiliary antenna, resolving a true integer ambiguity by searching candidates within the independent integer ambiguity search range, and obtaining a first baseline vector by using the true integer ambiguity;

(3-2) determining an integer ambiguity search .range for a second baseline by using an angle information between the first baseline vector and the second baseline vector between the reference antenna and a second auxiliary antenna, resolving the true integer ambiguity by searching candidates within the integer ambiguity search range,, and obtaining the second baseline vector by using the true integer ambiguity;

(3-3) if there is a third or more auxiliary antennas, applying the step (3-2) to the third or more auxiliary antennas to resolve the true integer ambiguity for the third or more baselines and obtaining a third or more baseline vectors; and

(3-4) determining 3-dimensional attitude of the vehicle by using any set of at least two baseline vectors.

An integer ambiguity resolution method in accordance with the present invention is performed by using an ambiguity resolution with constraint equation (hereinafter "ARCE") method, and therefore, searching for only independent integer number suffices for the integer ambiguity resolution.

According to the ARCE method, it is performed by first determining a search range for independent integer ambiguity, determining candidate integers within the search range, the candidate integers making the .value of an objective function to the minimum, as true independent integer ambiguity, and resolving dependent integer ambiguity by means of the true independent integer ambiguity. Furthermore ^' the true independent integer ambiguity can be resolved by a ratio- test for candidate integers making the value of an objective function below the threshold.

Detailed description for the ARCE method is disclosed in Korean patent application No. 1997-57696.

The search range determination of the independent integer ambiguity for each baseline vector may be performed by each of the several methods : a first method using a error covariance of code measurements of satellite signals; a second method using the code measurement error covariance of the satellite signals, carrier phase measurement error covariance, and the length of the first baseline, each obtained at the second process; a third method using attitude information from an IMU (inertial measurement unit) installed on a vehicle, and a fourth method using velocity information of the vehicle obtained by the GPS. A search range according to the first method is the largest, and thus searching time thereby is the longest. The third method can provide relatively small search range. Provided that the vehicle moves with high and constant velocity, however, the fourth method may provide small search range. Details of the first and second method are disclosed in Korean patent applications Nos. 1997-057696 and 2001-21788 respectively. The third and fourth methods will be sketched here.

In case there are at least two baseline vectors formed by at least three antennas, a fifth method for reducing the search range by using information, on an angle between two baseline vectors may be used. Korean patent application No.

2001-19813 discloses the fifth method in detail.

The present invention provides methods for validating the integer ambiguity by exploiting at least one information .on a _. length of the. baseline vector obtained by using independent integer ambiguity, an objective function obtained by using dependent integer ambiguity, a velocity information (velocity error) of the vehicle, a length of the baseline vector and angle information between the baseline vectors, and output information of an inertial measuring unit.

In accordance with the present invention, the vehicle attitude can be validated on the basis of a first feature that a yaw (or azimuth) and a pitch (or elevation) can be determined by using only velocity vector if a vehicle moves with high speed, and a second feature that the yaw and the pitch of the vehicle cannot be abruptly changed.

Brief Description of the Drawings

The above and other objects and features of the present invention will become apparent from the following description of a preferred embodiment given in conjunction with the accompanying drawings, in which:

Fig. 1 shows ^' a schematic diagram for a hardware structure of the GPS in accordance with the present invention; Fig. 2 shows a software structure of the GPS in accordance with the present invention;

Fig. 3 shows a function module diagram for a software structure of the GPS in accordance with the present invention; . . . Fig. 4 is a diagram for explaining a satellite allocation process of the GPS in accordance with the present invention;

Fig. 5 is a flow chart illustrating a satellite signal tracking process in accordance with the present invention; Fig. 6 shows a structure of the correlator section in accordance with the present invention;

Figs. 7 and 8 illustrate carrier phase measurement principle and measurement process;

Figs. 9 and 10 provide overall flow and detailed flow of integer ambiguity resolution process and algorithm;

Fig. 11 offers a flow chart for illustrating integer ambiguity validation process used in the present invention;

Fig. 12 shows a process for validating the integer ambiguity using angle information between baseline vectors; and

Fig. 13 illustrates attitude determination and validation process used in the present invention. Detailed Description of the Invention

Hereinafter, the preferred embodiments of the. present invention will be explained with reference to the accompanying drawings .

Fig. 1 shows a schematic diagram for configuration of the GPS receiving system for use in attitude determination in accordance with the present invention. The GPS receiving system in accordance with the present invention is largely divided into a hardware part and a software part.

As shown in Fig. 1, the hardware part comprises RF sections, correlator sections, and a central processing unit (CPU) . As shown in Fig. 2, the software part comprises a satellite allocation unit, a signal processing section, a measurement acquisition and processing unit, an attitude determination unit, and a position and velocity determination unit. As shown in Fig. 1, for contributing to attitude measurement, the hardware part of the GPS receiving system is configured to include four antennas, four RF/IF (radio frequency/intermediate frequency) sections, four correlator sections, each including a DSP unit, the CPU, a memory section, and an input/output (I/O) section.

Each of the RF/IF sections converts a RF satellite signal received at its corresponding antenna into an IF signal and then digitizes the IF signal. In this embodiment, GP2010 circuitry chip distributed from "Zarlink" may be used. As a reference clock, TCXO (Temperature Compensated Crystal Oscillator) of 10 MHz is used. Each of -the RF/IF sections supplies an operating frequency of 40 MHz to its corresponding one of the correlator sections DSPl, DSP2, DSP3, DSP4 in differential form and receives the GPS satellite signal from its corresponding one of the four antennas to down-convert it three times so as to form the IF signal. Further, the each of the RF/IF sections transmits the digitized signals MAG, SIGN digitized by using a sample clock of 5.714 MHz that is supplied from its corresponding correlator section.

Each of the correlator sections includes a tracking module having 12 channels to generate correlation values at the tracking module by using the digitalized signal from its corresponding RF/IF section for use in tracking the satellite signal.

In this embodiment, the correlator sections are configured with four GP2021 chips. The GP2021 chip includes the tracking module having 12 channels and generates correlation values at the tracking module by using the digitalized signal from its corresponding RF/IF sections for use in tracking the satellite signal. Employment of four DSPs makes it possible to track signals and acquire code and carrier measurements in 48 channels at a time.

The CPU calculates navigation solution including position, velocity of a vehicle and time of a vehicle by using the code and the carrier phase measurements that are acquired at the correlator sections, and determines attitude of the vehicle. In this embodiment, StrongARM (model name: SA-1100) CPU in ARM series is used. StrongARM includes a DUART unit, a MMU (Memory Management Unit) and a cache therein and has processing speed of about 268MIPS when a crystal oscillator of 220 MHz is used.

The memory section is configured with ROM/RAM, . ROM part having 512 Kbyte capacity using four 1 Mbit ROMs, RAM part having 2 Mbyte capacity using four 4 Mbit RAMs .

The I/O section for data communication between external devices and the CPU includes two asynchronous communication ports and a synchronous communication port. The asynchronous communication port contributes to transferring receiver monitor and commands and receiving DGPS signals. The synchronous communication port contributes to communicating with other sensors such as an IMU.

Fig. 2 shows a software structure of the GPS receiving system for use in attitude determinations in accordance with the present invention, which comprises an interrupt processing routine (1 ms Task) for a first process as a signal processing section executed by an interrupt signal within 1 ms, a satellite allocation and measurement acquisition routine (100 ms Task) for a second process executed at every 100 ms, and navigation solution (position and velocity) calculation and attitude determination routine (1 sec Task) for third process executed at every 1 sec.

The interrupt processing routine (the first process) as the signal processing section has its priority over other routines (100 ms and 1 sec Tasks) and executed by the interrupt signal that is generated at every 505 μs at the correlator sections. The interrupt processing routine performs reading and storing of correlation values that are generated . at the correlator sections, synchronizing data bit and frame, and tracking of the code and carrier. Among them, detailed description will be omitted for correlation value reading and storing and data bit and frame synchronization because they are commonly practiced in the art using conventional GPS receiving system. Code and carrier tracking employs a carrier tracking loop having a combined structure with FLL (Frequency Lacked Loop) and PLL (Phase Locked Loop) , of which, detailed description will be provided with reference to Fig. 5 below.

The satellite allocation and measurement acquisition routine (100 ms Task) as the second process includes tasks for measurement acquisition and satellite allocation to channels. The measurement includes code, carrier phase information and the number of carrier cycles generated at the correlator sections. This routine may be divided into a satellite allocation section and measurement acquisition and processing section, each of which will be described in detail in the following.

The navigation solution calculation and attitude determination routine (1 sec Task) as the third process is classified into acquisition of satellite information, communication with external devices, calculation of the navigation .solution including the position, the velocity and the time of the vehicle, and determination of 2-dimensional or 3-dimensional attitude of the vehicle. Among them, because they are similar to those practiced in the conventional GPS receiving system or DGPS receiving system, detailed description will be omitted for acquisition of satellite information, communication with external devices, and computation of the navigation solution, but it will be described in detail for attitude determination in the following.

Fig. 3 shows a block diagram for illustrating a software structure of the GPS receiving system for attitude determination with functional modules in accordance with the present invention. The software structure comprises a satellite allocation section, a signal processing section, a measurement acquisition and processing unit, a navigation solution calculation unit and an attitude determination unit. Further, a pseudorange filter is provided in order to obtain navigation solution including the position, the velocity, and the time. Further, an integer ambiguity resolution unit is further provided in order to resolve the integer ambiguity. The functional modules illustrated in Fig. 3 will now be described in detail. 1) The satellite allocation section : 100 ms Task The satellite allocation section is configured to manage allocation of the satellite to the correlator sections by confirming visible satellites and satellite allocation status. In this embodiment, the four antennas include one reference antenna (Antenna REF) and three auxiliary antennas (Antenna USERs) . Measurements at the reference antenna are utilized as a reference for obtaining position and attitude of the vehicle and measurements at the auxiliary antennas are utilized to generate a baseline vector for determining the attitude of the vehicle.

Signals received at each antenna pass through the RF/IF sections prior to being processed at the correlator sections. For attitude determination, signals from the same satellite group should be used. Thus, because each antenna should keep track on its corresponding satellite, satellite allocation process is made independently at each antenna. Fig. 4 shows a satellite allocation process in accordance with the present invention. First, after resetting all of the channels as idle state, it is determined whether all of the visible satellites are allocated to the 12 channels of the DSPl of the correlator section. If there is found any visible satellite that is not allocated, a visible satellite is allocated to an idle channel in case that there is any idle channel after checking whether there is any idle channel in the DSPl. When there is no idle channel, an unpredicted satellite vehicle that has already been allocated is disregarded and the visible satellite is allocated to the corresponding channel. Then, the process steps over to a state for_, waiting for a next visible allocation. After all of the visible satellites are allocated to the channels of the DSPl, it is determined whether there is an unpredicted satellite vehicle. When there is found any unpredicted satellite, it is allocated to the idle channel of the DSPl, which completes the process . for allocating a .visible satellite to the DSPl. By identically applying the above- described process for the same visible satellite group to the DSP2, DSP3 and DSP4, the visible satellite is allocated to all of DSPs. Such a satellite allocation process is repeated for every TIC, the TIC being set as 100 ms in this embodiment.

By executing the process as described above, each in same visible satellite group can be allocated to corresponding channel of all the DSPs . The satellites are allocated to the channels of the DSPs in order of their numbers when there is, initially, no information about positions of the satellite and the user. But, after the positions of the satellite and the user are obtained during the satellite signal processing, visible satellites are first allocated and unpredicted satellite vehicles are allocated next.

2) The signal processing section; Interrupter processing routine : 1 ms Task

The signal processing section tracks the satellite signal by using correlation values that are processed at the correlator sections and consists of a code tracking unit and a carrier tracking unit. The signal processing section determines noise characteristics of the code and carrier phase measurements and dynamic characteristic of the receiver as well as acquires and tracks the satellite signal. First of all, configuration of a precise, carrier tracking loop is considered as of highest importance because precise carrier phase measurements should be used for attitude determination.

In the present invention, for precise acquisition of the carrier phase measurement, a satellite signal tracking loop filter dualy having a FLL and a PLL is used. Characteristics of the satellite signal tracking loop is determined depending on the order of the loop filter, signal integration time, and bandwidth of the filter. The satellite signal tracking loop adopted in this embodiment employs a second-order loop filter as the FLL and a third- order loop filter as the PLL. The signal integration time and the loop filter bandwidth are adjusted according to experiment .

Fig. 5 is a flow chart illustrating a satellite signal tracking process that is performed at the signal processing section of the present invention.

First, after obtaining the correlation value from the correlator section, a lock indicator is calculated. At this point, a DLL (Delay Lock Loop) is used for code tracking and the combined loop filter having the FLL and the PLL is used for carrier tracking. The DLL is a module employed in the receiver for making coincidence between PRN codes of the satellite and the receiver, which make the PRN codes that are generated at the receiver coincide with the PRN code of the satellite by shifting the PRN codes of the receiver.

When the lock state is in a code lock, a code tracking loop is performed. If the lock state is in a carrier frequency lock without a bit synchronization, a carrier tracking loop is performed by using the FLL. When frequency error is determined to be lower than a threshold value after completion of the bit synchronization^ the carrier phase is regarded as locked and the carrier tracking loop is performed by using the PLL.

The frequency error should be kept low prior to performing the PLL loop for the PLL to duly operate, and therefore, the FLL is used for initial signal tracking and the PLL is operated when it is determined that the frequency error is sufficiently small. Further, a dual or combined structure is implemented so that if the frequency error becomes larger than the threshold, the FLL is used again.

3) The measurement acquisition and processing unit : 100 ms Task

For a GPS carrier, GPS solely can measure only the phase by measuring LI or L2. However, precise range calculation requires the number of the carrier cycles between the satellite and the receiver as well as the phase. This unknown number of the carriers is termed an integer ambiguity, of which determination is the key to obtain the position and the attitude of the vehicle by using the carrier phase measurement.

It is not easy to resolve the integer ambiguity contained in the carrier phase measurement. However, once. the integer ambiguity is resolved, the precise pseudorange can be obtained by continuously calculating variation of the carrier until the satellite signal gets lost. That is, assuming that the integer ambiguity N is resolved at a certain time to, the precise pseudorange can be continuously obtained by calculating variation ΔΦ of the carrier phase measurement.

Actually, in order to acquire the carrier phase measurement, the carrier phase and the number of carrier cycles in a unit time should be measured, which are summarized in table 1. Table 1

The values in table 1 can be generated from the correlator sections while the structure of the correlator sections is shown in Fig. 6.

In each of the correlator sections, the received signal is multiplied by the carrier and the code and then the correlation value is passed to the central processing unit at 1 kHz periods. Further, the number of carrier cycles and the carrier phase that are generated during carrier tracking are stored in a register. The number of carrier cycles are calculated at CARRIER_CYCLE_COUNTER (E in Fig. . 6) and are stored in two . registers CHx_CARRIER_CYCLE_HIGH and CHx_CARRIER_CYCLE_LOW, and then the value of CARRIER_CYCLE_COUNTER is reset. The carrier phase is constructed with upper 10 bits from CARRIER_DCO

(D in Fig. 6), which is stored in CHx_CARRIER_DCO_PHASE .

Storing of the measurements is performed with reference to the TIC. The TIC period is an adjustable value and set to 100 ms in this embodiment.

After calculating the number of the carrier cycles and the carrier phase, the carrier phase measurement is calculated based on these values. The carrier phase measurement is obtained by integrating continuous carrier Doppler, i. e., obtaining the amount of psuedorange variation from a starting point for calculation of Doppler to the current point. Equations for computing Doppler and the carrier phase measurement value are given in Eqs . (1) and (2)

Eq . ( 1 )

Eq. (2) where f_D is the Doppler frequency, v is the pseudorange variation, Li is the carrier frequency, c is the light speed, f_DCo is the frequency that is measured during signal tracking, and f_N0M is the frequency when there is no Doppler.. That is, ICP (Integrated Carrier Phase) can be obtained, by integrating Doppler and the carrier phase measurement process is illustrated in Fig. 7.

Using Eq. (3), the carrier phase measurement .(ICP) can be readily computed.

ICP_n ~ 2π^ Number in Carrier Cycle Counter

+ Final Carrier DCO Phase - Initial Carrier DCO Phase

= (K +PH₁ +l -PH_O) + (K₂ +PH₂ + 1 -PH_Ϊ ) +A +(K„ +PH_n +l~PE_ll__l)

= (-PH₀ +K₁ +PH_l) + (-PH_λ +K₂ +l+P ₂)+A + (-PH_n + K_n + PH_n ) Eq.

( 3 )

The carrier phase measurement that is calculated with

Eq. (3) is integrated ten times per 1 second and then passed to a navigation solution calculation task and the attitude determination unit. Such a carrier phase measurement process is shown in Fig. 8. ^• 4) The attitude determination unit and integer ambiguity resolution unit : 1 sec Task

. In order to determine a 3-dimensional attitude of the vehicle, in this embodiment, four antennas disposed on the vehicle constitute 3 baseline vectors. For a 3-dimensional attitude determination, integer ambiguity for at least 2 baseline vectors should be resolved. In the integer ambiguity resolution, double-differenced carrier phase measurement is used and ARCE is used as the algorithm therefor. The integer ambiguity resolution by using ARCE is disclosed in detail in another Korean patent application

(1997-57696) and thus, no further description will be made here . The ARCE for resolving the integer ambiguity uses the code measurement for initial integer ambiguity resolution and further uses known baseline vector length information as well as the code measurement for search range determination. The overall flow of integer ambiguity resolution process and algorithm are shown in Figs. 9 and 10.

Referring to Fig. 9, in case that three or more antennas are involved to thereby form two or more baseline vectors, error covariance of the code measurements and baseline vector length information for a first baseline vector are used in the search range determination, and angle information between the baseline vectors is used together in integer ambiguity resolution for a second and other baseline vector (s). The baseline vector length information and the angle information contributes to reducing the number of the integer ambiguity to fast perform^' the integer ambiguity resolution. The method including determining search range for integer ambiguity by using error covariance of the code measurement, error covariance of the carrier phase measurement and satellite position (line-of-light vector) , and determining a candidate that minimizes an objective function among candidates within the search range or passes the ratio-test among the candidates that generate the objective functions smaller than a threshold as a true integer ambiguity is disclosed in Korean patent application no. 1997-057696 and no further description will be made herein.

Further, the method for resolving fast the- integer ambiguity by using the baseline vector length information and the angle information between two baseline vectors is disclosed in detail in Korean patent application no. 2001- 019813 filed by the applicant of the present invention.

Further, in addition to the prescribed integer ambiguity resolution methods, as will be described below, there is provided a method, in case of an vehicle having a IMU( Inertial Measurement Unit), for resolving a true integer ambiguity by determining reduced integer ambiguity search range with information from the IMU and obtaining the true integer ambiguity through search using the objective function, and a method for resolving a true independent integer ambiguity ^'within a search range having 27 candidates when the speed of the vehicle is greater than a threshold (6 m/s) and velocity error due to acceleration is less than a threshold (10%) .

* An integer ambiguity validation process The resolved integer ambiguity is not always correct. Even if . the resolved integer ambiguity is . correct, the integer ambiguity tends to change if a cycle slip occurs or satellite signals are blocked. The integer ambiguity validation is a process for re-calculating a dependent integer ambiguity by using the independent integer ambiguity that has been already resolved and comparing it with a threshold. During this process, when the cycle slip occurs or measurement noise abruptly increases, it is determined that the measurement is not valid and the integer ambiguity is resolved again. At this time, removal of the troublesome measurement is needed since it takes long time to resolve the integer ambiguity again.

To cope with this, since there arises many problems in the dependent integer ambiguity resolved on the basis of the satellite having a low elevation, a validation process by using only the independent integer ambiguity (a step of an integer ambiguity validation using the independent integer ambiguity) is added. .Further, when the velocity of the car is high and steady, it is possible to use the obtained velocity information, more particularly, velocity error information (a step of an integer ambiguity validation using the velocity error) .

Further, in case that the vehicle is equipped with the IMU such as a Gyroscope and an accelerometer, integer ambiguity validation can be performed by using information from^' an IMU (a step of an integer ambiguity validation using the IMU) .

Further, in this attitude determination system, three baseline vectors should be determined and the integer ambiguity validation should^' be performed for all of the baseline vectors. For a second baseline vector and a third baseline vector, because the angle between the vectors does not change even if attitude of the vehicle changes, the validation can be performed by using the angle information between the baseline vectors (a step of an integer ambiguity validation using the angle information between baseline vectors) .

Fig. 11 shows a process for integer ambiguity validation using the independent integer ambiguities, information from the IMU, and the velocity error and another process for troublesome case. Fig. 12 shows a process for integer ambiguity validation using angle information between baseline vectors. First, referring to Fig. 11, there is an integer ambiguity validation step using the independent integer ambiguity comprising obtaining the length of the baseline vector with the independent integer ambiguity, comparing the obtained length value with the known length of the baseline vector, and checking whether the difference between them is greater than a threshold. When the difference is greater than the threshold, the process goes to a step of the integer ambiguity validation using the IMU and/or a step of the integer ambiguity validation using the velocity error. When the difference of the baseline lengths is less than the threshold, the dependent integer ambiguity is resolved by using the independent integer ambiguity (with ARCE) and it is checked whether the objective function calculated by using the dependent integer ambiguity is greater than a threshold. When the objective function is less than the threshold, it is determined that the integer ambiguity is valid and attitude determination is performed. When the objective function is greater than the threshold, the dependent integer ambiguity is re-resolved with the current position of the satellite for next validation process and it is checked again whether the difference between the baseline lengths that are obtained by using all of the integer ambiguity and the known baseline length is greater than the threshold. If the difference of the baseline lengths is less than the threshold, it is determined that the integer ambiguity is valid and attitude determination is performed. If the difference is greater than the threshold, it is determined that the integer ambiguity is not valid.

Of course, the present invention is not limited to the above-described order, but any combination of the integer ambiguity validation methods can be used depending on required attitude determination precision or processing time. As described above, in the. present invention, there are provided a validation method only using the independent integer ambiguity for rapid precise validation and a validation method using the angle between two baseline vectors in case of three or more antennas are further involved. Furthermore, for processing after validation, a method for resolving a true integer ambiguity within a reduced integer ambiguity search range by using information from the IMU and a method for resolving the true integer ambiguity by search through 27 integer ambiguity candidates when speed and velocity error satisfy a certain condition (faster than 6 m/s, less than 10%) are used.

Eq. (4) is used in the step of integer ambiguity validation using the independent integer ambiguity.

'ΆBJ ⁼ " ΆBJ^Ϊ'AB ^ ANABJ ^{+ W}AB Eq . ( 4 )

where 1_AB,_I is a double-differenced carrier phase measurement, H is a line-of-sight vector, λ is a carrier wavelength, N_AB,_I is a double-differenced independent integer ambiguity, and w is a measurement noise. Herein, 1, H, and w are measured values and the baseline vector can be obtained by substituting the resolved independent integer ambiguity, accordingly, the length of the baseline can be obtained. Because the actual length of the baseline is already known, the obtained baseline^' length is compared with the actual^' length to check whether the difference between the is greater than a prescribed threshold. The prescribed threshold may be determined depending on required degree of precision. For example, when the length of the baseline is about 1 m, the threshold may be defined as about 2 to 3 cm.

.The method . for resolving the dependent integer ambiguity from the independent integer ambiguity, and comparing the objective function calculated by using the dependent integer ambiguity with a threshold is disclosed in detail in Korean patent applications no. 1997-57696 and no. 2001-19813 and further detailed description will not provided herein.

The step for validation and processing using the IMU may be comprising : determining whether the IMU is installed in the vehicle, obtaining the 3-dimensional attitude of the vehicle using the information from the IMU, converting it to a relative position vector and resolving the integer ambiguity, more particularly, determining the integer ambiguity search range.

It will now be briefly described for the integer ambiguity search range determination using the IMU. The baseline vector r" in rectangular coordinate is converted to one in polar coordinate as Eq. (5) . Here, b, ψ, .and θ. indicate the length, the yaw (azimuth) and the pitch [elevation) of the baseline vector, respectively.

Eq. (5)

Linearization of Eq. (5) for a reference point b₀, ψo . and θ₀ results in Eqs. (6) and (7)

Eq. (6)

b₀cosψ₀co$θ_ϋ U_Q Sffi XJ/_Q Sin C7_Q ^— s oimn ψ ψ_0Q couso θu₀ δψ

-Z>₀sin^₀cos<9₀ - έ₀ cos ψ₀ sin 6>₀ cos ψ_ϋ cos <9_C δθ

0 έ₀ cos<9₀ sin θ_n δb

Eq. (7)

In Eqs. (6) and (7), b₀, ψo . , and θ₀ can be obtained from IMU data and δb, δψ., and δθ can be estimated based on experience or characteristics of the used IMU. From this result, integer ambiguity resolution reference value N and error δN can be expressed as Eqs. (8) and (9) ..

Eq. (81 δN = ~(HC:,δf - ω)

Eq. (9) ^'

From Eq. (9) , integer ambiguity error variance can be represented as Eq. (10) and this is used as the estimated integer ambiguity search range. σ =E[δNδN^τ]

= ιc τ^» _ψE[δ_Ψδψ^τ ]T;C?H^T + σl

Eq. (10) Consequently, the integer ambiguity search range using the IMU is determined as Eq. (11), where β is a reliability for the integer ambiguity presence range.

N-βσ_N ≤ N ≤ N+βσ_N Eq . ( 11)

Using the information of the IMU as described, time required for resolving the integer ambiguity can be reduced because the integer ambiguity search range can be reduced and the integer ambiguity search can be performed within the reduced range by using the objective function. The method for resolving the true integer ^' ambiguity among the integer candidates within the search range by using the objective function is disclosed in Korean patent application no. 1997- 057696 and further detailed description will not provided herein.

Next, it will be described in- detail for the validating step for the integer ambiguity using the velocity error.

It is necessary to analyze the precision of the velocity information that is obtained at the GPS in order to use the velocity information. Eq. (12) represents double- differenced carrier phase measurement. l-λN = Er' +w

Eq. (12) where 1 is the double-differenced carrier phase measurement, λ is a GPS carrier wavelength, N is the integer ambiguity, H is a matrix consisting of line-of-sight vectors, r^e is a relative position vector of the baseline, and w is a double- differenced carrier phase measurement noise. In Eq. (11) , unknown values are the integer ambiguity N and the relative position vector r^e. Accordingly, if the relative position vector can be known by using the velocity information, the integer ambiguity is readily resolved. That is, relation^" between the integer ambiguity estimated by using r^e = J (BLx(v^e /\v^e\))dt and the velocity vector can be represented as Eq. (13) .

Eq. (13) where BL is the length of the baseline vector, v^e is the velocity vector that is obtained at the GPS. Eq. (13) can be divided into two terms of a true component and an error component as presented in Eq. (14) .

Eq. (14)

Accordingly, the integer ambiguity search range can be determined as in Eq. (14) by using the velocity information that is obtained at the GPS and an integer making the objective function minimum among the integers within the search range is resolved as the true integer ambiguity. The time required for integer ambiguity resolution can be reduced because the integer ambiguity search range determined by using the velocity information is very narrow when the vehicle is at a steady velocity.

Here, the error component can be represented as Eq. (15) .

I +h_i3δv + (h δv_x +h_jXδv_y +h_nδv_∑))

< 2 x velocity error x baseline length l( speed x wavelength)

Eq . ( 15)

where hi is a line-of-sight directional vector between the antenna and the satellite i and δ_v is a respective directional component of the velocity vector error.

As shown in Eq. (15) , the estimated integer ambiguity error, is proportional to the velocity error and inverse- proportional to the speed, and the velocity error can be divided into an error due to measurement and an error due to acceleration.

The magnitude of the velocity error due to the GPS measurement is about 0.5 m/s in steady .velocity cruise. Further, error occurs between the actual velocity and the estimated velocity due to movement of a car. Particularly, when the acceleration is high, difference between the actual velocity vector and the estimated velocity vector increases because the velocity vector error increases . As found from experiment, when the speed of the vehicle is greater than 6 m/s, the integer ambiguity error is less than 1, and when the magnitude of the velocity error due to acceleration is less than 10% of the speed, the integer ambiguity error is less than 1. Accordingly, if the speed is greater than 6 m/s and the velocity error due to acceleration (speed variation/determination period) is less than 10% of the speed, the number of the search candidates is reduced to 27 so as to enable rapid integer ambiguity search. Herein, the number 27 is determined as 3 x 3 x 3 because the range is determined as 3 (n ± 1) that is wide enough for respective directional components for the 3 independent integers . Accordingly, in the step of validating the integer ambiguity and processing by^' using the velocity error, the magnitudes of the speed and the velocity error obtained at the GPS are compared with the threshold, i.e., it is checked whether the speed is greater than 6 m/s and the velocity error due to acceleration is less than 10% of the speed. If the velocity error is greater than the threshold, it is regarded that the integer ambiguity is not valid. If the velocity error is less than the -threshold, the true integer ambiguity is resolved by performing integer ambiguity search for 27 candidates.

Further, when the number of the baseline vectors is 2 or more, the .integer ambiguity validation may be performed by using the angle information between the two vectors, called ^xa step for validating the integer ambiguity using an angle information between baseline vectors' .

Fig. 12 is a flow chart of the step for validating the integer ambiguity using an angle information between baseline vectors, in which a first baseline vector and a second baseline vector are sequentially determined by using the integer ambiguity, the carrier phase measurement and the satellite position (line-of-sight directional vector). Because the lengths of the two vectors are known, the angle information can be calculated by performing inner product of the two vectors'. The calculated angle information is compared with the known actual angle to determine whether their difference is greater than a certain threshold. If the calculated angle information is greater than the actual angle, it is determined that the integer ambiguity is ^' not valid. If the difference is less than the threshold, it is determined that the . integer ambiguity is valid. The. threshold may be selected depending on required determination precision.

* Attitude determination process and attitude validity determination . . _. In the system of the present invention, 3 baseline vectors are formed by using 4 antennas to determine the 3- dimensional attitude of the vehicle. When 2 or more baseline vectors among the 3 baselines are determined, the 3-dimensional attitude can be determined, and when only 1 baseline vector is determined, the 2-dimensional attitude can be determined.

Because the carrier phase measurement is used for attitude determination, it is possible to perform attitude determination only after correct integer ambiguity is resolved.

Upon resolving the integer ambiguity, the baseline vectors between the antennas are determined by using the carrier phase measurements, the satellite position information and the resolved integer ambiguity and, accordingly, the 2-dimensional or the 3-dimensional attitude of the vehicle can be determined by using the baseline vectors. Accordingly, the integer ambiguity should be resolved within a short time and it should be checked whether the resolved integer ambiguity is correct.

As described above, for integer ambiguity validation, there are a method using the baseline vector length obtained by using the independent integer ambiguity in the integer ambiguity search process, a method by using the dependent integer ambiguity, a method by using the velocity error etc. and reducing the integer ambiguity, search range, and a method by using the IMU information and reducing the search range. However, there could be a case in that incorrect integer ambiguity is continuously determined as valid one. Thus, there is provided a method for validating attitude by using other information for precise attitude determination. The attitude validation may be performed considering the fact that the yaw (or azimuth) and the pitch (or elevation) of the vehicle can be obtained by using only the velocity vectors if the vehicle moves fast and the fact that the pitch (elevation) and the roll angle of the vehicle cannot have large values if the vehicle does not move.

In particular, any one of the following 3 methods may be used to validate the vehicle attitude: a first attitude validation method determining that the determined attitude is valid only if variation in time of pitch (elevation) of the vehicle between current epoch and previous epoch is less than a threshold; a second attitude validation method determining that the determined attitude is valid only if variation in time of vehicle yaw between current epoch and previous epoch is less than a threshold; and a third attitude validation method determining that the determined attitude is valid only if the difference between the vehicle yaw obtained from the velocity vector of the vehicle when the speed of the vehicle is greater than a threshold and the vehicle yaw obtained from the baseline vectors is less than a threshold.

. Fig. 13 illustrates the attitude determination and the attitude validation process.

It is checked whether the integer ambiguity is resolved for all of the 3 or more baselines vectors. When there is the integer ambiguity that is not resolved, the process goes through the integer ambiguity search routine as shown in Fig. 10. When the integer ambiguity for the 2 or more baselines are resolved, the baseline vectors are determined based on the resolved integer ambiguity and, accordingly, the 3-dimensional attitude of the vehicle represented by yaw (or azimuth) , pitch (or elevation) and roll is determined.

When the integer ambiguity for only one baseline vector is resolved, the baseline vector is determined based on the resolved integer ambiguity and the 2-dimensional attitude of the vehicle represented by the yaw and the pitch is determined.

Upon determining the 2-dimensional or 3-dimensional attitude of the vehicle, the validation for the vehicle attitude is performed by using the yaw and the pitch, and the velocity (speed) information of the vehicle.

In the attitude validation process, it is checked whether time variation of the pitch between the current epoch and the previous epoch or time variation of the pitch (elevation) is greater than a certain threshold. If it is greater than the threshold, the determined attitude is determined to be not valid. If the time variation of the pitch (or elevation) is less than the threshold, .it is further checked whether the speed is greater than a certain threshold. If the speed is less than the threshold, the attitude is determined to be valid and the process is finished. If the speed is greater than the threshold, it is checked whether an yaw error (difference) between the yaw obtained from the velocity vector and the yaw obtained from the baseline vector is larger than a certain threshold. If the yaw error is larger than the threshold, the attitude is determined not to be valid; and, if otherwise, the attitude is determined to be valid. The threshold for the time variation of the pitch, the threshold for the speed, and the threshold for the time variation of the yaw may be selected depending on required degree of reliability or precision of the validation. When the attitude is determined not to be valid, the above- described process for resolving the integer ambiguity and determining the attitude is repeated in order to determine a new attitude. It will be apparent to those skilled in the art that other particular embodiment can be made without changing the technical spirit or essential feature of the invention. Therefore, it should be understood that the embodiment, as described above is not intended to limit the invention but to provide an example. The scope of the invention is defined by the appended claims rather than the detailed description and any change or modification incited from meaning, scope and their equivalences of claims should be understood to be included in the scope of the invention.

Industrial Application

As described above, using the attitude determining method of the invention, the 2-dimensional or 3-dimensional attitude of the vehicle can be determined fast and precisely with a GPS system having 3 or more antennas basically.

Further, by employing the combined carrier tracking loop having a FLL and a PLL, satellite signal tracking performance is improved and search range for the integer ambiguity is reduced in various ways to, accordingly, reduce time required for integer ambiguity resolution.

Further, an invalid integer ambiguity that is generated due to cycle slip and the like can be readily found and processed by validating the resolved integer ambiguity. Further, the degree of precision of the attitude determination is improved by validating finally again the determined attitude of the vehicle, .

Claims

What is claimed:

A satellite navigation system comprising: a satellite signal receiving antenna section including a reference antenna and 2 or more auxiliary antennas;

3 or more RF/IF (radio frequency/intermediate frequency) sections for converting satellite signals that are respectively received at the antennas of the receiving antenna section into IF signals and digitizing IF signals;

3 or more correlator sections, each including 5 or more tracking modules, for generating correlation values at the tracking modules by using the digitized IF signals and for tracking the satellite signals; a central processing unit for obtaining navigation solution and determining attitude of a navigating vehicle by using code and carrier phase measurements obtained at the correlator sections; and an input/output section for data communication between external devices and the central processing unit.

An attitude determining method that is performed by using the system as recited in claim 1, the method comprising: a first process performed by using interrupt signals within 1 ms generated at the correlator sections for reading and storing of the correlation values generated at the correlator sections, synchronizing data bit and frame of the satellite signals, and tracking the code and the carrier; a second process performed in a period longer than that of the first, process with a priority lower than that of the first process for allocating satellites to channels of the respective correlator sections and acquiring the code and carrier phase information and the number of carrier cycles generated at _. the correlator sections; and a third process performed in a period longer than that of the second process with a priority lower than that of the second process for acquiring satellite information, communicating with the external devices, calculating navigation solution, and determining vehicle attitude through integer ambiguity resolution.

The method as recited in claim 2, wherein tracking the code and the carrier in the first process is performed by correcting error calculated at the central processing unit by using the correlation values obtained from the correlator sections, and a carrier tracking loop comprises a combined structure with a FLL (Frequency Lock Loop) and a PLL (Phase Lock^' loop) , the PLL being operated only if frequency error is less than a threshold during FLL operation, the FLL being operated again if the frequency error is greater than the threshold.

The method as recited in claim 2 or 3, wherein the satellite allocation in the second process includes the steps of:

(2-3) if there is any visible satellite which is not allocated, allocating it to an idle channel, and if there is no idle channel, a channel to which an unpredicted satellite is allocated is disregarded so as to allocate the visible satellite and then proceeding to a next TIC; and

(2-4) repeating the steps (2-1) to (2-3) until all of the visible satellites are allocated so that all of' the visible satellites are allocated to corresponding channels of the respective correlator sections. The method as recited in claim 2, wherein the vehicle attitude determination in the third process includes the steps of:

(3-1) determining an independent integer ambiguity search range for a first baseline between a reference antenna and a first auxiliary antenna, resolving a true integer ambiguity by searching candidates within the independent integer ambiguity search range, and obtaining a first baseline vector by using the true integer ambiguity;

(3-2) determining an integer ambiguity search range for a second baseline by using an angle information between the first baseline vector and the second baseline vector between the reference antenna and a second auxiliary antenna, resolving the true integer ambiguity by searching candidates within the integer ambiguity search range, and obtaining the second baseline vector by using the true integer ambiguity;

(3-4) determining 3-dimensional attitude of the vehicle .by using any set of at least two baseline vectors.

The method as recited in claim 5, wherein the integer ambiguity search range determination in the step (3-1) is performed by one of the methods of:

(3-1-1) using code measurement error covariance of the . satellite signals. and satellite position information, each obtained at the second process;

(3-1-2) using the code measurement error covariance of the satellite signals, carrier phase measurement error covariance, the satellite position information, and the length of the first baseline, each obtained at the second process;

(3-1-3) using the length of the first baseline, the satellite position information, and velocity information obtained in navigation solution computation of the third process; and (3-1-4) using attitude information from an

IMU (inertial measurement unit) installed on a vehicle, and the satellite position information.

7. The method as recited in claim 6, wherein the integer ambiguity search range of the (3-1-3) method is determined as follows:

where 1 is a double-differenced carrier phase measurement, H is a double-differenced line-of-sight vector, BL is a length of the first baseline, and v^β and δv^e are a velocity vector generated in navigation solution computation and .error component thereof, respectively.

The method as recited in claim 6, wherein the integer ambiguity search range of the (3-1-4) method is determined as follows:

N-βσ_N ≤ N ≤ N+βσ_N

N^(l-HC_nf!l)

■ E[δNδN^τ]

= j_ϊ ^C:T_s"_ψE[δψδψ^τ ]τζCfH^J + σ:

b₀ cos ψ_ϋ cos θ₀ b₀.cos ψ₀ sin θ₀ - sin ψ₀ cos θ_ϋ - b₀ sin ψ₀ cos 6^ - έ₀ cos ψ_ϋ sin < ₀ cos ψ₀ cos <9₀ δθ

0 Z>₀ cos θ₀ sin ft, δb

where C„ is a coordinate conversion matrix, r"₀ and δr" are a linearized value of the polar coordinate value of the vehicle attitude information obtained at the IMU with reference to a linearization point and error component .thereof, respectively, β is reliability for entire range of the integer ambiguity,, b , φ , and θ are the length, the yaw, and the pitch of each baseline vector, respectively.

The method as recited in one of claim 5 to 8, in order to validate the integer ambiguity resolved in the third process, further comprising at least one of:

a first validation method comprising the steps of obtaining the length of the baseline vector by using a resolved independent integer ambiguity, comparing the difference between the obtained length of the baseline vector and the actual length of the baseline vector with a threshold, and determining that the integer ambiguity is valid only if the difference is less than the threshold; a second validation method comprising the steps of obtaining an objective function by using the resolved dependent integer ambiguity and the position information of the satellite, comparing the obtained objective function with a threshold, and determining that the integer ambiguity is valid only if the objective function is less than the threshold; a third validation method comprising the steps of comparing the speed of the vehicle and a velocity error (velocity error due to acceleration) obtained by navigation solution calculation with predetermined first threshold and second threshold, respectively, determining that the integer ambiguity is not valid if the speed is less than the first threshold or the velocity error is greater than the second threshold, and resolving the true independent integer ambiguity through an search for 27 integer ambiguity candidates if the speed is greater than the first threshold and the velocity error is less than the second threshold; and a fourth validation method comprising the steps of comparing an angle between the first baseline vector and the second baseline vector determined by using resolved integer ambiguities with an actual angle between the first baseline vector and the second baseline vectors, and determining that the integer ambiguity is valid if the difference between the angles is less than a predetermined threshold. ^■ The method as recited in claim 5 to 8, in order to validate the vehicle attitude determined in the third process, further comprising at least one of: a first attitude validation method determining that the determined attitude is valid only if variation in time of the pitch (or elevation) of the vehicle between a current epoch and a previous epoch is less than a predetermined threshold; a second attitude validation method determining that the determined attitude is valid only if variation in time of an yaw (or azimuth) of the vehicle between a current epoch and a previous epoch is less than a predetermined threshold; and a third attitude validation method determining that the determined attitude is valid only if the difference between the yaw of the vehicle obtained from the velocity vector of the vehicle when the speed of the vehicle is greater than a threshold and the yaw of the vehicle obtained from the baseline vector is less than a predetermined threshold.