Method and device for detecting hazards at the entrance to a roadside intervention zone

The RADAR-based detection system addresses false alarms and manual intervention issues in construction site safety by automating vehicle detection and merging vehicles, ensuring accurate and timely alerts for imminent dangers.

FR3140698B1Active Publication Date: 2024-11-29AXIMUM
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
FR2022010404
Authority / Receiving Office
FR · FR
Patent Type
Patents
Current Assignee / Owner
Filing Date
2022-10-11
Publication Date
2024-11-29
Estimated Expiration
2042-10-11

AI Technical Summary

Technical Problem

Existing construction site safety systems, such as the Aximum-Yellow system, suffer from false alarms due to non-dangerous situations and require manual site personnel intervention to draw danger zones, leading to lack of confidence and practicality.

Method used

A method using a RADAR sensor and computer system to detect vehicle speed, trajectory, and distance to trigger alerts only when vehicles are potentially dangerous, eliminating the need for manual zone drawing and reducing false alarms by merging vehicles like trailers and tractors into single entities.

Benefits of technology

Enhances construction site safety by accurately detecting imminent dangers, reducing false alarms, and automating the detection process without manual intervention, thereby improving system confidence and efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000018_0000
    Figure 00000018_0000
  • Figure 00000018_0001
    Figure 00000018_0001
Patent Text Reader

Abstract

The invention relates to a method for detecting danger at the entrance to an intervention zone (Z1) of individuals on a road (20), using a mobile construction vehicle (10) comprising a telemetry sensor and a computer. According to the invention, the method comprises the steps of: - detecting a first vehicle (40, 51) and the speed of said first vehicle, - determining the trajectory (T40, T51) of said first vehicle, - triggering an alert based on the determined trajectory, and only if the first vehicle is located at a distance from the telemetry sensor less than a first threshold which varies according to the speed of the first vehicle. Figure for the abstract: Fig. 2
Need to check novelty before this filing date? Find Prior Art

Description

Title of the invention: Method and device for detecting danger at the entrance to a road intervention zone Technical field of the invention

[0001] The present invention relates generally to the securing of areas where individuals work on roads, and in particular the securing of construction sites on motorways.

[0002] It relates more particularly to a method for detecting danger at the entrance to an intervention zone of individuals on a road.

[0003] It also relates to a construction machine equipped with a computer adapted to implement this method. State of the art

[0004] Road construction sites, and particularly on motorways, are particularly dangerous areas for personnel working on these sites.

[0005] In France, for example, there have been several dozen deaths in these areas due to vehicles not having properly recognized these construction zones.

[0006] This is the reason why the applicant developed several years ago, under the reference Aximum-Yellow, a system for securing construction site areas.

[0007] This system comprises a camera on a construction site marking trailer which is used to signal in a very visible manner to road users that they are arriving at a construction site area. This camera makes it possible to film vehicles arriving at the area.

[0008] This system also includes a computer tablet allowing one of the site personnel to view the images filmed by this camera and to draw a danger zone on the tablet screen. Then, as soon as a vehicle enters this zone, an alarm is triggered for all personnel working on the site, to warn them of a possible danger.

[0009] This system unfortunately has two major drawbacks. The first is that many alarms are issued in situations that are actually not very dangerous for personnel, which generates a lack of confidence in the system. The second drawback, which stems in particular from the first, is that this system requires one of the personnel to take a tablet to the site and then draw the construction site area on it, which in practice rarely turns out to be the case due to a lack of practicality and confidence in the system. Presentation of the invention

[0010] In order to overcome the aforementioned drawbacks of the state of the art, the present invention proposes to use, instead of a camera, a remote sensing sensor. such as a RADAR sensor, backed by a computer to process the data emitted by this remote sensor.

[0011] More particularly, the invention proposes a method for detecting danger at the entrance to an intervention zone on a road, by means of a mobile construction machine comprising a telemetry sensor and a computer, which method comprises steps of:

[0012] - detection of a first vehicle and the speed of said first vehicle,

[0013] - determination of the trajectory of said first vehicle,

[0014] - triggering an alert depending on the speed and trajectory of said first vehicle, only if the first vehicle is located at a distance from the telemetry sensor less than a first threshold which varies according to the speed of the first vehicle.

[0015] Thanks to the invention, the telemetry sensor used makes it possible to obtain more information on vehicles arriving at the worksite. Typically, such a sensor will make it possible to detect a car driving behind a truck (thanks to the reflection effects of waves, particularly on safety barriers) or a vehicle hidden by the leaves of a tree.

[0016] The data from the sensor are further processed by the computer in order to take into account the trajectory of the vehicle, which makes it possible not to trigger an alarm if the trajectory of the vehicle is oriented in a safe direction, in particular if it is clear that the vehicle is leaving the traffic lane on which the site personnel are working.

[0017] They are further processed to take into account the speed of the vehicle, which makes it possible to adjust the time at which the alarm will be triggered depending on the speed of the vehicle. Typically, the alarm may be triggered earlier if the vehicle is traveling fast. In addition, it may be possible to provide for not triggering an alarm if the speed of the vehicle is low (thus, no alarm will be triggered if the vehicle is in a traffic jam and is traveling erratically).

[0018] Preferably, prior to the step of detecting the first vehicle, a step of determining standard trajectories of vehicles traveling on said road is provided by means of the telemetry sensor, followed by a step of deducing the position of lanes of said road, the alert then being triggered only if the first vehicle is traveling or heading towards one of the lanes on which the intervention zone is located.

[0019] Thus, thanks to the remote sensing sensor used, it is possible to monitor vehicles traveling on the road by observing their average trajectories in order to deduce the position and orientation of the traffic lanes relative to the construction machine. Therefore, it is not necessary to install the construction machine in the axis of the road but it can be offset from it. This then makes it possible, even if the road is not straight, to clearly determine which lane each approaching vehicle is traveling on, in order to deduce whether it is potentially dangerous or not.

[0020] Preferably also, the first vehicle being initially detected at a first distance from the telemetry sensor, if a second vehicle is subsequently detected at a second distance from the telemetry sensor which is less than the first distance by at least 10%, and if (at this instant when the second vehicle and / or the trajectory of this second vehicle is detected), the distance between the first and second vehicles is less than a predetermined distance threshold, provision is made to merge the first and second vehicles.

[0021] The applicant has in fact found that when a truck is detected by a remote sensor, initially, only the truck's tractor is detected. Significantly later, the sensor detects a second element, namely the trailer. This late detection of two separate vehicles is explained by the fact that the space between the tractor and the trailer becomes detectable only when the truck approaches the remote sensing sensor. The disadvantage is then that the trajectories of these two elements, when considered separately, can lead to a poor understanding of the situation by the computer and therefore to a false detection of danger. Here, it is then proposed to merge these trailers and tractors into a single vehicle, which avoids any false detection.

[0022] Other advantageous and non-limiting characteristics of the method according to the invention, taken individually or in all technically possible combinations, are the following:

[0023] - said merging comprises the selection of the first vehicle and the deletion of the second vehicle ;

[0024] - said fusion comprises creating a merged vehicle at a position which is depending on the positions of the first and second vehicles, said merged vehicle then acting as the first vehicle;

[0025] - the first threshold is equal to the product of the speed of the first vehicle and a duration safety, said safety duration preferably being between 1 and 5 seconds;

[0026] - the trajectory of the first vehicle is determined based on positions successive movements of the first vehicle detected by the telemetry sensor;

[0027] - the relative position of the first vehicle with respect to the telemetry sensor being determined axially by an axial coordinate and laterally by a lateral coordinate, at regular intervals, an operation of filtering the lateral coordinates of the successively determined positions is provided; - the alert is triggered only if the first vehicle accelerates, remains at a constant speed, or decelerates with a negative acceleration greater than a negative threshold.

[0028] The invention also relates to a mobile construction machine carrying:

[0029] - a telemetry sensor,

[0030] - an alert transmission device, and

[0031] - a computer adapted to implement a danger detection method such as aforementioned, and to activate the alert transmission device if an alert is triggered.

[0032] Preferably, the telemetry sensor is a RADAR sensor.

[0033] Preferably, the telemetry sensor is adapted to detect vehicles at a range greater than 200 meters.

[0034] Preferably, the telemetry sensor has an opening angle of more than 120° to detect vehicles.

[0035] Preferably, the alert emission device is adapted to emit an audible alarm.

[0036] Of course, the various features, variants and embodiments of the invention may be combined with each other in various combinations to the extent that they are not incompatible or mutually exclusive. Detailed description of the invention

[0037] The description which follows with reference to the appended drawings, given as non-limiting examples, will make it clear what the invention consists of and how it can be implemented.

[0038] In the attached drawings:

[0039] [Fig-1] is a schematic perspective view of a construction machine conforming to the invention;

[0040] [Fig.2] is a schematic top view of a motorway on which there are the construction equipment of [Fig.l] and vehicles;

[0041] [Fig.3] is a diagram illustrating different stages of implementation of a method of detection in accordance with the invention.

[0042] In [Fig.l], a mobile construction machine 10 is shown.

[0043] In this figure, this construction machine 10 is in the form of a trailer. 11 supporting traffic signs 12. We speak of “FLR”, for “Lighted Drop Arrow”.

[0044] Alternatively, this device could take a different form. It could thus be without a road sign, or it could be an automobile and take the form of a van (we would then speak of an “FLU” for “Emergency Light Arrow”).

[0045] A construction machine 10 such as that illustrated in [Fig. 1] makes it possible to indicate a danger to road users, so that they slow down or change lanes.

[0046] Here, this construction machine 10 supports several traffic signs 12, some of which are inert while others are luminous and controlled. In practice, the latter are formed of orange lights making it possible to draw arrows indicating to road users in which direction they must turn.

[0047] In the context of the invention, this construction machine 10 further comprises a remote sensing sensor.

[0048] Such a sensor makes it possible to detect objects at a distance, by interpreting radiation emitted and reflected by these objects.

[0049] This is preferably a RADAR sensor 15.

[0050] This RADAR 15 sensor is designed here to detect vehicles at a distance of more than 200 meters.

[0051] In order to be able to detect vehicles arriving in a direction very inclined relative to it, this RADAR sensor 15 has a detection cone open, in a horizontal direction, at more than 120°. Therefore, the construction machine 10 can be installed both in the axis of the road (situation of [Fig.l]) and at 45 degrees relative to the axis of this road.

[0052] This RADAR sensor will preferably be designed to be able to simultaneously detect at least twenty separate vehicles. It will not operate on a zone sectorization system (it will be called “non-sectorized”). Finally, it will include means of communication, typically means adapted to communicate on a CAN type network.

[0053] The construction machine 10 also includes an alert transmission device 18.

[0054] This device could take many different forms. For example, it could be a warning light suitable for emitting flashing lights to warn the driver of the vehicle causing the detected danger and / or the personnel present on the site of a danger.

[0055] It will be considered here that this is a loudspeaker suitable for emitting an alarm for the attention of at least the site personnel, and capable of emitting for this purpose an audible signal of more than 100 dB, preferably reaching 120 dB. Thus this audible signal makes it possible to indicate to the site personnel the presence of an imminent danger.

[0056] As a further variant, the device could be in the form of a wireless communication unit, for example in the form of a Bluetooth or Wifi chip, making it possible to send an alert signal to portable devices equipping the site personnel (typically to badges equipped with an alarm system or to mobile phones).

[0057] The construction machine 10 also comprises a computer processing unit for the data from the RADAR sensor 15, which is adapted to control the alert transmission device 18. This unit will hereinafter be called the computer 19.

[0058] This calculator 19 comprises a processor, a memory and different input and output interfaces.

[0059] Thanks to its input interface, the computer 19 is adapted to receive data from the RADAR sensor 15. This could be raw data or data preprocessed by the RADAR sensor.

[0060] We will consider here that this is preprocessed data, making it possible to signal any detected vehicle by assigning it an identifier and a position relative to the RADAR sensor.

[0061] Thanks to its output interface, the computer 19 is adapted to control the alert transmission device 18.

[0062] Thanks to its memory, the computer stores a computer application, consisting of computer programs comprising instructions whose execution by the processor allows the computer to implement the danger detection method described below.

[0063] Before describing this method, one can, with reference to [Fig.2], give an illustrative example of a situation in which the construction machine 10 can be used.

[0064] In this example, the construction machine 10 is installed on a road with several traffic lanes. In our example, this is a motorway 20 comprising four traffic lanes separated in pairs by a central safety rail, and emergency stopping lanes located on either side of the four traffic lanes.

[0065] In the following, only one of these two pairs of traffic lanes and one of these emergency stopping lanes 21 will be considered. We will speak of the “right lane 22” to designate the traffic lane adjacent to the emergency stopping lane 21 and of the “left lane 23” to designate the other lane.

[0066] Still in this example, we will consider that an intervention zone ZI of individuals is located straddling the emergency lane 21 on the right-hand lane 22.

[0067] An intervention zone is defined here as an area of ​​the road in which there are objects to be removed (broken-down vehicle, debris, etc.) and / or personnel who need to intervene (the intervention can be carried out on foot or in a low-speed vehicle).

[0068] This is typically a construction site area (work zone, road marking line area, etc.), an accident zone, an area where debris is located (burst tire, etc.).

[0069] This intervention zone ZI will be considered fixed in the remainder of this presentation, but as a variant, it could be mobile (for example when it is a road marking line tracing zone).

[0070] In our example and generally, the construction machine 10 is placed on the right-hand lane 22, upstream of the construction site or at the entrance to the latter (considering the direction of movement of vehicles on this lane).

[0071] For example, it is placed 50 meters upstream of the intervention zone Zl.

[0072] In the situation illustrated in [Fig.2], a car 40 is traveling upstream of the machine construction site 10, on the left lane 23, and it is followed by a truck 50 which has started to move onto this same lane and then which presents a trajectory appearing to be heading towards construction site machine 10. In [Fig.2], this same truck 50' has been shown in dotted lines at a previous moment.

[0073] In the example illustrated in [Fig.2], motorway 20 is straight but as will become clear below, the invention will apply in the same way if the motorway had a bend at the intervention zone Z1 or upstream of it.

[0074] In [Fig. 3], the danger detection method according to the invention is illustrated.

[0075] This process could be started as soon as the construction machine 10 is stopped and the activation a danger detection function (for example by pressing a button on the machine).

[0076] Alternatively, this method could also be implemented when the construction machine 10 is traveling at low speed, on mobile construction sites.

[0077] Preferably, this method will be implemented when approaching a construction site. More precisely, it can be activated as soon as the traffic signs 12 are activated (as soon as the signal arrows are raised and / or illuminated). In other words, the activation of the method will be done automatically, when the traffic signs 12 are activated.

[0078] Here, in the seconds following its start, this process is implemented in a degraded manner. A self-calibration process is then implemented in parallel. Once the calibration is carried out, the process can be implemented optimally.

[0079] The self-calibration process consists, for the computer 19: - to determine the standard trajectories of vehicles traveling on highway 20 using the RADAR 15 sensor, then - to deduce the positions of traffic lanes.

[0080] In practice, the computer 19 records the successive positions of the vehicles detected by the RADAR sensor 15 so as to determine the positions of the central lines of the different traffic lanes 22, 23. To do this, it deletes the trajectories of the vehicles which differ too much from most of the other trajectories (for example the trajectories of the vehicles which change traffic lane), then it groups the similar trajectories and takes an average for each group in order to find the position of each traffic lane.

[0081] Thanks to this calibration, the construction machine's computer can determine the position of the traffic lanes relative to this machine. It will then be possible to detect in which traffic lane each vehicle is located, whether the road is straight or curved, and whatever the orientation of the construction machine 10 relative to the road.

[0082] At this stage, the computer 19 detects that the construction machine is on the right-hand traffic lane 22. It deduces from this that only the left-hand traffic lane 23 can be used safely by vehicles.

[0083] During the calibration process, the hazard detection method which is implemented in a degraded manner does not know the positions of the traffic lanes. In this case, the computer 19 simply issues an alert as soon as a vehicle has a trajectory which is oriented towards the construction machine 10 or more to the right (towards the emergency lane) and is at a reduced distance from the construction machine 10 (this distance depending on the speed of the vehicle, as will be described below).

[0084] When the calibration process is completed, the hazard detection method is implemented in a loop as illustrated in [Fig.3], at a frequency greater than 2 Hz, for example every 80 ms.

[0085] During a first step E0, the computer 19 receives data from the RADAR sensor 15 and then determines whether a vehicle is detected.

[0086] If this is not the case, the process is repeated every 80 ms.

[0087] If this is the case, the method continues in a step E2 during which the The computer determines whether it has enough successive position data associated with the same vehicle (identified by a unique identifier) ​​to determine its speed and trajectory.

[0088] For example, it can be considered that 10 to 30 successive positions of the vehicle must be known to obtain a good estimate of its speed and trajectory. Here, the number of 20 successive positions is required.

[0089] If the data is not sufficient, the instantaneous position of the vehicle detected is recorded in the memory of the computer (step E4) and the method is repeated from step E0.

[0090] If the number of data is sufficient, the computer implements, during a step E6, an algorithm for characterizing the detected vehicle.

[0091] This algorithm consists of determining the speed of the detected vehicle and its trajectory T40, T50, taking into account its successive positions detected every 80ms.

[0092] This algorithm also consists of determining the future trajectory of the detected vehicle. For this, one solution is to determine the equation of the curve or straight line of the trajectory T40, T50, taking into account the last 20 successive positions of the vehicle. A simple approximation calculation is used here to determine the equation of a curve, in polynomial form for example. A mathematical calculation based on the least squares method can also be used.

[0093] At this stage, if a vehicle has been detected, the computer 19 therefore has its last 20 successive positions, its speed, and its predicted trajectory.

[0094] In the following, to clearly illustrate how this method works, we will consider the situation illustrated in [Fig.2]. We will consider that the car 41 will have been characterized at a time ti (in terms of positions, speed and predicted trajectory) and that the truck 50' will have been characterized at a time t2 (where it is represented in dotted lines).

[0095] During the following step E8, called “collision risk detection step”, the computer 19 implements an algorithm making it possible to detect whether there is a risk that each detected vehicle will enter the intervention zone Z1.

[0096] To do this, it determines whether the predicted trajectory intersects the intervention zone Zl. More precisely, it determines whether this trajectory intersects the tracks on which the intervention zone Zl is located, at the level of the construction machine 10 or downstream of the latter.

[0097] In practice, the construction machine 10 having been able to determine the positions of the traffic lanes, the computer 19 detects whether: - vehicles are on its lane or further to the right (on the emergency lane) and if their trajectories indicate that they are not going to leave these lanes a priori, and if - vehicles are in the left lane 23 but are moving towards the right lane 22.

[0098] If this is not the case, the process is interrupted and then repeated from step E0.

[0099] Typically, in our example, the car following a trajectory centered on the left road 23, the process is interrupted for this car.

[0100] The truck initially presents a trajectory which leaves the right lane 22, the process can also be interrupted.

[0101] However, upon a subsequent repetition of this process, it is detected that the truck appears to be returning to the right lane 22.

[0102] In the event that a risk of danger is detected, the process continues.

[0103] To do this, the computer 19 determines whether the vehicle (here the truck) is close to the construction machine 10, taking into account its speed.

[0104] More precisely, the computer 19 determines whether the vehicle is likely to reach the construction machine 10 within a given time, for example within the next 3 seconds. To do this, it calculates the ratio of the distance between the construction machine and the detected vehicle to the speed of the detected machine, and it compares this ratio with the duration of 3 seconds.

[0105] If this ratio is greater than this duration, the process is interrupted and then repeated from step E0.

[0106] Otherwise, it could be planned to issue an alarm directly.

[0107] However, one objective within the scope of the present invention is to reduce the number of alarm emissions by controlling as precisely as possible that the danger is imminent.

[0108] For this, various other operations are applied to the detected vehicle.

[0109] Thus, the method continues in a step E10 during which it is planned to filter the last 20 successive positions of the detected vehicle in order to smooth them. Indeed, since the RADAR has a very long range and a high opening angle, the detected vehicle positions may be slightly erroneous.

[0110] An error of a few centimetres or decimetres along the vehicle's axis of movement has no effect on the process. On the other hand, a lateral error can lead to the vehicle's incorrect trajectory being detected and therefore a danger being wrongly detected.

[0111] To describe this step E10, we can consider for example a reference frame centered on the construction machine 10, whose abscissa axis is oriented towards the detected vehicle and whose ordinate axis is oriented at right angles to the latter.

[0112] This step will then consist of smoothing the lateral coordinates of the successive positions of the vehicle along this ordinate axis.

[0113] More precisely, if one of these lateral coordinates differs from that(those) of the previous and following positions by a difference greater than a threshold, this lateral coordinate may be replaced by an average of the lateral coordinates of the following and previous positions of the vehicle.

[0114] After this filtering step E10, a step E12 similar to the aforementioned step E8 for detecting a risk of collision is provided.

[0115] In practice, this step E12 is implemented in an identical manner to step E8, but this time considering the filtered coordinates of the positions of the detected vehicle.

[0116] This step helps to reduce cases of “false positives”, i.e. cases where a danger has been wrongly detected.

[0117] At the end of this step E12, if no further danger is detected, the process is interrupted and then repeated from step E0.

[0118] Otherwise, the method continues in a step E14 during which it is planned to determine whether the last vehicle detected is located near another vehicle detected previously.

[0119] To better understand this step El4, we can place ourselves at a time t3 where the truck 50 has advanced from its position in which it was represented by dotted lines to its position represented by solid lines. We will also consider that the tractor 51 of this truck had been detected previously, when it was at a distance DI from the construction machine 10.

[0120] The idea here is to check whether the new vehicle detected in step E0 does not correspond to the trailer 52 of the truck 50 detected previously.

[0121] Indeed, taking into account the space between the tractor 51 and the trailer 52 of the truck 50, there comes a time when the RADAR sensor 15 no longer detects a single vehicle but two separate vehicles. It then assigns a new identifier to the second vehicle, i.e. to the trailer 52.

[0122] To check whether this configuration is present, the computer 19 determines the smallest distance AD ​​between the new vehicle detected and each of the other vehicles detected.

[0123] It then compares this smallest distance AD ​​with a threshold, which is for example of the order of a meter.

[0124] If no other vehicle is present within this radius, the method continues with step E26 described below.

[0125] If a third-party vehicle is indeed present in this radius at time t3, there is indeed a suspicion that the last vehicle detected forms a truck trailer and that this third-party vehicle forms the tractor.

[0126] To verify this suspicion, step E14 may possibly be repeated several times, at several successive times.

[0127] But above all, it is planned to determine whether the new detected vehicle was detected well after the third-party vehicle. Indeed, as explained above, given the precision of the RADAR sensor and its significant range, the latter would not be able to simultaneously detect the trailer and the tractor of the truck as two separate entities. It is only able to do so at a later stage. To check that this is indeed the situation, it is therefore possible to rely on the times of detection of the third-party vehicle and the new detected vehicle.

[0128] For this, we can take into account a time difference between the two detection moments, or a distance difference.

[0129] Here, with reference to [Fig.2], the computer 19 records the distance DI which separated the third-party vehicle from the construction machine 10 at the time of its detection and the distance D2 which separates the new detected vehicle from the construction machine 10 at the time of its detection. Here, if the last detected vehicle was spotted by the RADAR sensor 15 after the third-party vehicle has traveled more than a certain distance threshold since its detection, it will be considered that the new detected vehicle is indeed a trailer and that the third-party vehicle is its tractor.

[0130] The distance threshold is for example of the order of a few tens of meters. It is for example equal to 50 or 80 meters.

[0131] In practice, if the tractor 51 has traveled more than 10% of the distance remaining to reach the construction machine 10 between the time t2 when it was detected and the time t3 when the trailer 52 was detected, it can be considered that these two vehicles form a single truck 50.

[0132] In summary, two conditions are necessary to consider that the new vehicle detected is a trailer: - the first condition is that it was detected late after the third-party vehicle, and - the second condition is that it is located near the third-party vehicle.

[0133] Therefore, when these two conditions are met, it is intended to merge the two detected vehicles into one.

[0134] For this, it is first of all planned not to take into account the last vehicle detected, by no longer considering its identifier during the following step E16.

[0135] Taking this new data into account, the computer 19 implements a step E16 identical to the aforementioned step E8, in order to check whether the tractor 51 of the truck constitutes a risk for the site personnel. The trailer 52 is therefore not taken into account.

[0136] If this is not the case, the method is interrupted and resumes at step E0.

[0137] Otherwise, during a step E18, it is planned to combine the two detected vehicles (the trailer and the tractor) by calculating the position of a “virtual vehicle” based on the positions of these two detected vehicles, for example by averaging them.

[0138] Taking this new data into account, during a step E20 identical to the aforementioned step E8, it is checked whether there is still a risk of collision due to this virtual vehicle.

[0139] If this is not the case, the process is interrupted and resumes at step E0.

[0140] Otherwise, it is planned during a step E22 to reapply the filter from step E10 on the lateral coordinates of the last 20 positions of this virtual vehicle. Then, during a step E24, identical to the aforementioned step E8, it is checked whether there is still a risk of collision.

[0141] If this is not the case, the method is interrupted and resumes at step E0.

[0142] Otherwise, during a step E26, considering that the risk of collision is proven, the computer 19 commands the triggering of the alarm to warn the site personnel of this risk of collision, so that they can take advantage of the remaining seconds to take shelter.

[0143] At this stage, it can be noted that the use of a RADAR sensor will be preferred since it will make it possible to detect vehicles hidden by trees or by other vehicles. However, this last situation is undoubtedly the most dangerous. It happens that several trucks follow each other closely. When the first truck spots the work zone, it naturally moves towards the left lane as soon as possible (when no more vehicles are overtaking it). The following truck then perceives the danger later. The greater the number of trucks, the greater the risk that a truck perceives the danger too late. Thanks to the invention, this situation can be clearly perceived by the computer, which can emit an alarm early enough to warn the site personnel.

[0144] The present invention is in no way limited to the embodiment described and shown, but those skilled in the art will be able to provide any variant in accordance with the invention.

[0145] Typically, it will be possible to provide that the computer is not on board the construction machine and that the latter is equipped with means of communication with the computer. Thus the computer could be formed by a remote computer server.

[0146] It will also be possible, according to other variants, to program the calculator in such a way as to add other functionalities to it.

[0147] Typically, the computer could be programmed to detect traffic jams. It could rely on the density of the vehicles detected and / or their number and / or their speeds. Then, when a traffic jam is detected, it could be planned to deactivate the alarms. For example, the computer could be programmed to exclude any alarm triggering when the average speed of all the vehicles detected is less than 50 km / h.

[0148] As a further variant, the computer could be programmed to trigger an alarm according to the trajectory of the vehicle detected as potentially dangerous, but also according to the acceleration of the latter. For example, if it is detected that every second, the vehicle detected as potentially dangerous brakes by more than 10 km / h (which means that the brake of this vehicle is properly applied), it will be possible not to trigger an alarm (of course, in a variant, it will be possible to deactivate the alarm only if the vehicle brakes properly and if its speed is below a threshold beyond which it may be considered that the vehicle will not be able to stop in time).

[0149] More generally, the alert may be triggered only if the detected vehicle accelerates or remains at a constant speed or decelerates slightly. In other words, the alarm will be deactivated if the acceleration experienced by the detected vehicle is negative and below a predetermined threshold (as soon as the vehicle stops braking, the alarm may be triggered again).

Claims

Claims

1. Method for detecting danger at the entrance to an intervention zone (Zl) of individuals on a road (20), by means of a mobile construction machine (10) comprising a telemetry sensor (15) and a computer (19), comprising steps of: - detecting a first vehicle (40, 51) and the speed of said first vehicle (40, 51), - determining the trajectory (T40, T51) of said first vehicle (40, 51), - triggering an alert as a function of the speed and trajectory (T40, T51) of said first vehicle (40, 51), only if the first vehicle (40, 51) is located at a distance from the telemetry sensor (15) less than a first threshold which varies as a function of the speed of the first vehicle (40, 51), characterized in that, the first vehicle (51) being initially detected at a first distance (Dl) from the telemetry sensor (15), telemetry (15),if a second vehicle (52) is subsequently detected at a second distance (D2) from the telemetry sensor (15) which is less than the first distance (D1) by at least 10%, and if the distance (AD) between the first and second vehicles (51, 52) is less than a predetermined distance threshold, provision is made to merge the first and second vehicles (51, 52).,

2. Detection method according to the preceding claim, in which, prior to the step of detecting the first vehicle (40, 51), a step of determining standard trajectories of vehicles traveling on said road (20) is provided by means of the telemetry sensor (15), followed by a step of deducing the position of lanes (21, 22, 23) of said road (20), and in which the alert is triggered only if the first vehicle (40, 51) is traveling or heading towards one of the lanes (21, 22) on which the intervention zone (Zl) is located.

3. A detection method according to claim 1 or 2, wherein said merging comprises selecting the first vehicle (51) and deleting the second vehicle (52).

4. A detection method according to claim 1 or 2, wherein said merging comprises creating a merged vehicle at a position that is a function of the positions of the first and second vehicles. (51, 52), said merged vehicle then acting as the first vehicle.

5. Detection method according to one of the preceding claims, in which the first threshold is equal to the product of the speed of the first vehicle (40, 51) and a safety duration, said safety duration preferably being between 1 and 5 seconds.

6. Detection method according to one of the preceding claims, in which the trajectory (T40, T51) of the first vehicle (40) is determined as a function of successive positions of the first vehicle (40) detected by the telemetry sensor (15).

7. Detection method according to the preceding claim, in which the relative position of the first vehicle (40, 51) with respect to the telemetry sensor (15) being determined axially by an axial coordinate and laterally by a lateral coordinate, an operation of filtering the lateral coordinates of the successively determined positions is provided, said filtering operation consisting of smoothing the lateral coordinates of the successive positions of the vehicle.

8. Detection method according to one of the preceding claims, in which the alert is triggered only if the first vehicle (40, 51) accelerates, remains at constant speed, or decelerates according to a negative acceleration greater than a negative threshold.

9. Mobile construction machine (10) carrying: - a telemetry sensor (15), - an alert transmission device (18), and - a computer (19), characterized in that the computer (19) is adapted to implement a danger detection method in accordance with one of the preceding claims and to activate the alert transmission device (18) if an alert is triggered.

10. Mobile construction machine (10) according to the preceding claim, in which the telemetry sensor (15) is a RADAR sensor.

11. Mobile construction machine (10) according to one of the two preceding claims, in which the telemetry sensor (15) is adapted to detect vehicles at a range greater than 200 meters.

12. Mobile construction machine (10) according to one of the three preceding claims, wherein the telemetry sensor (15) has an opening angle of more than 120° for detecting vehicles.

13. Mobile construction machine (10) according to one of the four preceding claims, in which the alert transmission device (18) is adapted to transmit an audible alarm.