Session C1, Paper #1

Aircraft Landing System Utilizing a GPS Receiver with Position Prediction Capability/Functionality
M. Grzegorzewski, J. Bialy, A. Ciecko, J. Cwiklak, P. Koscielniak, S. Oszczak, The Polish Air Force Academy

The article presents theoretical foundations of the position prediction capabili-ty/functionality to be used by a GPS receiver onboard an aircraft in the event of an instanta-neous lack of position data occurring due to various reasons. So far mathematical methods have been most frequently used for finding solutions to the problem of discontinuity of posi-tion data availability. In the presented work, an alternative filter has been developed which relies mostly on the parameters related to the motion of an aircraft: position, velocity and position error statis-tics. The assumptions of our research study use the concept of the "position potential" based on the data received from the navigation services. According to Newton´s law of universal gravitation, a free mass particle is attracted by another mass. Statistical confidence regions (position error ellipsoids) of the most probable position fix are regarded as "sources of force", which should "attract" a trajectory of an object passing through them. The potential field of a single particle, which in our case is an aircraft, should reflect the observed position fix and the required force to be exerted on the particle, which should monotonically decrease when the particle approaches the assumed position fix. When the particle enters the position error ellipsoid, it will be attracted with a force whose magnitude will be proportional to the intensity of the potential. The potential monoton-ically decreases with the decreasing of the distance between the particle and the assumed fix. What is more, in order to allow the particle to continue moving after the position fix appears, the potential of the attracting source will be dissipated exponentially in time. If we select the position density function ("position potential"), which contains the dissipation exponent ? and the parameter G (positioning uncertainty, corresponding to Newton´s gravitation con-stant), we assume that the trajectory of the particle will represent the real trajectory of the ve-hicle. The first attempt at developing a navigation filter for 2D space that was based on the above assumptions was made by Inzinga and Vanicek . First of all, we have to select the proper potential function for 3D space. Using the se-lected function of position density we can establish the time-related position potential field for a sequence of 3D position fixes. Subsequently, we set up a model of motion of an individual particle and find the solution of its motion equation. In order to reflect the changing navigational environment, the potential function contains certain variable para-meters,? and G. Parameters and initial motion conditions in the motion model are related to previous ob-servations by means of "motion equations". The estimation of a given future position of the particle (aircraft), i.e. making a prediction, is possible owing to the model of motion having the determined parameters and on the basis of the present position of the particle. Our task was to create an algorithm, which may be implemented, and then used in prac-tice, in order to predict subsequent positions of an aircraft. The results of our work are the following: 1. Three different algorithms which predict the 3D position of an aircraft in the subse-quent instances on the basis of its positions at previous instances. 2. Comparison of the efficiency of the created algorithms when using various values of the parameters. The results, obtained during performed aircraft experiments, of these three different methods of position prediction were analyzed and compared .

The starting point was for us the method based on real physical assumptions which is called a ?Gn method - from its two key parameters ? and G.

?Gn Method Technically, its essence lies in expressing the solution of an ordinary second order linear dif-ferential equation as a function of two parameters ? and G, and, subsequently, selecting the optimum values of those parameters, which will minimize the error function, F0. Knowing those values facilitates the determining of the predicted values. Bessel special functions are present in the solution of the equation. We decided to use for its optimization a certain version of an evolutionary algorithm devel-oped by ourselves. The parameters affecting the course of action of the evolutionary algorithm are: population size, number of discarded candidate solutions (individuals), width (step) of the local method, number of steps and, finally ranges: ?, G.

linear n Method This is one of the simplest methods. It assumes linear relationship between height and time.ª Coefficients bo,bi are determined with the least squares method on the basis of the course of the flight so far (n measurements, n > 2) and used for making a prediction.

baryc n Method This method is in fact one of the versions of the ARIMA (p, d, q) method, namely A(1,1, 0), Coefficients ?1, ?2 are determined with the least squares method on the basis of the course of the flight so far (n measurements, n > 2) and used for making a prediction.

The real data recorded during the flight of a Cessna aircraft was used for comparing the me-thods mentioned above and for optimizing the ? and G parameters.

On the basis of the conducted experiments, we hold the view that the methods to be imple-mented in further, in-flight tests are:ª 1. ? G2 and ? G3 2. linear 2

However, only after implementing them in a real instrument, a GPS receiver, the real opera-tional efficiency of each of them may be assessed. In the event of substantial flight distur-bances, none of those methods - or any other method known to us - may be used without a running a great risk of making a serious error. During the flight, the 3D prediction may be used as necessary (lack of GPS data) provided that there are no sudden changes of flight parameters. The accuracy of the prediction decreas-es with time. Therefore, the pilot ought to be warned after a given time (we recommend 3 sec-onds) that the position of the aircraft being shown may differ considerably from its real position.

[Return to Program]

Session C1, Paper #2

Impacts of ADS-B Service Availability Prediction Requirements on Scheduled Airline Operations
M. Harris, T. Murphy, Boeing Commercial Airplanes

On May 22nd, 2010, the FAA released a final rule mandating that aircraft flown in US airspace have ADS-B capability by 2020. The rule included performance requirements specifying a Navigation Accuracy Category for position (NACp) of 8 and a Navigation Integrity Category (NIC) of 7. These requirements mean that the estimated 95% horizontal accuracy of the ADS-B position broadcast must be less than 0.05 nautical miles (92.6 meters), and the horizontal protection level that represents a bound on the position error with a 1-10-7 probability must be less than 0.2 nautical miles.

While the FAA´s new ADS-B rule did not include an explicit requirement for availability of ADS-B capability at the specified levels, the FAA has indicated that a pre-dispatch prediction of ADS-B service availability will be required after 2020. ADS-B Out service predictions are based on the expected GNSS satellite geometry and performance, for the scheduled flight path and time. If the predicted positioning performance does not meet the ADS-B requirements, then the scheduled flight must be cancelled or rescheduled for a time during which the requirements are predicted to be met.

This paper explores ADS-B service availability and the performance of the SAPT tool as proposed by the FAA. The paper also explores some of the operational implications of requiring dispatch prediction and illustrates the possible consequence of relatively frequent denial of dispatch for scheduled airline operations compared with dispatch availability enjoyed currently thanks to the radar infrastructure. Modeling and analysis has been done to quantify ADS-B service availability, both actual and as predicted by the proposed SAPT tool. The duration of outages is considered and finally, the potential impact of the use SAPT on scheduled airline operations is illustrated.

[Return to Program]

Session C1, Paper #3

Collision Avoidance with Integrity Using Raw Measurements in the Automatic Dependent Surveillance - Broadcast
M. Uijt de Haag, Ohio University; J.L. Farrell, Vigil, Inc.; S. Vana, Ohio University

Future collision avoidance methods will rely heavily on availability and integrity of the Automatic Dependent Surveillance - Broadcast (ADS-B). This paper discusses an alternative ADS-B implementation that has many advantages over the current ADS-B implementation, especially with respect to safety, accuracy, robustness, and integrity of the solution. The proposed method uses available provisions (Mode-S, UAT and GPS receivers) and existing GPS algorithms and techniques (modern estimation in combination with GPS double differencing).
To deal with the increasing demand on the National Airspace System (NAS), U.S. government and U.S. agencies have been planning the Next Generation Air Transportation System (NextGen) which will transform the National Airspace System (NAS). Similarly, Europe has been addressing the transformation of the European Air Space through EUROCONTROL. This transformation is referred to as Single European Sky ATM Research (SESAR). One important enabler in both the NextGen and SESAR architectures is the Automatic Dependent Surveillance - Broadcast (ADS-B).

In the current ADS-B implementation information regarding the aircraft state is transmitted through the Mode-S Extended Squitter (ES) or the Universal Access Transmitter (UAT) at an update rate of about 2Hz. The transmitted state information includes airborne position (latitude, longitude, altitude) and airborne velocity (ground track, groundspeed, NS velocity, EW velocity, vertical rate). Besides position and velocity data, the aircraft also transmits a set of performance-related parameters including the navigation accuracy category for position, NACP, the navigation accuracy category for velocity, NACV, the navigation integrity category, NIC, and the surveillance integrity level, SIL. The values of these performance parameters are completely determined by the navigation sensor that is used as an input to the ADS-B receiver such as GNSS Ground-based Augmentation System (GBAS) or Space-based Augmentation System (SBAS) receivers.

In the proposed ADS-B implementation the participant would transmit raw measurements or encoded raw measurements through the ADS-B datalinks (Mode-S Extended Squitter or UAT) rather than transmitting the position, velocity, accuracy and integrity information. The receiver can then combine the incoming measurements with his own measurements to establish a track of the target. Advantages of this method are increased observability, independence on of a datum reference, the availability of an intrinsic duality indicator through the estimator´s covariance matrix, a guaranteed incorporation of existing correlations in the estimator, optimal weighting of the measurements, and the ability to perform data screening and other integrity methods such as including both conventional and advanced receiver autonomous integrity monitoring (RAIM) algorithms.

Given the raw measurements from both the traffic target and ownship, the pseudorange and carrier-phase sequential differences and double differences can be used to accurately derive the baseline between ownship and target as well as the change in baseline. The latter quantity is directly related to the average velocity over a time epoch. This paper will address the alternative ADS-B implementation details and a collision avoidance method that uses this implementation. Furthermore, the use of the available data bits for transmission of the raw measurements is discussed. Finally, simulation results are shown based on a sensor simulation framework developed by Ohio University for the NASA Langley Research center.

[Return to Program]

Session C1, Paper #4

A Coverage Analysis Methodology for APNT Pseudolite Ground Network
E. Kim, Selex Systems Integration

The FAA is looking into a robust Alternative Positioning, Navigation, and Timing (APNT) solution that can provide seamless operation from en route to non-precision approach during expected and unexpected outages of the Global Navigation Satellites System (GNSS). A promising alternative of FAA´s APNT architecture is a pseudolite-based ground network that utilizes the existing infrastructure of widespread Distance Measuring Equipment (DME) and Automatic Dependent Surveillance -Broadcast (ADS-B) Ground Based Transceivers (GBT) over the Contiguous United States (CONUS). However, the technical and economic feasibility of the proposed APNT architecture against the intended operation has not been assessed. To evaluate the feasibility of the pseudolite-based APNT architecture, one method is to conduct a coverage analysis that shows the accuracy and integrity performance trade-off between ranging accuracy and user-to-ranging source geometry. From this analysis, a system developer would be able to see the level of ranging signal accuracy, the required number of ground stations, and the network lay-out. This paper discusses the formulation necessary to evaluate the coverage analysis against the APNT accuracy and integrity requirements and the typical coverage limitations resulting from the requirements. Additionally, given an existing DME/GBT infrastructure in select CONUS regions, the coverage differences with respect to varying levels of ranging accuracy are investigated to determine the level of ranging accuracy that yields the desired coverage. Further, a systematic strategy is presented to determine the placement of additional ground stations in optimal locations of an existing DME/GBT infrastructure to enhance an insufficient APNT coverage. The result from this systematic ground station placement strategy will provide the required number of additional stations and network geometry in the select area.

[Return to Program]

Session C1, Paper #5

Joint Precision Approach and Landing System Why JPALS?
E. Brown, NAVAIRSYSCOM

The Joint Precision Approach and Landing System (JPALS) is a GPS-based landing system. JPALS is an approved program through the Office of the Secretary of Defense. Radar-based landing systems are aging technology. Though there is no official timeline, the FAA will be removing ILS service. GPS-based approaches are the future of aviation. Rather than having a fleet of aircraft who can´t file IFR to many fields because there is no approved IFR approach that matches aircraft capabilities (much like the F/A-18 and lack of ILS today), OSD and JROC have selected JPALS as the future DoD precision approach and landing system. The aviation world has embraced GPS-based landing systems. Space Based Augmentation Systems (SBAS) and Ground Based Augmentation Systems (GBAS) are being developed and fielded worldwide for both civil and military applications. In the U.S. alone, Wide Area Augmentation System (WAAS) approaches can be used at over 2300 airports according to the FAA. As part of JPALS, SBAS and GBAS are part of the complete JPALS package. This means that with JPALS, users will have the ability to operate from most airfields around the world should the need arise. The JPALS ship´s system and carrier air wing integrations are fully funded, on cost and on schedule. The Ship´s System held a Critical Design Review in December 2010 and Engineering Design Models will begin testing on CVNs (Aircraft Carriers) in 2013. JPALS is replacing the aging legacy landing systems currently in existence in the US Military. Some of the key systems affected are Tactical Air Navigation System (TACAN), Instrument Landing System (ILS), SPN-46 (Aircraft Carrier Precision Approach and Landing System), and SPN-35 (Amphibious Ship Precision Approach and Landing System). A number of military aircraft programs have funded plans in the implementation phase for JPALS employment. The Joint Strike Fighter (JSF) is fielding with a JPALS landing system as their primary means of getting aboard ship. DDG-1000 is fielding with JPALS only, no TACAN. This paper addresses the system´s requirements, current program status, schedule, and future plans to include technologies employed, stakeholders, and systems/technologies that JPALS will interface with and/or replace. It also answers the question: Why JPALS from the perspectives of aircraft operators, Theater Commanders, and the taxpayers.
NAVAIR Public Release 213 SPR.11-110 Distribution: Statement A-"Approved for public release; distribution is unlimited."

[Return to Program]

Session C1, Paper #6

Accounting for Tropospheric Anomalies in High Integrity and High Accuracy Positioning Applications
S. Khanafseh, M. Joerger, Illinois Institute of Technology; A. Von Engeln, EUMETSAT, Germany; B. Pervan, Illinois Institute of Technology

This paper describes and quantitatively evaluates integrity risk contributions of tropospheric anomalies in high accuracy differential satellite-based navigation systems. Two types of tropospheric phenomena are considered. First, we discuss severe weather fronts, which are already known to cause unusually large differential ranging errors. Second, the issue of tropospheric ducts is introduced and investigated using ten years of tropospheric data. A parametric tropospheric duct error model is defined and a threat space is established. It is shown that tropospheric ducts can produce differential ranging errors on the order of 50 cm. New dynamic error modeling and probabilistic bounding approaches are introduced to account for integrity risk resulting from both types of tropospheric anomalies. Performance evaluations are performed for an example aircraft shipboard landing application.
The troposphere is the lower part of the earth´s atmosphere, extending from earth surface to about 16 km altitude. It is made of electrically neutral gases that are not uniform in composition, including water vapor and dry gases. Refraction in the troposphere delays the transmission of satellite signals. The tropospheric delay consists of a largely predictable dry component, and of a wet component that varies with latitude, altitude, season, and weather condition but represents a much smaller fraction of the error. Therefore, the majority of the tropospheric delay can be removed by troposphere modeling (e.g., using a Modified Hopfield Model). Several models exist that describe the tropospheric delay under nominal conditions. Although some of these models exploit regularly updated weather parameters, they do not account for anomalous conditions.

The first part of this paper addresses severe weather fronts. A weather front is characterized by an abrupt change in refractivity (measured by variations in temperature, pressure, and relative humidity) over short horizontal distances. Published research reports experimental observations of tropospheric delays of up to 40 cm over a 5-km distance [1]. Further published work, with application to ground based augmentation systems (GBAS), assumed a ´weather-wall´ model to investigate the impact of high stationary fronts on differential ranging measurements (between a user and a local reference station) [2]. Non-nominal differential tropospheric errors were accounted for using a bounding tropospheric decorrelation parameter (i.e., a tropospheric delay gradient over user-to-reference separation distance). However, for applications that require the integrity of GBAS, but higher accuracy, a tighter upper bound on the spatial decorrelation parameter is needed. Also, in GBAS, the tropospheric decorrelation parameter is computed at specific ground stations. In this work, establishing a tighter upper bound is achieved by reconsidering the assumption of worst-case conditions on all satellites at all times (worst-case conditions are defined in [2] by worst-case weather-wall parameters, on the worst-case satellite at the worst-case elevation). In addition, we explore the possibility of establishing a bounding decorrelation parameter that is not location dependent.

The second part of the paper investigates tropospheric ducts. Under nominal conditions, pressure drops exponentially with height, and temperature decreases with altitude at an approximate rate of 1K/100m (over the first few kilometers above sea level). As a result, the computed refractivity gradient with respect to altitude is approximately -40 /km. However, this standard behavior does not apply under anomalous atmospheric conditions, and tropospheric ducts can be generated. Ducts tend to form when either temperature is increasing, or water vapor concentration is decreasing with height (due to climatologic mechanisms such as temperature inversion, evaporation ducts, air subsidence and air advection). Published literature shows that ducts appear at varying altitudes with likelihood of occurrence depending on location. Severe refractivity gradients (up to two orders of magnitude larger than nominal) were observed in localized layers of the troposphere at various altitudes and with different thicknesses.

In this work, ten years of data from the European Center for Medium-range Weather Forecasts (ECMWF) was analyzed. Duct thickness, altitude and refractivity gradient statistics were quantified and their probability of occurrence was established for a worldwide grid of 1 degree resolution. Using this data, it is shown that ducts can cause differential ranging measurement errors of up to 50 cm. In addition, the probability of occurrence of a duct is much higher than integrity and continuity risk requirements used in aviation applications (up to 90% in certain locations). Because ducts are not particularly rare and can cause non-negligible ranging errors, they are potential threats to navigation integrity for high accuracy applications.

The impact of ducts on differential ranging measurements is analyzed for a benchmark aviation application that requires tight integrity and accuracy requirements during final approach and landing phases. Differential measurements are influenced by portions of the duct located between the aircraft altitude and the reference receiver antenna height. Therefore, the ECMWF data is screened for ducts occurring at altitudes lower than 200 m and 500 m (below the aircraft decision height). This data is then used to define a threat space that describes the worst case parameters for tropospheric ducts. Within this more limited threat space, we determine that the differential ranging errors caused by ducts can reach up to 10 cm.

To ensure integrity in the presence of this anomalous source of error, various methods are considered, including inflating the measurement error standard deviation, deriving a robust dynamic model of the duct error profile (to be incorporated in the position estimation process), and bounding the positioning error in protection level and integrity risk computation. These approaches are further complicated in high-accuracy applications that require carrier phase cycle ambiguity resolution due to the nonlinear nature of the fixing process.

Availability analysis for the example shipboard landing application is carried out to compare the performance of these integrity risk mitigation approaches. The methods must be robust to varying atmospheric conditions to ensure navigation integrity, but they cannot be overly conservative or navigation availability will be adversely affected. Trade space parameters are identified and performance sensitivity is evaluated for multiple user locations.

[1] Huang, Jidong, van Graas, Frank, Cohenour, Curtis, "Characterization of Tropospheric Spatial Decorrelation Errors Over a 5-km Baseline", NAVIGATION, Vol. 55, No. 1, Spring 2008, pp. 39-53.

[2] van Graas, F., Zhu, Z., "Tropospheric Delay Threats for the Ground Based Augmentation System," Proceedings of the 2011 International Technical Meeting of The Institute of Navigation, San Diego, CA, January 2011, pp. 959-964.

[Return to Program]

Session C1, Paper #7

A Methodology to Elaborate Aircraft Localization Requirements for Airport Navigation
A. Guilloton, J-P. Arethens, Thales Avionics, France; C. Macabiau, A-C. Escher, ENAC, France; D. Koenig, GIPSA-LAB, France

The general increase in air traffic and the complexity of modern airport layouts have conducted to think about new technologies to assist pilots during maneuvers on airport surface. In compliance with ICAO recommendations, the safety of aircraft surface movement is currently ensured by the principle of "see and be seen" [1]. This basic principle becomes difficult to follow particularly in low visibility condition.
In the 90´s, ICAO develops the A-SMGCS (Advanced Surface Movement Guidance and Control System) concept to take into account new technologies and to cope with the air traffic increase and the airport complexity. A-SMGCS is a surface management system interfacing with the ATM (Air Traffic Management) system. It considers four connected main functions: Surveillance, Control, Routing and Guidance.

Surveillance function concentrates and displays to the ATC (Air Traffic Control) the position and the identification of all aircraft and vehicles on the airport surface. The aim of Control function is to detect conflicts (between moving aircraft and authorized vehicles on the movement area) and provides to each aircraft the constraints and alerts relative to its followed path. Routing function designates the most efficient route to follow for each aircraft or vehicle. Guidance function gives indications to the pilot to follow the assigned route.

The mid term new concepts concern guidance* and steering applications. These applications include airport navigation in low visibility condition (airport in LVP - Low Visibility Procedure - conditions).

A guidance application will provide the pilot indications for navigation over the airport surface in accordance with the ATC routing instructions. The aircraft position will be displayed on an airport layout as EMM (Electronic Moving Map) or HUD (Head Up Display). Steering application will allow the pilot to drive in all weather condition using steering indications relative to synthetic vision.

In the final stage of airport navigation, automatic control of the aircraft will combine the guidance and steering capability to enable an automatic control of the aircraft on the airport surface. It is also called autotaxi. Some studies were conducted about automatic control of the aircraft on the ground considering actuators but all make the assumption that the aircraft position is known without errors [2].

The common point of all these new functionalities is the use of the aircraft current position estimated by the onboard positioning system. This position is mandatory to generate the 2D map or 3D virtual reality displayed to the pilot and to compute guidance and steering information. An aircraft position is characterized by navigation performance parameters: accuracy, integrity, continuity and availability.

Therefore, it may be felt that this navigation performance will depend on the operational condition in which the system is used. Integrity requirements on position will not be the same if the system is used for situation awareness, or if it is used for navigation or control of the aircraft.

Precise requirements on aircraft localization service on the airport surface, in terms of accuracy, integrity, continuity and availability, have not been completely expressed by existing standards.

This paper proposes a methodology to derive these requirements based on FHA (Functional Hazard Assessment) method. FHA is a predictive technique and its goal is to explore the functional failure effects on parts of the system. The primary aim of FHA is to identify hazardous function failure conditions. FHA is recommended by the Aerospace Recommended Practice - ARP 4754 - to perform hazard identification and it is defined as one of the preliminary activities in the safety assessment process. The main difficulty to derive these requirements is the definition of the operational context relative to the forecast application within a global A-SMGCS vision.

The organization of the paper is the following. First section recalls the different operational phases on airport surface and the main characteristics of airports according to their classification. In the second part, the two considered applications, guidance and steering, are detailed. The third section introduces the FHA-method and a description of the risk classification according to the effects on aircraft and/or passengers. The last part presents the aircraft localization requirements, in terms of accuracy, integrity and continuity based on FHA-method, for the two considered applications. This results in the suggestion of 95% accuracy, integrity risk, HAL (Horizontal Alert Limit) and TTA (Time To Alert) requirements according to the visibility conditions for the different operations.

References [1] ICAO, Doc 9830 AN/452, Advanced Surface Movement Guidance and Control Systems (A-SMGCS) Manual, 2004 [2] F. Villaum‚, Contribution … la commande des systŠmes complexes: application … la l´automatisation du pilotage au sol des avions de transport, PhD thesis, University of Toulouse III, 2002

*Use of guidance terminology is as per A-SMGCS concept definition: it corresponds to the management of the navigation on the airport surface during the taxi operation.

[Return to Program]

Session C1, Paper #8

Virtual Receiver to Enhance GNSS-based Curved Landing Approaches
F. Kube, S. Schon, Leibniz Universitat Hannover, Germany; T. Feuerle, Technische Universitat Braunschweig, Germany

The civil aviation of the future should be ecological, economical and safe. The citizens in metropolises have on the one hand the need to travel efficiently between airports near the city. On the other hand they do not want to suffer from noise and emission of airplanes starting and landing at city near airports. Consequently, the future airplane should harmonize these needs. In order to fulfil these requirements new concepts for start- and landing approaches are mandatory. Curved approaches are one solution. These complex maneuvers can only be carried out with the help of satellite navigation systems. One major challenge of curved approaches is the obstruction of satellite signals due to shadowing effects by the aircraft itself. Existing GPS landing systems might therefore suffer degraded performance (accuracy and continuity) during curved approaches, especially with higher roll angles at low latitudes.
To avoid this problem and to fulfil the requirements of integrity, accuracy, continuity and availability for future GNSS landing systems the concept of the so called "virtual receiver" was developed in the framework of the research project "Buergernahes Flugzeug" (Metropolitan Aircraft). The main idea is to combine GNSS observations of several antennae, which are optimally installed on the airplane to compute one common position solution. Besides the flexibility of the approach, a further advantage is, that there is no necessity to develop new GNSS hardware. Existing antennae and receivers, which are already certified for aviation applications can immediately be used. Finally in future not only GPS, but also Galileo signals will contribute to determine an accurate position of the airplane. Until the full operability of Galileo pseudolites will be used at the research airport Braunschweig, which are installed in the vicinity of the airport for test and development purposes and send Galileo-like signals.

In a first step the mathematical model for the combination of observations from different antennae was developed. To reduce the influence of receiver noise and multipath, the C/A code observations were smoothed using carrier phase measurements. A smoothing time constant of 100s according to GBAS CAT I approaches was chosen. In particular the application of the smoothing filter leads to larger gaps because smoothed pseudoranges cannot be used for the navigation solution immediately after (re)acquisition of the satellite signal. No differential corrections were applied to the pseudorange observations, but the undifferenced observations are corrected for the satellite clock error, relativistic effects, tropospheric (Hopfield model) and ionospheric refraction (Klobuchar). The observations were weighted according to the elevation angle of the incoming satellite signal. Two solutions are computed: (i) the position of each antenna was individual computed with an epoch by epoch least squares solution, (ii) the position of a virtual receiver was computed combining the observations of multiple antennae and receivers. In the latter case for each receiver a new clock offset is set up in the combined adjustment and the lever arm w.r.t. an arbitrary chosen master antenna is corrected.

Before analysing real data, a simulation with a 3D model of an Airbus A320 was set up. With given attitude angles and coordinates of typical curved landing approaches, different positions for antennae on the airplane were compared with regard to the obstruction and continuity of the satellite signals. Using GPS satellite positions and varying the time and place for the landing approach it was shown, that it would be advantageous to mount at least two antennae sideways slanted on the fuselage in the front section of the airplane. This installation was intended to show the benefits and drawbacks of additional antenna installations at commercial airliners.

Using real data from a 30 minutes test flight with several maneuvers with the research aircraft Dornier Do 128-6 of the Institute of Flight Guidance (IFF) of the Technische Universitaet Braunschweig the developed algorithms were successfully tested. Two antennae were mounted on the fuselage, the two others on the wings of the airplane. The sampling rate was 1Hz. An elevation mask of 5 degree was applied.

First results show that the continuity is increased by combing observations of two antennae. The number of epochs with no position solution could not be completely eliminated due to a few flight maneuvers with extreme high roll (-40 to 55 degree) and pitch angles (-5 to 35 degree). However such attitude angles are not typical and expected for approaches of commercial airplanes. A further advantage is given by the higher number of observations, which are available per epoch. Besides an increased redundancy and capacity of fault detection, the HDOP as an indicator for the quality of the positioning result decreases by about 30%. Finally the virtual receiver enables a flexible choice and combination of observations from several antennae and GNSS for future GNSS-based curved start and landing approaches.

[Return to Program]

Session C1, Alternate #2

A New Design of FOG-SINS/Doppler Radar/ Baro-Altimeter Integrated Navigation System for Helicopters
H. Bai, X. Xue, Nanjing University of Science and Technology, China

To complete their missions successfully, helicopters must have accurate, jam-resistant navigation systems that are able to operate effectively and independently in all mission scenarios, these conditions may require high speed, low-level flight over featureless, unknown terrain like cruising over a dense forest, jungle or rescuing people in cities around with tall buildings. Although GPS is the most common and accurate navigation system being used on helicopters, it is very easy to be interfered and its signal is usually very weak in those areas. Meanwhile, helicopters because of their low cost and large number,can not afford the more sophisticated costly navigation systems like high accuracy inertial navigation system(INS) used in large fixed-wing aircraft. In order to solve the problem that precision and stability of GPS or GPS/INS navigation system will be highly affected when helicopters need to operate in those remote areas, along with the technical development of Doppler radar and fiber optic gyroscope strapdown inertial navigation system(FOG-SINS), a new type of Doppler radar/FOG-SINS/Baro-meter integrated navigation system(DFBINS) for helicopters was designed, which has many design features that are of particular importance to helicopters, such as low cost, easily reconstructive, self-contained and no external aids required, thus it can be an ideal navigation choice for helicopters when GPS is not available in some areas. From 1980s and 1990s, simple Doppler radar dead-reckoning navigation systems(DNS) of medium accuracy(1%-2%) were developed, and were applied extensively to helicopters, like AN/ASN-128, which was developed by Kearsoft company, has a navigation accuracy of 1.3%(CEP) when supplied with heading information from the helicopter´s magnetic compass with an accuracy of one degree(one sigma). Unlike DNS developed in the past, we used a low weight strapdown Doppler radar developed in recent years and a middle accuracy fiber optic gyroscope inertial measurement unit(FOG-IMU) which has a 0.2 degree per hour Bias stability performance, combined with a Baro-meter to build a new type of autonomous integrated navigation system for helicopters. These primary sensors were connected to an embeded navigation computer via interfaces designed by FPGA and DSP, FOG-IMU was interfaced via a RS422 serial data bus while Doppler radar and Air Data System(ADS) were interfaced via ARINC429 data buses, a dual MIL-STD-1553B data bus was also used to provide all data to external control and display systems of helicopters. Based on the hardware architecture, this paper systematically analysed this integrated navigation sytem and its related questions, three different models named loosely coupled, tightly coupled and ultra tightly coupled DFBINS models were introduced. In loosely coupled model, FOG-SINS was assisted by transfered velocity value from Doppler radar, while in tightly coupled model, both FOG-SINS and Doppler radar can be modified by each other, in order to reduce the computation complexity, a useful reduced tightly coupled DFBINS digital model was also provided in this model. Although Doppler radar is independent and can detect the body referenced 3-axile velocities of helicopters accurately, it may be invalidated when helicopters are flying over river, sea or desert due to the effect which is called the mirror effect, so an innovative ultra coupled model was introduced that using four named virtual beams from FOG-INS to assist Doppler radar when mirror effect appears. To validate the rationality of this model, related digital simulation experiments were carried and the results were showed in a table to compare the navigation results when using different accuracy inertial measurement units. Finally, a series of long endurance vehicle tests were conducted on both closed loop and open loop highway. The excellent results of simulation and vechicle tests indicate that the designed system is effective, it has a navigation accuracy of 0.1%~0.15%(CEP), which can well satisfy the requirement of navigation system for helicopters when GPS signal is not available in some remote areas. Due to the fact that GPS signals are denied in some remote areas and the low cost, high accuracy requirements of navigation system for helicopters, this paper designed a new type of DFBINS and the related integration models were introduced and analysed in detail. Correlative hardware, data buses and interfaces of the designed system were also introduced. Finally, in order to validate our designed system, a series of digital simulations and vehicle tests were conducted on both closed loop and open loop highway. Excellent results of these experiments show that our system is effective, accurate, it can be an ideal navigation choice for helicopters when satellite signal is not available in some area.

[Return to Program]