Dane wiarygodności SBAS w systemach e-nawigacji SBAS integrity data in e-navigation system

Pełen tekst

(1)PRACE NAUKOWE POLITECHNIKI WARSZAWSKIEJ z. 113. Transport. 2016. " 6!

(2) Akademia Morska w Szczecinie, W Nawigacyjny. SBAS INTEGRITY DATA IN E-NAVIGATION SYSTEMS The manuscript delivered: March 2016. Summary: The paper presents the usable form of SBAS integrity data presentation in e-navigation systems. The usability testing of ECDIS display augmented with innovative SBAS integrity data tool was based on the IMO’s Guidelines on Software Quality Assurance and Human-centred Design for e-Navigation. By incorporating the eye tracking techniques into the procedure it was possible to measure the visual attention distribution and the cognitive workload. The applied usability evaluation method proved that the seafarers were able to successfully perform primary operations of the systems reasonably upgraded with SBAS integrity tool and raise their situation awareness, regardless of the type and specifications of the system and their knowledge and experience with the system. Keywords: SBAS, ECDIS, usability testing. 1. INTRODUCTION E-navigation systems are expected to enhance safety of navigation. With each new generation of navigational equipment new features and extended functionality are incorporated. However, additional e-navigation functions can make it more difficult to understand systems’ primary information and may hamper the operation of the primary functions of the system with poor usability. For example, additional information in an Electronic Chart Display and Information System (ECDIS) may impede the route monitoring function if it is poorly presented in a raw alphanumeric form. Therefore, the new features presentation should be designed in a way to ensure that the seafarers are able to successfully perform primary operations of upgraded systems, regardless of the type and specifications of the system and users’ knowledge and experience with the system. A proposal of the Marine Vessel Protection Area (of the acronym MVPA) as the means of satellite based augmentation system’s (SBAS) integrity data presentation in an electronic chart system’s (ECS) has been developed in Maritime University of Szczecin as the result of the EMPONA Project financed by ESA [7]. The MVPA model takes into account several factors, influencing the presented figure’s shape and dimensions. These include Global Navigation Satellite Systems (GNSS) signal aspects (measurements’ errors depend on geometry of visible satellites and the signal propagation), the ship’s size, the ship’s heading and its estimated accuracy, position of the GNSS/SBAS antenna relative to the.

(3) 496.

(4) C . ship’s hull, and SBAS integrity data. This graphical representation of the marine ship’s position and protection level of the electronic position/course fixing equipment equivalent to the horizontal protection level (HPL) used in aviation has been tested in accordance to the IMO’s Guidelines on Software Quality Assurance and Human-centred Design for e-Navigation [4] and is currently analyzed for its potential benefits on navigation safety in different navigational tasks.. 2. SBAS UNCERTAINTY ELLIPSE DETERMINATION Satellite-based augmentation systems (SBAS), such as the European Geostationary Navigation Overlay Service (EGNOS) system, provide ranging signals transmitted by GEO satellites, wide area differential corrections and additional parameters aimed to guarantee the integrity of the GNSS. Integrity monitoring according to IMO [3] is the process of determining whether the system performance (or individual observations) allow use for navigation purposes. Overall GNSS system integrity is described by three parameters: the threshold value or alert limit (AL), the time to alarm (TAL) and the integrity risk (IR). The output of integrity monitoring is that individual (erroneous) observations or the overall GNSS system can or cannot be used for navigation or, in other words, alerting the user that he can experience a position error larger than the fixed AL value. To enable the user to estimate the position error, or its equivalent protection level (PL), the SBAS integrity data consist of estimations of each satellite ranging errors. These data have been successfully used for calculations of instantaneous point positioning PL which is the standard in aviation [2,6]. For example, in case of the EGNOS system, this concept is based on the broadcast of differential GPS / GLONASS corrections in message types MT1-5,7,9,17-18,24-26 and corresponding integrity data in MT2-6,10,24,26-28 [1,6,8]. The input quantities derived from the GNSS and SBAS messages for the integrity algorithm on the user side are: 1) The geometry between GNSS satellites and user derived position from observations of the GNSS satellites (the geometry matrix G of size n×4):. G. ª cos el1 sin Az1 « cos el sin Az 2 2 « « « ¬ cos el n sin Az n. cos el1 cos Az1 cos el 2 cos Az 2 cos el n cos Az n. sin el1 1 º sin el 2 1 »» » » sin el n 1 ¼. (1). Where: eli and Azi are the elevation and azimuth angles between the receiver antenna and the ith satellite (i=1,2,...,n), and n is the number of visible satellites, respectively.. 2) The weight matrix W built under assumption of uncorrelated, SBAS corrected, measurements characterized by the inverse variances of the distances to the observed satellites:.

(5) SBAS Integrity Data in e-Navigation Systems. W. ª 1 «V 2 « 1 « 0 « « « « 0 «¬. 1. V 22 0. º 0 » » 0 » » »» 1 » V n2 »¼ . 0. . 497. (2). V i2 V i2, flt V i2,UIRE V i2,tropo V i2,mr. (3). where, in (3): 2i,flt is the estimated variance for the residual error associated to user differential range error i,UDRE, which can be calculated per analogy to the model of [6] [m2], 2i,UIRE is the estimated variance for the slant range ionospheric error associated to grid ionospheric vertical error i,GIVE, which can be calculated per analogy to the model of [6] [m2], 2i,tropo is the estimated variance for the residual tropospheric error, which can be calculated per analogy to the model of [6] [m2], 2 i,mr is the estimated variance of marine (shipborne) receiver error depending on the receiver’s properties, and site-specific GNSS signal propagation effects like multipath, which must be locally evaluated (this alone variance cannot be derived from the SBAS message) [m2].. Based on (1) and (2) the covariance matrix can be found: ª s E2 « « s EN «s « EU ¬« s ET. s EN. s EU. s N2 s NU. s NU sU2. s ET º » s NT » sUT » » sT2 ¼». G WG

(6) T. 1. (4). s NT sUT where: s2E is the variance of the receiver’s antenna Easting measurement in the local reference frame centered on the GNSS antenna (East, North, Up, ENU) [m2]; s2N is the variance of the receiver’s antenna Northing measurement in the local reference frame (ENU) [m2]; s2U is the variance of the receiver’s antenna vertical measurement [m2]; is the variance of the receiver’s time correction measurement multiplied by the speed of light s2T [m2]; and, finally, the mixed terms (e.g. sEN etc.) are the co-variances of the respective measurements [m2].. The “elliptical” assessment of the SBAS point positioning user integrity can be given as a protection ellipse (PE), which is specified by 4 parameters, i.e.: 1) semi-major axis of the estimated position error ellipse, da [m]; 2) semi-minor axis of the error ellipse, db [m]; 3) orientation of the error ellipse, ; and 4) coverage factor, k, based on the confidence intervals. The integrity risk (or probability of Misleading Information, MI) is the probability that the user will experience a true position outside the protection ellipse, PEmr. It can be formulated as follows (see Fig. 1):.

(7) 498.

(8) C . PE. kd a ° ®kd b °¯. (5). where:. da. § s 2 s N2 s E2 s N2 ¨¨ E 2 © 2. 2. · 2 ¸ s EN ¸ ¹. (6). 2. db. atan2. § s2 s2 · sE2 sN2 2 ¨ E N ¸ sEN 2 2 © ¹. S 2. 1 atan 2 2 EN , E2 N2

(9) 2. (7). (8). is a clockwise angle of rotation from North either of the semi-major ellipse’s axis (if sE>sN) or of the semi-minor axis (if sN>sE); is the four-quadrant inverse tangent (arctangent) function of the real parts of two arguments (y, x) in Cartesian reference frame; and the formulae (6), (7) are derived from the square root of the eigenvalues of the covariance matrix (4) confined to:. CPA. ª sE2 « ¬ sEN. sEN º » sN2 ¼. and (8) is the direction of the eigenvector of (9).. Fig. 1. Elliptical representation of an SBAS point positioning protection area. (9).

(10) SBAS Integrity Data in e-Navigation Systems. 499. The “elliptical” presentation of a protection area provides navigator with the extra benefit coming from knowledge of N-E variances and their covariance resulting in changes of the ellipse’s orientation and shape. Still the elaborated concept treats the moving vessel as a point. That is why the new concept of Marine Vessel Protection Area (MVPA) has been developed for marine ECS and/or ECDIS systems where a vessel is shown as a 2dimensional spatial object (a model ship’s contour) [7].. 3. MARINE VESSEL PROTECTION AREA DETERMINATION BASED ON SBAS DATA The mathematical model describing how the vessel is presented on the ECS display can be expressed by two observation equations: x j,N. xN xGPS d j cos \ D j

(11). (10). y j,E. yE yGPS d j sin \ D j

(12). (11). x j2 y j2. (12). atan 2 x j , y j

(13). (13). where:. dj. Dj xj, yj. xGPS, yGPS xj,N, yj,E xN, yE dj j. S 2. are the calculated coordinates of consecutive j points of ship’s contour in the body-fixed reference frame (this is fixed to the marine vessel at the common reference point of aft perpendicular with positive x axis to fore, y axis to starboard, following the convention used in marine craft hydrodynamics and simulations – see fig. 2); are the coordinates (offsets from 0 at aft perpendicular) of EGNOS augmented GPS receiver antenna in the body-fixed reference frame; are the calculated coordinates of consecutive j points of ship’s contour in the local reference frame (ENU); are the recorded positions of EGNOS augmented GPS receiver antenna in the local reference frame (ENU); is the heading of marine vessel counted clockwise from North in the local reference frame (ENU); , is the jth distance between GPS antenna and jth point of ship’s contour; and is the jth angle between GPS antenna and jth point of ship’s contour counted clockwise from x-axis in the body-fixed reference frame..

(14) 500.

(15) C . Fig. 2. Model marine vessel’s contour consisting of 14 points in the body-fixed reference metric frame and GNSS antenna’s position in the fore part. The errors of parameters in equations (10) and (11) will propagate into the final MVPA according to the Gauss’s Error Propagation Law. The systematic errors of xGPS, yGPS, dj and j can be minimized to a negligible magnitude by a precise dimensional control. Therefore, only the propagation of other parameters’ errors (xN, yE, ) is taken into account in the MVPA determination according to the formula: C j , PA. J j CJ j. T. (14). where: Cj,PA is the covariance matrix of derived quantities:. C j , PA s2j,E s2j,N sj,EN Jj. s j , EN º » s 2j , N ¼». (15). is the Easting variance of consecutive j points of ship’s contour in the local reference frame (ENU) [m2]; is the Northing variance of consecutive j points of ship’s contour in the local reference frame (ENU) [m2]; is the covariance of j points respective coordinates [m2], is the Jacobian matrix (matrix of all first-order partial derivatives) of equations (10) and (11) excluding xGPS, yGPS due to their negligible errors: Jj. C. ª s 2j , E « ¬« s j , EN. ª1 0 sin(\ D j ) d j cos(\ D j ) º «0 1 cos(\ D ) d sin(\ D )» j j j ¼ ¬. (16). is the covariance matrix of observations:. C. ª s E2 « « s EN « 0 « ¬« 0. s EN s N2 0 0. 0 0º » 0 0» 0 0» » 0 s\2 »¼. (17). Where: s2 is the marine vessel heading variance, relevant to the marine-specific attitude / heading equipment (the typical values for marine gyros in transport vessels are in range 0.5°-1°); and: JjT is the transposed Jacobian matrix (16)..

(16) SBAS Integrity Data in e-Navigation Systems. 501. Estimated error of each jth ship’s contour point involves the errors of two jointly distributed variables of xj,N and yj,E coordinates. Thus, the positional error follows a bivariate normal distribution. Taking above into account, to fully describe the estimated error of each jth point, it is necessary to determine the orientation j and lengths of the semi-major dj,a and semi-minor axes dj,b of the jth error ellipses according to the formulas analogical to (6)–(8):. d j ,a. d j ,b. j. s 2j , E s 2j , N 2. s 2j , E s 2j , N 2. S 2. . 2. § s 2j E s 2j , N ¨ , ¨ 2 ©. · 2 ¸¸ s j , EN ¹. § s 2j , E s 2j , N ¨ ¨ 2 ©. · 2 ¸¸ s j , EN ¹. (18). 2. 1 atan 2 2 j , EN , 2j , E 2j , N

(17) \ 2. (19).

(18). (20. where j is a clockwise angle of rotation from the ship’s body-fixed x-axis either of the semi-major ellipse’s axis (if sj,E>sj,N) or the semi-minor axis (if sj,E<sj,N).. Each of the determined j ellipses can be further enlarged to the established confidence level by multiplying dj,a and dj,b by a coverage factor k, in analogy to the formula (5). Knowing the parameters (18) – (20) of uncertainty ellipses centred in j points of ship’s contour the next step is to find the extreme outer points of these ellipses in order to construct the MVPA. In order to do this, the maximum vertical values for a generalised (rotated) ellipse in the Cartesian reference frame fixed to the jth segment of ship’s contour (xcaxis between consecutive j and j+1 points), i.e., the upper bounding line or tangent of such an ellipse has to be calculated (see fig. 3).. Fig. 3. Construction of two tangent points (gray circles) to the jth error ellipse.

(19) 502.

(20) C . The algorithm is as follows: 1) The angle j of the line leading through j and j+1 points (j=1 is set as j+1 in case of max. j achieved) counted clockwise from x-axis in body-fixed reference frame is determined according to the formula: S atan 2 x j x j 1 , y j y j 1

(21) Ej (21) 2. 2) Tangent points of ellipses with lines of slope j are determined according to the formulas: S j ,c j E j (22) 2. Rj. sin E j º cos E j »¼. d 2j ,b sin 2 j ,c d 2j ,a cos 2 j ,c. (24). A j ,2. d 2j ,b cos 2 j ,c d 2j ,a sin 2 j ,c. (25). cos 2 j ,c sin 2 j ,c d 2j ,b d 2j ,a. ª « t j ,1 « « « t j ,1 «¬.

(22). A j ,1 A j , 2.

(23). 2. t j ,1 cos 2 j ,c sin 2 j ,c d 2j ,b d 2j ,a º » A j ,1 » t j ,1 cos 2 j ,c sin 2 j ,c d 2j ,b d 2j ,a » » A j ,1 »¼. T j ,3. ª y j ,tp «y ¬ j ,tn. (23). A j ,1. A j ,1d 2j ,a d 2j ,b. t j ,1. T j ,2. ªcos E j « sin E j ¬. x j ,tp º x j ,tn »¼.

(24). (27). (28). T j ,2 R. ª T j ,3 1,1

(25) y j «T 2,1

(26) y j ¬ j ,3. (26). T j ,3 1,2

(27) x j º T j ,3 2,2

(28) x j »¼. (29). where: j,c is the counter-clockwise angle of the jth ellipse rotation to x-axis in standard Cartesian 0xy reference frame, xj,tp, yj,tp, xj,tn, yj,tn are the coordinates of consecutive j tangent points in body-fixed reference frame, the extreme outer points are either xj,tp, yj,tp if j>0 or xj,tn, yj,tn if j´=]. The MVPA is constructed by linear connection of the resultant tangent points (see Fig. 4). This way the bounding spline representing the furthest points of ellipses in respect to the ship’s hull is found. In order to achieve acceptable level of ellipses’ areas coverage by the MVPA (or to minimize the linear spline approximation error) the number of tangent points can be increased adding extra tangent lines of slope angles in the range between j and j+1..

(29) SBAS Integrity Data in e-Navigation Systems. a). b). 503. c). Fig. 4. Examples of MVPA around the ship heading to 45° in the body-fixed reference metric frame and the antenna position in fore part. Prior to presentation of the vessel contour and MVPA in ECS the following requirements should be met: 1) Vessel’s contour points (xj, yj) need to be set from the shipyard data and uploaded to the navigation system. Depending on the size and shape (curvature) of the vessel hull the number of points describing vessel’s contour may vary. For most applications 14 to 16 points (j=1,2,...,14) have been assumed as sufficient. The origin of the vessel’s body-fixed coordinate frame is referred as CCRP (consistent common reference point) and usually in ECS-es it is fixed either to the craft’s aft perpendicular or to the navigator’s conning position. The reason for CCRP fixing to the conning position (frontcentre of the bridge) is to minimize the parallax error resulting from different placement of various electronic and optical position fixing devices in the ship. 2) GPS antenna offset should be measured with the highest possible accuracy in the prevailing circumstances, preferably with use of land survey techniques (e.g. a Total Station Theodolite, TST). The presented ship’s body-fixed MVPA can be updated periodically as new valid EGNOS integrity data and heading error estimation are received (Fig. 5)..

(30) 504.

(31) C . Fig. 5. SBAS integrity data as graphical MVPA in the ECS. 4. USABILITY TESTING. In March 2015, International Maritime Organization has issued the circular on “Guideline on Software Quality Assurance and Human-centred Design for e-Navigation”, officially introducing UT methods into future electronic equipment for marine navigation [4]. The appendix 3 of this guideline presents an UT process based on the ECDIS example as a one closely aligned with testing of future e-navigation systems. This UT example aligns with the integration and testing stage of a Human-Centred Design (HCD) process for evaluating the performance of essential tasks by competent users. Based on this recommendation the usability testing of ECDIS with SBAS integrity data presentation (MVPA) and without this data presentation has been performed. The framework for usability testing included 4 stages, is given in table 1 based on [5]. Table 1 Four stages used for ECDIS usability rating Stage 1 Stage 2 Stage 3. Achievement level Time Eye tracking measures. Stage 4. Scanpath analysis. Graded on scale 1 – 4 Total time required to accomplish given task All relevant data captured with the eye tracker Detailed analysis of participant’s visual attention distribution. In the stage 1 each task was rated in a scale from 1 to 4 according to IMO guidelines [4] by the experienced instructor. The stage 2 was concerned with the time required for accomplishing each task. In the stage 3 all relevant eye tracking measures are taken into ac-.

(32) SBAS Integrity Data in e-Navigation Systems. 505. count and evaluated. This data provides a basis for a cognitive evaluation and should help in identification of those tasks that are the most demanding and result in increased workload for the participant. In the study following measures were considered: total number of fixations per task, fixations frequency, fixations duration, location of fixations in a given area of the interface, gaze duration, number of blinks, duration of blinks. Via stage 4 analysis it was possible to identify all distractors and errors during a given task, by close examination of scanpaths. For example, during a task of measuring a bearing and a distance to a given landmark, the scanpath showed precisely where the participant’s attention was focused in any given moment (Fig. 6). On a typical scanpath fixations are represented as circles, where size of each circle corresponds to the fixation’s duration and colour intensity is used for ordering – more recent fixations are shown with vivid and opaque colour. Without the eye tracking technique, it is only possible to register participant’s actions and evaluate if those were either correct or incorrect. By incorporating the eye tracker into the study it was possible to register and evaluate participant’s attention distribution. This showed not only actions but also intentions of the participant and made it possible to recreate the search process on a cognitive level. The results were as follows: in the 24 scenarios simulated in the Full Mission Bridge Simulator (FMBS), each run by 5 experienced navigators with and without SBAS integrity data presentation, there were no statistically significant differences in ratings of stage 1, 2 and 3 (as defined in the table 1). The detailed analysis of stage 4 showed intensified visual attention to MVPA while close to safety margins (like safe isobaths or infrastructure) or in times of evident changes of its dimensions.. Fig. 6. Scanpath for the task: bearing and distance measurement.

(33) 506.

(34) C . 5. CONCLUSIONS. The Marine Vessel Protection Area (MVPA) concept has been developed for presentation of SBAS integrity data in marine ECS and/or ECDIS systems. By incorporating the eye tracking techniques into the procedure of usability testing based on the IMO’s Guidelines on Software Quality Assurance and Human-centred Design for e-Navigation [4] it was possible to measure the visual attention distribution and the cognitive workload. The applied usability evaluation method proved that the seafarers were able to perform primary operations of the systems upgraded and not upgraded with SBAS integrity tool on a similar cognitive workload level. The extra attention given by navigators to the MVPA during critical phases of manoeuvres in restricted waters seems to justify its presentation for the safety reasons. Presently all simulated runs are processed statistically in order to obtain probability density function of ship’s maximum distances from the centre of the waterway and accident probability calculation in given conditions. By comparing these measures for runs with and without SBAS integrity data presentation one should get another quantified measure of the MVPA benefit to navigator.. References 1. ESA, CNES: User Guide for EGNOS Application Developers, Ed. 1.1, 07/30/2009, 2009. 2. ICAO: Standards and Recommended Practices (SARPs), Volume 1 – Annex 10, Amendments 1-81, 2006. 3. IMO: Revised Maritime Policy and Requirements for a Future Global Navigation Satellite System (GNSS), Resolution A.915(22) from 22nd Session of the Assembly of International Maritime Organization, adopted on 29 November 2001, London, 2001. 4. IMO MSC.1/Circ.1512, Guideline on Software Quality Assurance and Human-Centred Design for e-navigation, IMO, London, 2015 5. $

(35) G]X

(36) C

(37) ]Y

(38) "

(39) !

(40)

(41) #

(42) ?-navigation systems, Technika Transportu Szynowego TTS 2015/12, sekcja nauka,

(43)

(44)

(45) #

(46) ,

(47) <=>] 6. RTCA, Inc. SC-159: RTCA DO-229D Minimum Operational Performance Standards (MOPS) for Global Positioning System / Satellite–Based Augmentation System Airborne Equipment, 2006, Change 1, 2013 7. C

(48) ]X

(49) B

(50) ]X

(51) G

(52) ']X

(53) "

(54) ]X

(55) +

(56) +]X

(57) G

(58) ']X

(59)

(60) $]Y

(61) '

(62) #

(63)

(64) Mission Ship Bridge Simulation Including EGNOS, Conference Proceedings of 2015 International Association of Institutes of Navigation World Congress, Prague, Czech Republic, 20–23 October 2015. 8. http://www.navipedia.net/index.php/The_EGNOS_SBAS_Message_Format_Explained. 1-*;$\02123;'BAS W SYSTEMACH E-NAWIGACJI Streszczenie: (

(65)

(66)

(67)

(68)

(69) %

(70)

(71)

(72)

(73) !

(74) ^! _

(75) GNSS pozyskiwanych z SBAS w systemach e-Nawigacji. ,

(76) %

(77)

(78) ?;@*+

(79)

(80)

(81) +G'+

(82)

(83) oparty o wytyczne IMO: „Guidelines on Software Quality Assurance and Human-centred Design for e-Navigation”.

(84)

(85)

(86)

(87)

(88)

(89) ! #

(90) %

(91)

(92)

(93)

(94)

(95) %

(96)

(97) !

(98) %

(99) wczego. Zastoso

(100)

(101)

(102)

(103) &X

(104) %

(105) !

(106)

(107)

(108)

(109) &

(110)

(111)

(112)

(113)

(114) %

(115)

(116)

(117) +G'+

(118)

(119)

(120)

(121)

(122)

(123) !

(124)

(125)

(126)

(127)

(128) &

(129) . 'kluczowe: SBAS, ECDIS,

(130) % .

(131)