Device for aircraft ground collision avoidance with turn prediction

The invention relates to the general field of aids to navigation and aviation safety.

In the patent specification EP-A-0 565 399, the Applicant has proposed a device comprising:<ul id="ul0001" list-style="dash" compact="compact"><li>an input for receiving status indications, suitable for predicting the trajectory of the aircraft,</li><li>a working memory, suitable for containing a temporary local map, from which an envelope in the terrain altitude can be defined, in the area where the aircraft is operating,</li><li>first trajectory prediction means, for establishing, from said status indications and chosen flight parameters, a vertical limit trajectory of the aircraft (in the vertical "plane"),</li><li>first comparison means for making a first comparison between this vertical limit trajectory and the corresponding part of the envelope at altitude, to determine a risk of collision with the ground, and</li><li>alarm means for establishing at least one alarm state, as a function of the result of the first comparison.</li></ul>

This solution has undeniable advantages, in that it provides an entirely on-board anti-collision device, and above all capable of taking into account future situations of the aircraft.

However, it turned out that certain places, in particular the surroundings of certain airports, have a topology which is likely to always cause an alarm, within the framework of standard approach maneuvers. Although it is conceivable to admit such cases, exceptionally, this type of exception is harmful: indeed, an alarm system should in principle always intervene wisely; otherwise, there is a risk that an alarm will be ignored, at a time when it must on the contrary be imperatively taken into account.

This problem, known for a long time, has never been completely solved in known anti-collision systems.

The present invention comes precisely to offer a solution, starting from the aforementioned device, described in EP-A-0 565 399.

According to a general aspect of the invention, the device further comprises:<ul id="ul0002" list-style="dash" compact="compact"><li>second prediction means for establishing, from said status indications and chosen flight parameters, a lateral limit trajectory of the aircraft, on one side of the vertical of the flight trajectory,</li><li>second comparison means for making a second comparison between this lateral limit trajectory and the corresponding part of the envelope at altitude,</li></ul>    and the alarm means are arranged to modify at least one of the alarm states relating to the first comparison, if the second comparison means do not indicate a risk of collision with the ground.

Although such a device, operating only on one side, already provides progress, it is considered preferable to work on both sides. To this end, the system is supplemented by:<ul id="ul0003" list-style="dash" compact="compact"><li>third prediction means for establishing, from said status indications and chosen flight parameters, another lateral limit trajectory of the aircraft, on the other side of the vertical,</li><li>third comparison means for making a third comparison between this other lateral limit trajectory and the corresponding part of the envelope at altitude,</li></ul>    and the alarm means are arranged to inhibit the alarm state relating to the first comparison which indicates the maximum risk, at least in the case where both the second and the third comparison means do not indicate any risk lateral collision with the ground.

Other characteristics and advantages of the invention will appear on examining the detailed description below, and the appended drawings, in which:<ul id="ul0004" list-style="dash"><li>Figure 1 is the very general block diagram of the prior device, described in EP-A-0 565 399;</li><li>Figure 2 is a more detailed block diagram of the same prior device;</li><li>Figure 3 is an even more detailed, but partial block diagram of the same prior device;</li><li>FIGS. 4A and 4B are two other detailed diagrams, which, taken together (taking into account their overlap), define a particular embodiment of the same prior device;</li><li>FIG. 5 (FIG. 18 of EP-A-0 565 399) illustrates an example of a known anti-collision protection curve;</li><li>FIG. 6 (FIG. 21 of EP-A-0 565 399) is an example of a known avoidance trajectory;</li><li>Figure 7 is a very general block diagram of a device according to the present invention;</li><li>FIGS. 8A and 8B more completely illustrate an avoidance maneuver in the vertical plane;</li><li>FIGS. 9A to 9D illustrate an avoidance maneuver when cornering, in different situations; and</li><li>FIG. 10 is a diagram defining anti-collision procedures according to the present invention.</li></ul>

The accompanying drawings are essentially of a certain nature, and therefore form an integral part of the present description. They can not only serve to make it better understood, but also contribute to the definition of the invention, if necessary.

In addition, and taking into account the technical nature of the material, the descriptive content of EP-A-0 565 399 is to be considered as fully incorporated into the present description. The same is true for the following document:<ul id="ul0005" list-style="dash" compact="compact"><li>Note from DASSAULT ELECTRONIQUE and ROCKWELL Avionics - Collins entitled "A New Approach to CFIT Prevention: GCAS GROUND COLLISION AVOIDANCE SYSTEM", publicly released from June 1995.</li></ul>

The detailed description below relates to the case of civil aircraft, it being noted that the invention can be applied to other categories of aircraft.

Reference is made to Figures 1 to 4B. The device described in EP-A-0 565 399 is essentially intended to be installed on board an aircraft. This comprises equipment 2 capable of supplying, in the form of electrical signals, indications of flight parameters (FIG. 3), such as:<ul id="ul0006" list-style="dash" compact="compact"><li>an inertial unit 20 or INU,</li><li>a radionavigation instrument, here a GPS 21 receiver, with its antenna,</li><li>a radio altimeter 22, with its antenna,</li></ul> or other on-board navigation sensors.

The inertial unit 20 supplies the components of the speed (V) and acceleration (GAMMA) vectors of the aircraft. We can deduce the associated characteristic angles: incidence, wander, slope, pitch, heading, heel, in particular. However, insofar as the inertial unit measures and / or uses some of these angles to determine the speed and acceleration vectors, it is preferable to collect directly the values ​​of the inertial unit for said angles, where we are is used for the implementation of the invention. These angular values ​​can be displayed and / or used at the control station.

For altitude, the inertial unit cooperates with a barometric altimeter (not shown), in a known manner.

The following notations will be defined:<ul id="ul0007" list-style="dash" compact="compact"><li>Zb is the barometric altitude given by the measurement of atmospheric pressure, and varies according to altitude and weather conditions,</li><li>Zi is the inertial altitude calculated by the double integration of the vertical acceleration measured by the accelerometers of the inertial unit (long-term variations),</li><li>Zbi is the baro-inertial altitude, i.e. Zb filtered by Zi (3rd order loop, for example),</li><li>Zc will be the calculated altitude (HRS + Zta), where HRS is the radio probe height given by the radio altimeter (s) of the aircraft (accuracy of a few meters), and Zta will be the altitude of the terrain under the aircraft given by the terrain file (defined below)</li><li>Zgps is the altitude provided by the GPS.</li></ul>

Most of the time, aircraft are equipped with a battery of inertial units, and decision logic taking into account all the indications from these units. For the implementation of the invention, the information can come from one or more central offices (as far as they are confirmed).

The radionavigation instrument 21 provides raw measurements of latitude L1, longitude G1, and altitude Z1 (= Zgps), refreshed at a rate p1 from a few seconds to a few minutes. By integration on the speed and acceleration vectors, the inertial unit 20 provides other measurements of latitude L0, longitude G0, and altitude Z0 (= Zbi), precise but drifting over time. A block 25 compares the two types of measurement, and validates the quantities L1, G1, Z1, if they are consistent with L0, G0, Z0. Such validation techniques are known. The validated L2, G2, Z2 measurements are available at rate p1. However, they are refined from the inertial unit at a rate p2 of approximately one second.

A block 28 extrapolates the information between the last instant of measurement by the instrument 21 and the current instant (this extrapolation is useful in particular in the event of a problem with the rate of supply of information, which may be too low).

The radio altimeter 22 delivers the height above the ground, denoted HRS.

Block 3 contains a field file, established in a manner which will be described below. Depending on the quantities L and G, a part of this file, known as the local map, is accessed and stored in a local memory 40 (FIG. 4A).

From this local map, and of the quantities L, G, Z as well as HRS, block 4 performs anti-collision calculations, preferably accompanied by terrain avoidance calculations.

In the presence of a risk of collision, an alarm (51) is emitted. A director of orders 53 can suggest an avoidance maneuver. This is for the control station.

The local map can also be used for generating a synthetic image (60), with its display device 55.

It will also be used for functions (7) for validating the navigation and for memorizing deviation information (or differences), denoted MDF.

These deviations are transferred to the ground from block 79 to block 10, then sorted and are used for the total or partial updating of the database 30.

The memory 70 retains the mapped overflown profile, according to the sequence of data L, G, Z. The memory 71 retains the real overflown profile, taken from the height HRS. The difference between these two profiles is calculated in 72.

More precisely (FIG. 4B), the difference is compared with two thresholds at 74 and 75, respectively. It is then analyzed (77), from where:<ul id="ul0008" list-style="dash" compact="compact"><li>actuation of a compliance indicator 57 confirming the proper functioning of the device;</li><li>the production of a signal VAL.NAV., to validate the calculations of block 4; and</li><li>according to other criteria, the storage of MDF information in a memory 79, in correspondence with the values ​​L and G for which they have appeared. This MDF information can then serve as a basis, after the flight, for the update carried out in the system 10 of FIG. 4A.</li></ul>

In practice, the analysis block 77 will make a correlation of the two inputs, but taking into account:<ul id="ul0009" list-style="dash" compact="compact"><li>the history (recent past) of the flight,</li><li>estimated confidence in the radio altimeter data (operating limits taking into account the attitude of the aircraft, self-monitoring),</li><li>the fleeting nature of crossing the threshold,</li><li>analysis of the result (bias or noise or other),</li><li>possibly adequate stored information.</li></ul>

For its part, the database management block 35 makes use of at least one index which can be stored in the mass memory 30, so as to allow identification and rapid access to the data contained in the BDT database.

The control station 90 also provides information, which includes an "input status word" indicating at least:<ul id="ul0010" list-style="dash" compact="compact"><li>whether the piloting is automatic or manual;</li><li>the characteristics (in particular of flight) of the aircraft.</li></ul>

One of the essential bases of EP-A-0 565 399 is the fact that the Applicant has perceived the possibility of storing on board an aircraft a terrain file likely to represent almost the entire land block, within the limit contour and resolution suitable for the needs of an aircraft.

The analysis of the instantaneous and predicted situation of the aircraft can then be summarized as a set of curve tests, capable of generating two types of alarms:<ul id="ul0011" list-style="dash" compact="compact"><li>a pre-alarm (or "alert") indicating a dangerous configuration in the medium term, and</li><li>an alarm indicating a configuration requiring immediate action by the pilot, since flight safety is at stake.</li></ul>

To this end, two surfaces (curves in space) are provided for protecting the aircraft with respect to the terrain, defined according to the same principle but with different parameters, and comprising (FIG. 5; FIG. 6, for one of these areas):<ul id="ul0012" list-style="dash" compact="compact"><li>a short-term CCT surface, mainly intended to avoid an accident. As soon as a point on the ground enters the surface or the upper envelope of the CCT surface, the pilot must intervene (alarm) by performing an avoidance maneuver.</li><li>a medium-term surface CMT, mainly intended to warn the pilot that the trajectory of his aircraft will encounter an obstacle if it continues as it is, and that he must consider an avoidance maneuver (pre-alarm).</li></ul>

These surfaces, which constitute an important element of the protection system, can be developed from numerous static and dynamic parameters of the aircraft, in particular:<ul id="ul0013" list-style="dash" compact="compact"><li>the pilot transfer function of the aircraft, that is to say its ability to maneuver,</li><li>delays TR₀ due to the reaction time of the pilot of the aircraft,</li><li>the horizontal speed V_h from the aircraft,</li><li>the rate of climb V_z from the aircraft,</li><li>the admissible load factor ng,</li><li>the planned safety height, and</li><li>the roll of the aircraft.</li></ul>

For the rest of the description, some definitions are necessary:<ul id="ul0014" list-style="dash" compact="compact"><li>the instantaneous axis of the trajectory of the aircraft is called, in a predefined manner, either the tangent to the instantaneous trajectory (direction of the instantaneous speed vector), or the axis of the past and predicted trajectory (if the aircraft is in turn, the axis is then curved), that is to say an axis (in principle intermediate) defined for example a weighted linear combination between the two preceding ones;</li><li>"vertical plane" is a surface (not necessarily planar) which contains the vertical passing through the aircraft and an instantaneous axis of the trajectory of the aircraft; with regard to the axis of the past and predicted trajectory, the "vertical plane" is a curved surface if the aircraft is in a turn; the maneuvers whose main component is in a vertical plane are called "vertical";</li><li>the horizontal plane passing a reference point of the aircraft (center of gravity for example) is called "horizontal plane" and the maneuvers whose main component is in a horizontal plane are qualified as "horizontal" or "lateral"; here again, the horizontal "plane" could be a surface curved in space, defined according to the trajectory of the aircraft;</li><li>one distinguishes, among the horizontal maneuvers, those which go to the left and those which go to the right of the predicted trajectory of the aircraft;</li><li>finally the words "vertical", as well as "horizontal" or "lateral" will also be used to qualify in particular the obstacles and the risks which can be encountered during the maneuvers.</li></ul>

The limit avoidance curve in the vertical plane will be defined by three sections (Figures 6 and 8A):<ul id="ul0015" list-style="dash" compact="compact"><li>from T₀ at T₁, the continuation of the trajectory as it is for a time equal to the delay RT0 = T₁ - T₀ (corresponding to a reaction time),</li><li>from T₁ at T₂, a transition period (SVRM1 in FIG. 8A) due to a possible reduction in the roll and to the change in the radius of curvature of the trajectory passing from infinity to the ascending radius RT,</li><li>from T₂ at T₃, the avoidance trajectory proper (SVRM2 in FIG. 8A), whose radius of curvature RT is directly a function of the square of the aircraft's linear speed, divided by the load factor actually applied, ie<maths id="math0001" num=""><math display="block"><mrow><msub><mrow><mtext>R</mtext></mrow><mrow><mtext>T</mtext></mrow></msub><msub><mrow><mtext> = (V</mtext></mrow><mrow><mtext>h</mtext></mrow></msub><msub><mrow><mtext>)2/</mtext></mrow><mrow><mtext>ng</mtext></mrow></msub></mrow></math><img file="EP0802469A1_D0001.tif" /></maths></li></ul>

The foregoing has, for the most part, been indicated in FIG. 6 of EP-A-0 565 399.

Compared to Figure 6, it is desirable (Figure 8A) to add a fourth section to the boundary curve:<ul id="ul0016" list-style="dash" compact="compact"><li>beyond T3, a straight line SVRM3 of slope linked to the characteristics (performance) of the aircraft.</li></ul>

FIG. 8B (top view) shows that in practice the limit avoidance curve is a surface in space. Thus, the limits VT20, VT5 and VRP of FIG. 8A become curvilinear segments, defined here in the horizontal planes passing through the points VT20, VT5 and VRP as illustrated in FIG. 8A.

FIGS. 8A and 8B repeat, by adding thereto, the elements of FIG. 6, in a treatment of a different nature:<ul id="ul0017" list-style="dash" compact="compact"><li>a trajectory corresponding to a standard avoidance maneuver in the vertical plane SVRM ("Standard Vertical Recovery Maneuver") is defined first;</li><li>by taking the instantaneous axis of the trajectory of the aircraft and / or according to the orientation of the predicted trajectory (or their linear combinations), one can slide on this axis the SVRM, to the point where it meets the land envelope;</li><li>we can then define a time or vertical reference point VRP ("Vertical Reference point"), which is the beginning of SVRM;</li><li>upstream of this point VRP on the predicted trajectory, two times VT5 and VT20 are defined, with for example VT5 = VRP - 5 seconds, and VT20 = VRP - 20 seconds;</li><li>a "vertical" pre-alarm and an alarm are then defined, respectively as soon as the aircraft passes point VT20, and as soon as it passes point VT5 (the alarm naturally "overwrites" the pre-alarm).</li></ul>

The difference in treatment is due to the fact that, as shown in FIG. 8A, the avoidance maneuver is defined with respect to the terrain T, rather than with respect to the present position of the aircraft.

The technical means which have just been defined give satisfaction in most of the situations encountered in practice.

In short, they make it possible to provide the pilot (s) of the aircraft with a "pull up" alarm signal, if the predicted trajectory suggests a certain risk towards screw the neighboring terrain overflown, so that the pilot can initiate an emergency maneuver to avoid this terrain with a minimum safety margin. The concept of minimum safety margin is understood both in terms of human reaction time and in distance vis-à-vis the terrain avoided. The expression "neighboring overflown terrain" takes into account not only the terrain directly encountered in the axis of the aircraft trajectory, but also its neighboring parts.

With this alarm signal is associated a prior signal (called "alert"), which intervenes a few seconds before, in order to warn the crew of the proximity of this potential risk.

However, there are cases where it is normal for the aircraft to exceed the end point to perform the standard avoidance maneuver. However, beyond this point, no "vertical" avoidance maneuver is no longer possible.

These cases correspond in particular to trajectories that must be carried out in a fairly steep mountain environment. These trajectories are initially directed towards a relief which rises much higher than the current altitude of the aircraft. Consequently, once the ultimate point VRP has been passed, the aircraft can no longer cross this relief with certainty by maneuvering in the vertical plane along the axis of the predicted trajectory. The trajectory continues with changes in orientation of the aircraft. These are encountered in particular during approaches to airports surrounded by reliefs, forming deep valleys.

One might think that, this type of problem having already been encountered with military aircraft, it is easy to transpose the means used by the latter, to apply them to anti-collision devices for civil aircraft. Practice has shown that this transposition is far from immediate, for many reasons, including the fact that strong accelerations are both much less accessible to the aircraft, and very quickly intolerable, even disabling, for passengers.

It is also possible to define for civil aircraft a standard lateral avoidance maneuver (SHRM), capable of being initiated urgently to avoid a risk of collision with the surrounding terrain with minimum safety margins (in reaction time and in lateral distances MH and in height MV with respect to the terrain avoided).

We will first describe a particular example, with reference to FIGS. 9A to 9D.

The actual SHRM avoidance path begins at an HRP point on the predicted path. Before this point, two anticipated points HT5 and HT20 are also provided, which are also located on the predicted trajectory, respectively at 5 and for example 20 seconds upstream of the point HRP, taking into account the present speed of the aircraft.

We first consider the case where the aircraft is in horizontal flight (Figures 9A and 9B):<ul id="ul0018" list-style="dash" compact="compact"><li>the first segment SHRM1 (of t₁ at t₂) consists of an accentuated turn, with for example a turn radius RR of the order of 4 kilometers (2 nautical miles), compatible with the performance of the aircraft,</li><li>the second segment SHRM2 (from t₂ at t₃) consists of continuing the steep turn, so as to return to the HRP point,</li><li>a third segment SHRM3 (of t₃ at t₄) may consist of a repetition of the first two segments (without turning), so as to return to the HRP point, with altitude gain, if necessary.</li></ul>

An actual trajectory of the aircraft could be HFPa (dashed line), because the pilot has room to maneuver to control an even more accentuated turn than that of the SHRM.

An axis line represents a trajectory SHRMx, which is the limit of trajectory SHRM obtained in the event of an instantaneous reaction of the pilot (as soon as the point HT20 is exceeded).

A constraint is fixed according to which all the trajectories comprised between SHRM and SHRMx must remain at a distance at least MH from the reliefs, horizontally, and at a distance MV from the reliefs vertically. MH can be chosen on the order of for example 1 kilometer (0.5 nautical miles); MV can be chosen in the range of, for example, 150 meters (450 feet).

In the case of FIGS. 9A and 9B, this amounts to considering all the points of the domain (plane, or curve) comprised between SHRM and SHRMx, in order to apply to them the conditions of distance on MH and MV. This can be carried out without difficulty by checking that no relief exceeds the corresponding planar domain from the local map taken from the on-board terrain file, the accuracy of which is in principle increased near airports and areas of approach trajectories. in a steep environment.

In cases where the aircraft is descending (FIG. 9C), or ascending (FIG. 9D), the phase SHRM1 is modified to allow it to recover a horizontal flight condition. It may be lengthened over time.

In all cases, the constraints of lateral distance and height apply over the entire SHRMx + SHRM domain; therefore, in the event of a climb, the vertical distance constraint applies from point HT20 (Figure 9D).

If the aircraft is in a turn, the predicted trajectory takes into account the initial situation of the aircraft.

More generally, the SHRM maneuver is defined as being a trajectory which:<ul id="ul0019" list-style="dash"><li>breaks down into:<ul id="ul0020" list-style="none" compact="compact"><li>. a continuation of the present trajectory taking into account the initial situation (speeds, roll, slope, ...) during an initial reaction time, which leads to an HRP point,</li><li>. a turn from an HRP reference point at the instant noted t₁ in FIG. 9A, at an inclination greater than 20 ° (with a roll rate of 15 ° / s for example), and<ul id="ul0021" list-style="none" compact="compact"><li>.. in the event of a downhill flight, with the cancellation of the vertical speed for a time t₂ (FIG. 9A), for example the first 90 degrees of the turn, so as to keep the aircraft level;</li></ul></li><li>. continuation of the flight in a turn so as to make a complete circle (or even several),<ul id="ul0022" list-style="none" compact="compact"><li>.. in the case of a level or uphill flight, at an altitude equal to or at least equal to the present altitude,</li><li>.. in the case of a descent flight, at an altitude at least equal to the present altitude minus the loss of altitude necessary to level off;</li></ul></li></ul></li><li>avoids the surrounding terrain for:<ul id="ul0023" list-style="none" compact="compact"><li>. a vertical margin (MV) at least equal to a predefined value (for example 450 feet) relative to the horizontal part of this trajectory (in the most unfavorable case)</li><li>. a horizontal margin (MH) at least equal to a predefined value (for example 0.5 NM) relative to the trajectory.</li></ul></li></ul>

These margins must take into account the initial reaction time.<ul id="ul0024" list-style="dash"><li>is compatible with the capabilities of the aircraft at the start of this maneuver,</li><li>in principle requires a more severe inclination than that which would be implemented for a standard turn carried out in the absence of risk of collision,</li><li>authorizes a prior reaction time to initialize this maneuver before an HRP reference point.</li></ul>

It is to highlight that :<ul id="ul0025" list-style="dash"><li>HRP is seen as the ultimate point from which to turn.</li><li>The purpose of this maneuver is to allow an aircraft to return to its present place deemed "safe" and to be able (if it deems it useful) to climb uphill to get out of this potentially dangerous situation.</li><li>The trajectory described corresponds to the ultimate trajectory for performing a horizontal avoidance maneuver.</li></ul>

Consequently, it must allow a pilot to be able to perform a horizontal avoidance maneuver, which, normally, will<ul id="ul0026" list-style="dash" compact="compact"><li>tighter than the SHRM (so with a tighter turn), and</li><li>will include a turn before HRP.</li></ul>

As in vertical, the processing is "inverted", that is to say that we first determine what is the limit SHRM (and SHRMx) set which meets the constraints of lateral and vertical distance relative to the relief of the local map.

This is done for the extrapolation of the present trajectory, and for a range of trajectories surrounding it, in the same way as that is illustrated in FIG. 8B for the SVRM.

From there, we know how to determine the points HT20 and HT5.

In fact, there will generally be two different possibilities, respectively to the right (r) and to the left (ℓ) of the present trajectory of the aircraft. This results in two points HRPr and HRP_ℓ. Where there is no need to distinguish the right and the left, we will call (i) one of the sides, and (j) the other. The corresponding notations are HRP_i and HRP_j.

For the implementation of these functions, the present invention provides means which are illustrated in FIG. 7, in a manner comparable to FIG. 1. The unit 4 is now subdivided into three chains, which include:<ul id="ul0027" list-style="dash"><li>first trajectory prediction means TPM1 referenced 411, followed by first comparison means in the vertical plane VCM1, referenced 421;</li><li>second trajectory prediction means TPM2 referenced 412, followed by second comparison means operating laterally, on one side of the vertical plane HCM2, denoted 422; and</li><li>third TPM3 trajectory prediction means referenced 413, followed by third comparison means HCM3, denoted 423, operating laterally, on the other side of the vertical plane.</li></ul>

From there, the alarm decision unit 5 operates in a manner which will now be described.

The instant present in the trajectory of the aircraft is designated by t. To determine whether the aircraft has passed point VT20 in FIG. 8, the sign of the expression is therefore determined: (t - VT20). It is specified that notations like VT20 designate both the instant provided at point VT20 upstream from the reference point, which is here VRP, and the point of the trajectory which corresponds to this time interval. The argument (t-VT20) is positive if the point VT20 is exceeded, negative otherwise.

To avoid weighing down the drawing in FIG. 10, which represents the decision mechanism implemented in the example described, several functions relating to expressions of this kind are used.

The result of the logic function "No Negative Transition", noted NT / (), is, for a numerical argument "Arg":<ul id="ul0028" list-style="dash" compact="compact"><li>YES (True) if the last zero crossing of Arg is towards positive values, or if Arg has always remained positive (since an adequate origin time);</li><li>NO (False) if the last zero crossing of Arg is towards negative values ​​or if Arg has always remained negative (since an adequate origin time).</li></ul>

The origin time can be the start of the flight, from which it is preferable to withdraw the takeoff proper, phase of flight on the runway and just after, by criteria such as non-contact of the landing gear, or re-entry of the train , or passage at 15 m high above the runway threshold, for example.

The NT () function is the logical opposite of the previous one, that is, the NT () function is false if NT / () is true, and vice versa. It is only used to simplify the writing of the conditions of figure 10.

In some parts of the mechanism in Figure 10, it is necessary to know which is the most recent of two events. An event will be the transition to positive values ​​(with the definition above) of an expression of the form (t - VT20).

FIG. 10 uses for this purpose, and for illustrative purposes only, an LNT () function, which tests which is the last event encountered among several. For example, the LNT function (HT20, VT5) answers "yes" if the last event encountered is the change to positive values ​​of the expression (t - VT5), and "no" if the last event encountered is the change to values expression positive (t - HT20).

To carry out this test, one way of doing this is to date, up to an appropriate level of temporal precision, all the transitions encountered over a recent period of the flight of the aircraft. This can moreover be an arrangement of a function such as NT (), which in this case gives the first logical result defined above, and, moreover, as a second result the instant at which the last positive transition was observed. , and, for example, time originates in the absence of any transition.

From there, the LNT () function is a simple comparison of the times of two events.

Other ways of doing this are possible, such as for example providing a flag which would remain true (yes) as long as the most recent event is, in the example LNT (HT20, VT5), the passage of (t - VT5) to a positive value.

Reference is now made to FIG. 10. The first test 1001 determines whether no passage from point VT20 has been encountered so far, that is to say no positive transition of the expression (t - VT20). If yes, the state obtained in 1003 is the state AL0, which we will see that it represents the absence of alarm.

In the opposite case (no), we pass to test 1101, which determines whether the point VT5 has been passed, that is to say whether or not the aircraft is in the presence of a "vertical risk". On the "yes" output of test 1101, the situation is that of a vertical prerisk, as illustrated in block 1103. In this case, test 1201 aims to determine whether on each of the left and right sides of the aircraft, we crossed the lateral prerisk point HT20_i. We denote by the index i the side concerned, i can take one of the left values ​​"ℓ" or right "r". If at least one of the HT20 points_i has not been crossed, we go to state 1203 or AL100 (we will see that it can be associated with the same message as state AL0 of no alarm.

Otherwise, the situation (block 1205) includes a vertical prerisk and two lateral pre-risks (or risks). A corresponding alarm AL12 is then issued in 1207.

We now return to test 1101. Its other "no" output means, as indicated in 1105, that a vertical risk be encountered, that is to say that the point VT5 is crossed. In this case, test 1301 determines whether no positive transition has yet been observed at the lateral points HT20, on one side (ℓ) and on the other (r) of the vertical flight plane of the aircraft. If there is no lateral prerisk, the state AL200 is taken in 1303, which may be the same as the state AL0.

Otherwise, we go to test 1305. (The function is NT / () and not NT () in this test 1305). It checks whether the point HT20 has been crossed on both sides "ℓ" and "r". considering test 1301, the answer "no" to test 1305 means that a lateral prerisk (perhaps a lateral risk) has been encountered on one and only one side of the vertical plane, as indicated in the status block 1401. This side is noted by the index j, j can take one of the values ​​(ℓ) for the left and (r) for the right.

Test 1403 then determines whether, on the side j considered, a lateral risk has been encountered, by the expression NT / (t-HT5_j). The "yes" output indicates that one is in lateral prerisk on one side only, and the alarm indicated is then AL210 (block 1405), with messages to which we will return later. Otherwise, there is a lateral risk from one side, which in 1407 produces the state AL220.

We now come back to the other output (yes) of test 1305. In this case (block 1501), the situation is that of a vertical risk and prerisks on both sides (which can be risks).

Test 1503 determines whether the two NT functions (t-HT5_ℓ) and NT (t-HT5_r) answer "yes", that is to say if you stay at the level of a prerisk on both sides. If yes, it is only a vertical risk and a prerisk on both sides (status block 1601). In this case, test 1603 determines whether the most recent transition encountered was that relating to the vertical plane, that is to say the vertical risk. If yes, an AL211A alarm is produced in block 1605. Otherwise, an AL211B alarm is produced in block 1607.

At the other end of test 1503 ("no"), we went to the lateral risk, on at least one side of the vertical plane. The 1701 test determines whether one has gone to the lateral risk on both sides of the vertical plane of the aircraft. If this condition is not verified, that is to say that there is a prerisk on one side and a risk on the other side (noted j), it is determined in test 1703 if the last event encountered was that of vertical risk, that is to say that relating to vertical risk. If yes, alarm AL221A is produced on unit 1705. Otherwise, it is alarm AL221B of unit 1707 which is generated.

Finally, for the other output of test 1701, there is a vertical risk and a bilateral risk, and the alarm, which corresponds in this case to the maximum risk, is designated by AL222 in block 1709 .

FIG. 10 thus made it possible to describe the general mechanism of transition of the different situations, without risk, prerisk, risk, and this for vertical, left lateral, right lateral.

The messages addressed to the pilot are susceptible to different general options. Preferred examples of these will now be described, with reference to Table I appended at the end of this description.

In the first column, Table 1 shows the symbols of the alarms in Figure 10. The three columns which follow recall the corresponding states, respectively in the axis, on the side at maximum risk, and on the other side. In the fifth column appear the messages provided according to the prior art.

The last three columns respectively define options 1, 2 and 3 for the messages displayed to the pilot according to an example of implementation of the invention. These messages are in English, as is customary in civil aviation. A French translation is provided here:<tables id="tabl0001" num="0001"><table frame="all"><tgroup cols="2" colsep="1" rowsep="0"><colspec colnum="1" colname="col1" colwidth="78.75mm" /><colspec colnum="2" colname="col2" colwidth="78.75mm" /><thead valign="top"><row rowsep="1"><entry namest="col1" nameend="col1" align="left">ENGLISH</entry><entry namest="col2" nameend="col2" align="left">FRENCH</entry></row></thead><tbody valign="top"><row><entry namest="col1" nameend="col1" align="left">Field deposit (ahead)</entry><entry namest="col2" nameend="col2" align="left">Field alert (front)</entry></row><row><entry namest="col1" nameend="col1" align="left">Pull Up</entry><entry namest="col2" nameend="col2" align="left">ascend</entry></row><row rowsep="1"><entry namest="col1" nameend="col1" align="left">Avoid terrain</entry><entry namest="col2" nameend="col2" align="left">Terrain avoidance</entry></row></tbody></tgroup></table></tables>

The alarm absence states AL0, AL100 and AL200 naturally do not give rise to any message.

The AL12 state corresponds to a vertical prerisk and a unilateral prerisk (at least). In this case, the three options provide that a message is sent to the pilot informing him of a risk with the terrain.

The AL210 state is that of a vertical risk and a unilateral prerisk, which also does not give rise to any message in options 1 and 2. Option 3 provides that the pilot is informed of a vertical prerisk by the indicated alarm.

The AL220 state is that of a vertical risk and a unilateral risk, with no pre-risk on the other side. In the three options, the message sent informs the pilot of the vertical risk.

The AL211A and AL211B states are that of a vertical risk and a bilateral prerisk. Option 2 makes no difference, and in both cases we see the vertical prerisk message. On the other hand, options 1 and 3 distinguish according to whether the most recent transition occurred at the level of vertical risk, in which case the pilot is sent a message inviting him to pitch up ("pull up"); in the opposite case where the most recent event is the bilateral prerisk, the pilot is sent a message inviting him to a lateral evasion ("avoid terrain").

AL221A and AL221B have in common the fact that it is a vertical risk, a unilateral risk, and a pre-risk on the other side. In both cases, options 2 and 3 stipulate that the pilot be sent a vertical pre-risk message ("terrain guarantee ahead"). Under option 1, an emergency prompt is issued to the pilot for:<ul id="ul0029" list-style="dash" compact="compact"><li>in state AL221A, make a vertical evasion, or</li><li>in state AL221B, make a lateral evasion.</li></ul>

Finally, the last state AL222 is that of a vertical risk and a bilateral risk. In this case, option 1 invites the pilot to pitch up urgently. Options 2 and 3 invite him to an emergency lateral evasion.

In the foregoing, the expression "lateral evasive" refers to the possibility of circling uphill to gain altitude in order to escape from dangerous reliefs situated in the axis of the predicted trajectory, as indicated above.

The choice between the different options is made a priori and will depend in particular on the following considerations:<ul id="ul0030" list-style="dash"><li>maneuvering capabilities of the aircraft,</li><li>definition of the main parameters of the anti-collision system, and in particular the effective values ​​of the times VT5, VT20, as well as HT5 and HT20. Incidentally, it should be noted that the times are not necessarily the same on both sides of the aircraft, although the identity of these times is currently considered to be largely preferable.</li><li>predefinition at the start of flight (s), taking into account special cases that the aircraft may encounter, in particular when approaching its destination or stopover airports.</li></ul>

Of course, neither the alarm messages, nor the conditions for applying them, as they have just been defined with reference to FIG. 10 and to Table I, are not limiting. The alarms can be detailed or supplemented more generally, for example as already described according to considerations on the most recent of the events encountered, or of several of the most recent events encountered, or on other conditions applicable to these events.

For example, in the AL220 state, it is possible to envisage issuing a vertical “pull up” alarm, instead of the “ground advance ahead” pre-alarm. It is possible, in other cases, to use the messages provided according to the prior art, but after having taken into consideration the additional conditions added according to the present invention.

The alarm messages themselves are not limiting.

In the examples given in Table I, the lateral avoidance invitation messages do not indicate which side the pilot should go to. It is of course conceivable to inform him more completely, by indicating to him in particular that evasion is impossible on the left ("left side terrain unclear"), or on the right (right side terrain unclear ") by decomposition for example, of AL100 status.

In addition, it is also conceivable to add a level of "pre-pre-risk", given to the pilot for information ("advisory").<tables id="tabl0002" num="0002"><img file="EP0802469A1_D0002.tif" /></tables>