7.1 FAILURE DETECTION
There are two types of failure detection, one to be performed when integrity is guaranteed by WAAS and the other to be performed to guarantee integrity with Receiver Autonomous Integrity Monitoring (RAIM)
i) Failure detection performed when integrity is guaranteed by WAAS
Basically, only satellites in WAAS HEALTHY state and guaranteed by WAAS are used, but it is required to detect local variation of these satellites so that the receiver can detect failure in these satellites. While the MOPS requires the failure detection only for the PA mode, our Receiver is designed to execute the failure detection for the navigation modes of NPA and En Route/Terminal as well in addition to the PA mode.
The residual of pseudo-range after computation of position-fixing with the weighted least square is divided by weight of each satellite to be normalized. Root Sum Square (RSS) of all satellites is calculated and then defined as detected statistical value. That value is compared with the pre-calculated threshold value. If that value exceeds the threshold value, the satellite is judged to fail and an alert of “FD Abnormality” is given to the satellite. The threshold value is derived from the failure detection probability and the number of satellites used for position-fixing.
Selection of the method of position-fixing is not directly affected by the occurrence of alert.
ii) Failure detection performed to guarantee integrity with RAIM
When integrity cannot be secured by WAAS, the Receiver alone must guarantee integrity. This is executed when the method of position-fixing is either one of 4) FDE Position-fixing, 5) FD Position-fixing or 6) FD Position-fixing using baro-altimeter. RSS of the residual of the pseudo-range after computation of position-fixing is calculated and then defined as detected statistical value. That value is compared with the pre-calculated threshold value. If that value exceeds the threshold value, the satellite is judged to fail. In case of 4) FDE Position-fixing, the satellite in which such an failure is found is identified and excluded. In case of the other methods of position-fixing, an alert of “FD Abnormality” is given.
The threshold value is derived from the failure detection probability, SA presence/absence and the number of satellites used for position-fixing. SA presence/absence is judged from URA of ephemeris.
7.2 EXCLUSION OF FAILED SATELLITE
This process is only executed for the case of the method of 4) FDE Position-fixing. Since the number of satellites for position-fixing and HELFD have already been checked at the phase of selection of method of position-fixing, it is assured that the failed satellite can be excluded.
Firstly, the failure detection is carried out using all satellites in accordance with subclause 7.1. If it turns out to be something wrong, the detected statistical value is calculated every time the satellite is removed one by one from the group of satellites used for position-fixing. This procedure is repeated until the last satellite remains in the group (This procedure gives the same number of detected statistical values as the one of satellites used for position-fixing).
When the detected statistical value shows minimum in the result of the calculation, the excluded satellite is registered in the list as failed satellite. The satellites registered in the list will not be used for position-fixing.
7.3 MONITORING OF EXCLUDED SATELLITES FOR RESTORATION
If there exists satellite excluded as the result of process of Exclusion of failed satellite (subclause 7.2) or pseudo-range step detection, in order to judge whether such satellite may be restored, the failure detection is performed for the satellite including all other excluded satellites. For this purpose, the same algorithm as the one of subclause 7.1 is used. To provide the judgement as normal with hysteresis, the threshold value is set to false alert probability of 0.05 and the prerequisite to judge to be normal is made more stringent.
7.4 CALIBRATION OF BARO-ALTIMETER
The calibration of baro-altimeter is the process in which the difference between the result of GPS position-fixing and baro-altitude is memorized as a correction amount which will be used when the baro-altitude is used for position-fixing. To estimate the error in the baro-altitude correction amount, the error in the vertical direction of the result of GPS position-fixing is calculated and memorized. In addition, to obtain the increase of error since calibration, altitude, latitude and longitude and time when the calibration is carried out are also memorized.
The error in the vertical direction of the result of GPS position-fixing is calculated using VPLFD obtained under the condition that false alert probability is 0.05 and missed alert probability is 0.32.
