Method for determining parking space type

Abstract

A method for determining the type of parking space for a motor vehicle comprises: a phase of detecting an object (P1), a phase of evaluating the orientation of the object relative to a road (P3), and a phase of determining the type of parking space based on the evaluated orientation (P4). The detection phase (P1) is implemented by a front camera of the vehicle.

Description

Technical Field

[0001] The present invention relates to the field of parking assistance for motor vehicles and more particularly to determining the type of parking space. Background Art

[0002] Many vehicles today are equipped with automatic parking systems. These systems make it possible to perform parking maneuvers without the involvement of the driver. In order to increase the convenience of the driver even further, the automatic parking systems are coupled with a device for detecting the vehicle's surroundings in order to determine whether there is an available parking space. Reference may be made, for example, to document EP 2 327 608, which describes such a system.

[0003] One disadvantage of this solution is that it does not determine the type of maneuver to be performed in order to park the vehicle. For example, the automatic parking system may perform a parallel parking maneuver to park the vehicle in several perpendicular parking spaces. This results in inconvenient parking of the vehicle.

[0004] To address this problem, the driver is typically asked to select the type of maneuver that the automated parking system should perform to park the vehicle in an available parking space. Thus, the problem of improving driver convenience is clearly not addressed.

[0005] Document EP 2 982 572 describes a device for determining the type of parking space by analyzing the ground markings defining the parking space. This solution is not completely satisfactory because it only works if the parking space is well defined by the ground markings and the ground markings have not been erased.

[0006] Another solution is to use ultrasonic sensors to determine the type of parking space. A disadvantage of this solution is that ultrasound detects objects at a very short distance. As a result, when the type of available parking space is determined, the parking space is often passed. Summary of the invention

[0007] In view of the above situation, an object of the present invention is to overcome the aforementioned disadvantages.

[0008] More specifically, it is an object of the invention to make it possible to determine the type of parking space sufficiently early and even in the absence of ground markings defining the parking space.

[0009] To this end, a method for determining the type of a parking space for a motor vehicle is proposed, the method comprising:

[0010] - The stage of detecting the target,

[0011] - a stage of evaluating the orientation of the object relative to the road, and

[0012] A phase of determining the type of parking space based on the assessed orientation.

[0013] According to a general feature of the method, the detection phase is implemented by a front camera of the vehicle.

[0014] The vehicle's front camera is particularly suitable for identifying objects corresponding to passenger cars or commercial vehicles. The type of parking space can then be determined based on the already parked vehicle. Therefore, there does not need to be ground markings that define the parking space. The object may be located at a relatively far distance in front of the vehicle. Therefore, the type of parking space can be determined before the vehicle drives past the available parking space.

[0015] In one implementation mode, in the detection phase, data associated with the target are determined, and the data associated with the target include a target continuous detection count signal, the position of the target, a variable representing the accuracy of the position of the target, the speed of the target, the orientation of the target, and a variable representing the accuracy of the orientation of the target. The method also has a filtering phase between the detection phase and the evaluation phase, and the filtering phase includes at least one filtering step selected from the following: an inaccurate position filtering step, an inaccurate orientation filtering step, a moving vehicle filtering step, and a phantom detection filtering step.

[0016] In the present application, the expression "continuous detection count signal" is understood to mean, in an embodiment, a signal that is incremented each time a target is detected and reinitialized when the target is no longer detected. Furthermore, in the present application, it is considered that the term "filtering" will be used according to its usual definition, that is, sorting is performed, and at the end of filtering, the filtered targets can be retained or rejected.

[0017] In the imprecise location filtering step, if the variable representing the accuracy of the location exceeds a location threshold, the target is rejected, and if the variable does not exceed the location threshold, the target is retained.

[0018] In the inaccurate orientation filtering step, the target is rejected if the variable representing the accuracy of the orientation exceeds an orientation threshold, and is retained if the variable does not exceed the orientation threshold.

[0019] These inaccurate position filtering steps and inaccurate orientation filtering steps make it possible to reject objects for which data detection is inaccurate. These steps are preferably performed only at the beginning of object detection and not again thereafter.

[0020] Preferably, the variables representing the accuracy of the respective position and orientation are the position standard deviation and the orientation standard deviation.

[0021] In the moving vehicle filtering step, if the speed of the target exceeds a speed threshold, the target is rejected, and if its speed does not exceed the speed threshold, the target is retained.

[0022] This filtering step makes it possible to reject objects corresponding to moving vehicles, thereby determining the type of parking space based solely on parked vehicles. Preferably, this step is applied as long as objects are detected, in order to eliminate objects that were initially retained but then began to move, such as vehicles that were initially stopped at a stop sign or traffic light.

[0023] In the ghost detection filtering step, if the target continuous detection count signal is less than the count threshold, the target is rejected, and if the signal is not less than the count threshold, the target is retained.

[0024] The ghost detection filtering step makes it possible to reject ghost detections made by the front camera. Preferably, this step is only performed on objects that are located in the field of view of the front camera of the vehicle, and not on other objects. Alternatively, this step is only performed at the beginning of object detection and not again thereafter.

[0025] According to an advantageous embodiment, during the detection phase, the position of the target is determined, the method further comprising a filtering phase between the detection phase and the evaluation phase, during which the target is rejected if the distance between the position of the target and the vehicle exceeds a distance threshold, and the target is retained if the distance does not exceed the distance threshold.

[0026] This filtering makes it possible to limit the length of the area in which the objects are considered. This is advantageous because it is more likely that the type of object far from the vehicle is different from the type of parking space close to the motor vehicle.

[0027] Advantageously, in the evaluation phase, the orientation of the road at the location of the object is determined and the orientation of the object relative to the road is evaluated based on the orientation of the road at the location of the object.

[0028] When the road has a curve, using the orientation of the road at the location of the object allows a simple evaluation of the orientation of the object in order to determine the type of parking space.

[0029] In one implementation mode, if the road has ground markings, the road orientation at the position of the target is determined by taking into account the ground markings; if the road does not have any ground markings and if the target is located behind the vehicle or at the same position as the vehicle, the past trajectory of the vehicle is determined and the road orientation at the position of the target is evaluated based on the past trajectory at the position of the target; and if the road does not have any ground markings and if the target is located in front of the vehicle, the possible trajectory of the vehicle is determined and the road orientation at the position of the target is evaluated based on the possible trajectory at the point closest to the target.

[0030] Thus, the road orientation can be determined even if the road does not have any ground markings and both in front of and behind the vehicle. When the road has ground markings, the ground markings are taken into account in order to determine the road orientation more accurately.

[0031] Preferably, when determining the possible trajectory of the vehicle, at least one scenario selected from the following scenarios is accepted:

[0032] - Speed ​​scenarios where the vehicle's speed remains constant;

[0033] - acceleration scenarios where the vehicle's acceleration remains constant,

[0034] - a steering wheel angle scenario in which the vehicle's steering wheel angle remains constant; and

[0035] - The yaw angle of the vehicle is kept constant in the yaw angle scenario.

[0036] Such scenarios constitute a compromise to allow simple and precise determination of possible trajectories.

[0037] In one implementation mode, in the detection phase, a plurality of targets are detected and in the determination phase, whatever the target is, at least one first term and one second term are associated with the target, the terms being selected from the following:

[0038] -Parallel parking type items,

[0039] -Vertical parking type items, and

[0040] - Herringbone parking type entry,

[0041] And then calculate the first sum of the first terms of each target and the second sum of the second terms of each target, and determine whether one of the first sum and the second sum is greater than the other sum, and preferably determine whether one of the first sum and the second sum is greater than the other sum plus a strictly positive offset term.

[0042] The type of parking space corresponding to the highest sum is most likely the type of parking space within the vehicle's area.The offset term allows the confidence in the determination of the type of parking space to be adjusted.

[0043] Preferably, whatever the objective is, at least one coefficient is associated with that objective and before calculating the sums, the terms associated with that objective are weighted by the coefficient associated with that objective.

[0044] For example, the coefficient may represent the distance of the target from the vehicle.

[0045] It is then taken into account that the further away an object is from the vehicle, the more likely it is that the object is a vehicle parked in a different type than the parking space located near the vehicle.

[0046] For example, the coefficient represents taking into account ground markings on the road to assess the orientation of the target relative to the road.

[0047] Targets whose orientations have been determined to be relatively imprecise may then be given lower weights.

[0048] For example, the coefficient may represent the lateral clearance between the target and the trajectory of the vehicle.

[0049] Then to a lesser extent, targets with a high risk of being located on another road or another row of parking spaces are considered.

[0050] According to another aspect, a computer program is proposed comprising code configured to implement the method as defined above when executed by a processor or an electronic control unit.

[0051] According to yet another aspect, a device for determining the type of parking space of a motor vehicle is proposed, the device being receivable onboard a motor vehicle, the device comprising:

[0052] a detection module, the detection module is used to detect the target, the detection module can be connected to the front camera of the motor vehicle,

[0053] an evaluation module capable of evaluating the orientation of the object relative to the road, and

[0054] A determination module which is able to determine the type of parking space by taking into account the orientation evaluated by the evaluation module. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Other objects, characteristics and advantages of the present invention will become apparent from a reading of the following description given by way of non-limiting example only and with reference to the accompanying drawings, in which:

[0056] - Figure 1Schematically illustrating a device according to one aspect of the invention,

[0057] - Figure 2 is a graph showing a mapping of parking space type term values ​​as a function of the orientation of the target, and

[0058] - Figure 3 is a flow chart illustrating a method according to another aspect of the invention. DETAILED DESCRIPTION

[0059] refer to Figure 1 , schematically showing a motor vehicle 2. The vehicle 2 is associated with a direct orthogonal vector basis 3 conventionally used in motor vehicle design. The orthogonal basis 7 comprises vectors vector and with vector and Vertical vector The vehicle 2 has an onboard computer 4 which is in information connection with a front camera 6 .

[0060] The camera 6 also forms part of an advanced driver assistance system of the vehicle 2. Such a system is also referred to as an advanced driver assistance system or the corresponding abbreviation ADAS. The function of the camera 6 is to collect a plurality of images of the surroundings of the vehicle 2. More specifically, the camera 6 detects images within a field of view located in front of the vehicle 2, the field of view having a maximum distance of 60 m and an aperture angle ranging around the vector In the present application, the term "image" is understood according to its usual definition in the field of optics, that is to say it indicates a representation of a solid object given by an optical system.

[0061] Camera 6 is able to reject objects corresponding to recreational vehicles and commercial vehicles in the collected images. In particular, camera 6 does not detect pedestrians, trucks, buses, motorcycles, bicycles, or even urban facilities that may be in its field of view as objects. For each object, camera 6 is able to determine data associated with the object, which includes:

[0062] - the position of the target relative to vehicle 2,

[0063] - the speed of the target,

[0064] - The orientation of the target relative to the vehicle 2. In the present application, the orientation of the target relative to the vehicle 2 corresponds to the longitudinal direction of the target and the vector The angle between.

[0065] The camera 6 is also configured to detect the presence of ground markings. In particular, the camera 6 detects ground markings defining spaces for traffic flow on the road and ground markings defining parking spaces. The ground markings defining spaces for traffic flow may include a center line, a hard shoulder line, or a line defining a lane. The ground markings defining parking spaces may include a line forming a "T" in the corner of the parking space. Advantageously, other cameras may be provided, such as cameras with a wide-angle field of view, which make it possible to build a 360° view around the vehicle. Such cameras are generally referred to as panoramic monitoring imaging systems or the corresponding abbreviation AVM.

[0066] The vehicle 2 has a device (not shown) for detecting available parking spaces. Such a device (known per se) can provide the onboard computer 4 with information about the existence of an available space for parking the vehicle 2.

[0067] The vehicle 2 has a device 8. The function of the device 8 is to determine the type of parking space. More specifically, the device 8 makes it possible to determine the type of parking space even when the parking space is not defined by ground markings. To this end, the vehicle 2 may also include a device (not shown) for determining the type of parking space based on ground markings. The combination of these two devices increases the robustness of the determination made when the parking space is defined by ground markings, and allows determination to be made when the space is not defined by ground markings. However, it is certainly not outside the scope of the present invention to envision that the device 8, which is not coupled to another device, determines the type of parking space. In the example considered, the device 8 is configured to determine the type of parking space corresponding to a parallel parking space, a perpendicular parking space, and a herringbone parking space. However, it is certainly not outside the scope of the present invention to envision other types of parking spaces.

[0068] The device 8 has a detection module 10 which is directly or indirectly connected to the onboard computer 4 and the camera 6. The module 10 collects data associated with various objects detected by the camera 6. The module 10 also collects odometer data of the vehicle 2. More precisely, the module 10 collects the position of the vehicle 2 in the reference system, the orientation of the vehicle 2 in the reference system, the steering wheel angle, the speed of the vehicle 2 and the speed of the heading angle of the vehicle 2.

[0069] For each target, the detection module 10 collects target position accuracy information. In this case, the position accuracy information is the position standard deviation. This data is delivered directly by the image processing algorithm integrated into the camera 6. Whatever the target, the target's position standard deviation corresponds to the standard deviation of the target's position acquired several times by the camera 6. For each target, the detection module 10 also collects target orientation accuracy information, in this case, the standard deviation of the target's orientation. This data is also delivered by the image processing algorithm integrated into the camera 6. Whatever the target, the target's orientation standard deviation corresponds to the standard deviation of the target's orientation acquired several times by the camera 6.

[0070] The module 10 has a counter 11. Whatever the target is, the counter 11 determines the consecutive detection count signals associated with that target.

[0071] The device 8 has a filtering module 12. The function of the module 12 is to filter the objects detected by the camera 6 and collected by the module 10. To this end, the module 12 is informationally connected to the module 10. The module 12 is equipped with hardware and software devices for implementing the steps of filtering the objects collected by the module 10 based on the data received by the module 10.

[0072] The device 8 comprises an evaluation module 14. The role of the module 14 is to evaluate, for each target target_i collected by the module 10 and not rejected by the module 12, the orientation θ of the target relative to the road. i / rte To this end, the module 14 notably receives the respective orientations of the objects collected by the module 10. The module 14 also receives the odometer data received by the module 10 and receives signals representative of the detection and shape of the ground markings defining the space for the flow of traffic.

[0073] The device 8 has a determination module 16. The function of the module 16 is to determine, based on the orientation evaluated by the module 14, the type of parking space under consideration.

[0074] To this end, the module 16 has a map 18 storing the values ​​of the parallel parking type item, the perpendicular parking type item and the herringbone parking type item as a function of the orientation of the target relative to the road. Figure 21. The mapping 18 is shown in FIG. The mapping 18 includes a first curve 20 represented by a thin dashed line, which corresponds to the value of the parallel parking type item. A second curve 22 represented by a thick dashed line corresponds to the value of the herringbone parking type item. A third curve 24 represented by a solid line corresponds to the value of the perpendicular parking type item. The curves 20, 22 and 24 are formed by a plurality of refined functions whose images are between -1 and 1. Regardless of the target, when the mapping 18 delivers a parallel parking type item, a perpendicular parking type item or a herringbone parking type item equal to 1, the target is parallel parking, perpendicular parking or herringbone parking, respectively.

[0075] By means of the device 8, it is possible to implement Figure 3 The method shown in the figure includes a first stage P1, a second stage P2, a third stage P3 and a fourth stage P4. The method is intended to be implemented periodically, for example, once per second.

[0076] Phase P1 comprises a first step E11 in which the module 10 detects targets by means of the front camera 6. Each target is denoted target_i, where i is an integer. Whatever the target target_i is, its position relative to the vehicle 2 is denoted x i / veh , the standard deviation of the target position is denoted as σ x_i , the speed of the target is represented by v i , the orientation of the target relative to vehicle 2 is denoted as θ i / veh , the standard deviation of the orientation of the target is denoted by σ θ_i , and the count signal associated with the target is denoted as Σ i In the exemplary embodiment of the considered method, ten targets are detected:

[0077] -target_1 is a vertically parked vehicle 15m in front of vehicle 2,

[0078] -target_2 is a vertically parked vehicle 15m behind vehicle 2,

[0079] -target_3 is a vehicle parked vertically on a curve 25m in front of vehicle 2.

[0080] -target_4 is a vehicle parked in a V-shaped pattern on a curve 25 meters behind vehicle 2.

[0081] -target_5 is detected with poor detection quality of vehicles,

[0082] -target_6 is detected with poor detection quality. of vehicles,

[0083] -target_7 is the vehicle moving in the opposite direction of vehicle 2,

[0084] - target_8 is a parked vehicle that was not detected by the previous two implementations of the method,

[0085] - target_9 is a parked vehicle 45m ahead of vehicle 2, and

[0086] - target_10 is a parked vehicle 60m behind vehicle 2.

[0087] Step E11 is followed by step E12. Whatever i is, in step E12 the front camera 6 and the module 10 detect the data associated with the target target_i, that is to say the position x i / veh , standard deviation σ x_i , speed v i , orientation θ i / veh and the standard deviation σ θ_i . It also collects the signal Σ delivered by the counter 11 i .

[0088] In a subsequent step E13, whatever target target_i was detected in the previous embodiment of the method but not in the current embodiment of the method, previous data are collected, ie data associated with the target target_i detected in the previous embodiment. Next, the updated data x is determined. i / veh , σ x_i 、v i ,θ i / veh and σ θ_i In order to determine the data x i / veh 、v i and θ i / veh , the previous data is updated by taking into account the data associated with the vehicle, such as the moving speed of the vehicle, the steering wheel angle of the vehicle, or the yaw angle of the vehicle. x_i and σ θ_i Same as the corresponding previous data. Finally, the updated data is stored for use in subsequent steps of the method.

[0089] Phase P2 comprises a step E21 in which the standard deviation σ x_i Compare with threshold threshold_1. If the standard deviation σ x_i If the threshold value threshold_1 is exceeded, the target target_i is rejected in step E21. In the example considered, the threshold value threshold_1 is between 2m and 4m, preferably substantially equal to 3m. In this case, the position x of the target target_5 is rejected.5 / veh The quality of the detection is poor, making the standard deviation σ x_5 Threshold threshold_1 is exceeded. Target target_5 is rejected in step E21.

[0090] Phase P2 has a step E22. Step E22 takes into account the standard deviation σ θ_i To perform a filtering similar to that in step E21 on the remaining targets. In the example considered, no matter what i is (except 5), if σ θ_i If the threshold threshold_2 is exceeded, the target target_i is rejected. The threshold threshold_2 is between 2° and 6°, and is preferably substantially equal to 4°. In this case, the quality of the detection of the orientation of the target target_6 is poor, so that the standard deviation σ θ_6 The threshold value threshold_2 is exceeded. As a result, the target target_6 is rejected in step E22.

[0091] Phase P2 comprises a step E23. In step E23, whatever i is (except 5 or 6), the speed v i is compared with the threshold value threshold_3, and if the speed v i If the threshold value threshold_3 is exceeded, the target target_i is rejected in step E23. In the example shown, the threshold value threshold_3 is between 8 km / h and 4 km / h and is preferably substantially equal to 6 km / h.

[0092] As an alternative, instead of considering the velocity v i , the highest speed among the speeds of the targets target_i measured by the camera 6 can be considered. In this way, it is avoided that already stopped vehicles (for example, vehicles stopped due to congestion, traffic lights or stop signals) are considered when determining the type of parking space.

[0093] In this case, the target target_7 moves in the opposite direction to the vehicle 2. Therefore, the speed v7 is much greater than the threshold threshold_3. As a result, the target target_7 is rejected in step E23.

[0094] Phase P2 has a step E24 in which, whatever i is (except 5, 6 or 7), the signal Σ i Compare with the counting threshold threshold_4. If the signal Σ iStrictly less than the threshold value threshold_4, the target target_i is rejected in step E24. In the example considered, the threshold value threshold_4 is equal to 5. However, completely different values ​​can certainly be envisaged without departing from the scope of the invention. In this case, the target target_8 was not detected in the second embodiment of the method before the current one. As a result, the signal Σ8 is equal to 2. The target target_8 is then rejected in step E24.

[0095] Phase P2 has a step E25 in which, whatever i is (except 5, 6, 7 or 8), the distance between the target target_i and the vehicle 2 is determined. The determined distance is then compared with a threshold threshold_5. In the example considered, the threshold threshold_5 is equal to 30 m. If the determined distance exceeds the threshold threshold_5, the target target_i is rejected in step E25. In this case, the distance determined for the target target_9 is 45 m. This distance exceeds the threshold threshold_5. The target target_9 is rejected in step E25. Similarly, the distance for the target target_10 is 60 m. The target target_10 should be rejected in step E25. At the end of step E25, phase P2 ends.

[0096] Phase P3 has a first step E31 of evaluating the orientation of the targets that are retained and that are located on the portion of the road that is on the straight line on which the vehicle 2 is travelling. In step E31, whatever i is, it is determined whether the target target_i is located along a straight line corresponding to the longitudinal direction of the vehicle 2. If this is the case, the orientation θ i / rte Equal to orientation In the example considered, the targets target_1 and target_2 are not rejected in phase P2 and are parked along the straight line portion that the vehicle 2 is traveling. Therefore, in step E31 , the orientations of the targets target_1 and target_2 are defined as follows:

[0097]

[0098] Phase P3 has a second step E32 in which it is determined whether there are objects that are reserved and parked along the road portion forming a curve. If the answer is "no", phase P3 ends. In this case, the objects target_3 and target_4 are both parked on the curve. Therefore, the answer to step E32 is "yes".

[0099] In this case, step E33 is applied, in which it is determined whether the bend on which the target is located has ground markings. For this purpose, camera 6 and module 10 can be used. If the answer to step E33 is "yes", step E34 is applied. If the answer to step E33 is "no", step E35 is applied.

[0100] In step E34, the orientation of the road is determined based on the ground markings. More precisely, the orientation of the road corresponds to the orientation of the road relative to the vector basis 3 at the point closest to the target target_i, regardless of whether the target_i is located on a curve. Then, the process proceeds to step E38, which is described below.

[0101] In step E35, the past trajectory of the vehicle 2 is collected. In this case, the past trajectory corresponds to the last thirty meters traveled by the vehicle. For this purpose, a memory (not shown) storing the trajectory of the vehicle 2 incorporated in the onboard computer 4 can be used.

[0102] Step E35 is followed by step E36 in which the possible trajectory of vehicle 2 is predicted. In order to determine the possible trajectory of vehicle 2, the speed v of vehicle 2 is collected. 2 / rte , the steering wheel angle α2 of vehicle 2 and the yaw angle β2 of vehicle 2. By assuming the speed v 2 / rte , angle α2 and angle β2 are kept constant over the entire possible trajectory to define a possible trajectory. Of course, only considering one or two of these three scenarios does not depart from the scope of the present invention.

[0103] In step E37, the road orientation at the location of the target is defined as the tangent of the past trajectory or possible trajectory of vehicle 2 at the point closest to the target. In this case, target_4 was not rejected in stage P2 and is parked on a curve behind vehicle 2. The road orientation at the location of target target_4 is indicated by the tangent of the past trajectory at the point closest to target target_4. Similarly, target target_3 is located on a curve in front of vehicle 2 and does not have any road markings. The road orientation at the location of target target_3 is indicated by the tangent of the possible trajectory at the point closest to target target_3. At the end of step E37, step E38 is applied.

[0104] In step E38 , whatever i is, the orientation θ is calculated corresponding to the target target_i retained in phase P2 and positioned along the curve. i / rteFor this purpose, the road orientation at the location of the target determined in step E34 or step E37 is used. In the example shown, the targets target_3 and target_4 are located on a curve without any ground markings. Therefore, the orientation determined in step E37 is used. At the end of step E38, phase P3 ends.

[0105] In the example presented, phase P3 implements two different modes of evaluating the orientation of the target relative to the road, depending on whether the target is located on a straight section on which the vehicle is travelling. However, it is certainly conceivable to use only one of these two evaluation modes without departing from the scope of the invention. In a first variant, phase P3 comprises only step E31. The advantage of this variant is that its design is simpler and that it works in the same way on a straight line as in a conventional parking lot. According to a second variant, phase P3 comprises only steps E33 to E38. Although this variant is more complex, it makes it possible to take into account the orientation of the road in order to take into account targets located on a curve and to adopt differences in orientation between the vehicle and the road.

[0106] Phase P4 has a first step E41 of reading the mapping 18. More precisely, whatever i is, an orientation θ is entered at the input of the mapping 18 corresponding to a target target_i that was not rejected in phase P2. i / rte , and the term t_c is read on curve 20 i , read the term t_b on curve 22 i , and the term t_e is read on curve 24 i If target_i is parallel parked, then term t_c i is equal to 1, and if it is not parallel parked, the term is equal to -1. Similarly, if the target target_i is parked vertically and in a herringbone pattern, the term t_b is equal to -1. i and t_e i is equal to 1. In this case, targets target_1, target_2, and target_3 are parked vertically, while target_4 is parked in a herringbone pattern. Result:

[0107]

[0108] And therefore:

[0109]

[0110] Phase P4 comprises a step E42 of calculating a coefficient c_d representing the distance of the object from the vehicle i Whatever i is, corresponding to the target target_i that was not rejected in stage P2, the coefficient c_d iis determined as a decreasing function of the distance between target_i and vehicle 2. In the example considered, if the distance between vehicle 2 and target_1 is less than 20m, the coefficient c_d i is equal to 1; and if the distance between target_i and vehicle 2 is greater than 20m, the coefficient is equal to 0.25. In this case:

[0111]

[0112] Phase P4 comprises a step E43 of determining a coefficient c_θ representing the orientation of the road taking into account ground markings on the road to evaluate the orientation of the road at the location of the target. i In the example considered, whatever i is, corresponding to a target target_i that has not been rejected in phase P2, if the orientation of the road at the location of the target target_i is determined in step E37, the coefficient c_θ i is equal to 0.25, or if not determined, it is equal to 1. In this case:

[0113]

[0114] Phase P4 comprises a step E44 in which a coefficient c_e representing the lateral clearance between the trajectory of the target and the vehicle is determined. i In step E44 , the lateral clearance between the trajectory of the vehicle 2 and the position of the target is determined. Whatever i is, corresponding to a target target_i that has not been rejected in phase P2 , if the lateral clearance for the target target_i is less than the threshold threshold_6 , the coefficient c_e i is equal to 1. If it is not less than the threshold, the coefficient c_e i is equal to 0.25. In the example shown, the threshold value threshold_6 is equal to 1.5 times the length of the vehicle 2. The trajectory of the vehicle 2 considered may be a past trajectory and / or a possible trajectory determined in steps E35 and E36. In this case, the targets target_1, target_2 and target_3 are parked along the road, while the target target_4 is parked on the side road. Therefore, the lateral distance between the trajectory of the vehicle 2 and the target target_4 exceeds the threshold value threshold_6:

[0115]

[0116] Phase P4 includes steps E451 and E452. In step E451, various coefficients c_d i 、c_θ i and c_e i For item t_bi ,t_e i , t_c i The weighting is performed and in step E452 the weighted terms are summed. More precisely, in steps E451 and E452, three sums ∑b, ∑e and ∑c are calculated:

[0117]

[0118]

[0119]

[0120] In this case, using the values ​​established previously, the calculated sum has the following value:

[0121] ∑b=2.046875

[0122] ∑e=-2.046875

[0123] ∑c=-2.078125

[0124] Phase P4 comprises a step E46 of determining the type of parking space. In step E46 the calculated sums are compared. More precisely, for each sum ∑b, ∑e or ∑c, the sum is determined by adding a strictly positive offset term t offset To calculate the offset sum ∑b′, ∑e′ or ∑c′. In the example shown, the term t offset is equal to 1. If there are three sums ∑b, ∑e or ∑c that are greater than the offset sum calculated from the other two sums, then the type of parking space corresponds to the parking space of this sum:

[0125] - If ∑b>∑c+t offset And ∑b>∑e+t offset , then the parking space is a vertical parking space,

[0126] - If ∑c>∑b+t offset And ∑c>∑e+t offset , then the parking space is a parallel parking space,

[0127] - If ∑e>∑c+t offset And ∑e>∑b+t offset ,, then the parking space is a herringbone parking space.

[0128] If such a sum does not exist, the type of parking space next to vehicle 2 is undefined.

[0129] In this case, the sum ∑b is greater than the sum of each of ∑e and ∑c plus the term t offset . Therefore, the parking space is a perpendicular parking space.

[0130] Item offset This makes it possible to increase the reliability of the determination of the type of parking space. However, the presence of this term has the consequence that no conclusions can be drawn about the type of parking space. The value of this term can of course be modified without departing from the scope of the invention in order to adjust the trade-off between reliability and driver convenience. It is also possible to eliminate the offset, in which case one parking space type is practically always determined.

[0131] Without departing from the scope of the invention, different offset terms may also be envisaged for each comparison. In this case, an asymmetric compromise may be made to increase the confidence level in determining the type of parking space.

[0132] In view of the above, the present invention makes it possible to reliably determine the type of parking space before a vehicle drives past an available parking space, and can function even in the absence of any ground markings defining the parking space.

Claims

1. A method for determining a parking space type for a motor vehicle (2), the method comprising: - Phase of detecting the target (P1), - Evaluate the orientation of the target (θ i / rte ), if the target is on a portion of the road that is in a straight line with which the vehicle is travelling, evaluating the orientation of the target relative to the portion of the road on which the vehicle is travelling, and if the target is not on a portion of the road that is in a straight line with which the vehicle is travelling, evaluating the orientation of the target relative to the portion of the road at the position of the target, and - Based on the estimated orientation (θ i / rte ) a stage of determining the type of parking space (P4), It is characterised in that the detection phase (P1) is carried out by means of a front camera (6) of the vehicle (2).

2. The method of claim 1, wherein: In the detection phase (P1), data associated with the target is determined, and the data associated with the target includes a target continuous detection count signal, a position of the target (x i / veh ), a variable representing the accuracy of the target's position, and the target's speed (v i ), the orientation of the target (θ i / veh ) and a variable representing the orientation of the target, the method further comprising a filtering phase (P2) between the detection phase (P1) and the evaluation phase (P3), the filtering phase (P2) comprising at least one filtering step selected from the following: an imprecise location filtering step ( E21 ) in which the target is rejected if the variable representing the accuracy of the location exceeds a location threshold and is retained if the variable does not exceed the location threshold; an imprecise orientation filtering step (E22) in which the target is rejected if the variable representing the precision of the orientation exceeds an orientation threshold, and is retained if the variable does not exceed the orientation threshold; - A moving vehicle filtering step (E23), in which if the speed of the target (v i ) exceeds a third threshold, rejecting the target, and retaining the target if its speed does not exceed the third threshold; and - A phantom detection filtering step (E24), in which the target is filtered out if the continuous detection count signal is less than a count threshold, and the target is retained if the signal is not less than the count threshold.

3. The method according to claim 1 or 2, wherein: In the detection phase (P1), the position of the target (x i / veh ), the method further comprises a filtering phase (P2) between the detection phase (P1) and the evaluation phase (P3), in which, if the position (x i / veh ) and the vehicle (2) exceeds a distance threshold, the target is rejected (E25), and if the distance does not exceed the distance threshold, the target is retained.

4. The method according to claim 1 or 2, wherein: In the evaluation phase (P3), the orientation of the road at the location of the target is determined and the orientation of the target relative to the road is evaluated based on the orientation of the road at the location of the target.

5. The method of claim 4, wherein: If the road has ground markings, the road orientation at the position of the target is determined (E34) by taking into account the ground markings; if the road does not have any ground markings and if the target is located behind the vehicle (2) or at the same position as the vehicle (2), the past trajectory of the vehicle is determined (E35) and the road orientation at the position of the target is evaluated (E37) based on the past trajectory at the position of the target; and if the road does not have any ground markings and if the target is located in front of the vehicle (2), the possible trajectory of the vehicle (2) is determined and the road orientation at the position of the target is determined based on the possible trajectory at the point closest to the target.

6. The method of claim 5, wherein: When determining the possible trajectory of the vehicle (2), at least one scenario selected from the following scenarios is accepted: - a speed scenario in which the speed of the vehicle (2) remains constant; - an acceleration scenario in which the acceleration of the vehicle (2) remains constant; - a steering wheel angle scenario in which the steering wheel angle of the vehicle (2) remains constant; and - A yaw angle scenario in which the yaw angle of the vehicle (2) remains constant.

7. The method according to claim 1 or 2, wherein: In the detection phase (P1), a plurality of targets are detected (E11), and wherein, in the determination phase (P4), whatever the target is, at least one first term and one second term are associated (E41) with the target, these terms being selected from the following: -Parallel parking type item, -Vertical parking type items, and - Herringbone parking type entry, And then calculate (E452) the first sum of the first items of each target and the second sum of the second items of each target, and determine (E46) whether one of the first sum and the second sum is greater than the other sum, and determine whether one of the first sum and the second sum is greater than the other sum plus a strictly positive offset term.

8. The method of claim 7, wherein: Whatever the goal is, at least one coefficient is associated with the goal (E42, E43, E44), the at least one coefficient being selected from the following: - a coefficient representing the distance of the target from the vehicle (2), - means taking into account the ground markings on the road to estimate the orientation of the target relative to the road (θ i / rte ), and - a coefficient representing the lateral clearance between the target and the trajectory of the vehicle (2), And before calculating these sums, the terms associated with the objective are weighted by the coefficient associated with the objective ( E451 ).

9. A computer program product comprising code configured to implement the method according to any one of claims 1 to 8 when executed by a processor or an electronic control unit.

10. A device (8) for determining the type of parking space of a motor vehicle (2), the device being capable of being accommodated onboard the motor vehicle (2), the device (8) comprising: a detection module (10) for detecting an object, the detection module (10) being capable of establishing an information connection with a front camera (6) of the motor vehicle (2), - an evaluation module (14) capable of evaluating the orientation (θ i / rte ), if the target is on a portion of the road that is in a straight line with which the vehicle is travelling, evaluating the orientation of the target relative to the portion of the road on which the vehicle is travelling, and if the target is not on a portion of the road that is in a straight line with which the vehicle is travelling, evaluating the orientation of the target relative to the portion of the road at the location of the target, and a determination module (16) capable of determining the orientation (θ) evaluated by the evaluation module (14) by taking into account i / rte ) to determine the type of parking space.

Images