Method for determining the occupant weight of an occupant of a vehicle

Abstract

The invention relates to a method, in particular a computer-implemented method, for determining the occupant weight of an occupant on a vehicle seat, the method comprising detecting, using a body posture sensor device, one or more body postures of the occupant, determining, by means of a processing device, whether the body posture corresponds to a predetermined body posture, and if the body posture corresponds to the predetermined body posture, reading out, by means of an occupant weight sensor device, sensor measured values and determining, by means of the processing device, the occupant weight from the sensor measured values, and wherein in particular sensor measured values are not read out and in particular the occupant weight is not determined if the body posture does not correspond to the predetermined body posture.

Description

[0001] Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0002] Max-Planck-Str. 11 Industriestraße 20

[0003] 85716 Unterschleissheim 73553 Alfdorf

[0004] 301438-WO-PCT

[0005] October 21, 2025

[0006] METHOD FOR DETERMINING THE WEIGHT OF AN OCCUPANT OF A VEHICLE

[0007] TECHNICAL BACKGROUND

[0008] The invention relates to a method for determining the weight of an occupant sitting on a vehicle seat.

[0009] STATE OF THE ART

[0010] Numerous methods for determining an occupant's weight are known in the art. For example, in these known methods, one or more weight sensors detect a weight composed of the weight of a vehicle seat and the weight of the occupant, and then use this determined weight, particularly for controlling the vehicle's safety functions. According to an alternative method, a contactless sensor determines whether a vehicle seat is occupied, and if the seat is occupied, a drive device is operated to adjust the seat. The occupant's weight is then determined from the drive current required by the drive device during the seat adjustment.

[0011] The methods known from the prior art for determining occupant weight meet the requirements regarding accuracy in determining Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0012] 301438- WO- PCT October 21, 2025 of the occupant weight, especially since occupant weight is a crucial parameter for controlling safety functions to ensure optimal occupant protection. Therefore, there is a great need to provide a method and device for determining occupant weight that allows for increased accuracy in determining occupant weight.

[0013] SUMMARY OF THE INVENTION

[0014] The invention is based on the objective of providing a method and a device with which the weight of an occupant of a vehicle, in particular sitting on a vehicle seat, can be determined with the highest possible accuracy.

[0015] Another task is to provide a processing device for processing sensor measurements to determine the occupant weight according to the above method, as well as a computer program product and a computer-readable storage medium, comprising instructions which, when executed, cause a processing device to perform the above method.

[0016] The problem is solved by the method according to claim 1. Advantageous embodiments are specified in the respective dependent claims. The problem is further solved by the device with the features of claim 10. Further advantageous embodiments are specified in the respective dependent claims.

[0017] One aspect of the invention relates to a method, in particular a computer-implemented method, for determining the occupant weight of a vehicle occupant, comprising:

[0018] • Detecting, using a posture sensor device, one or more body postures of the occupant,

[0019] • Determine, using a processing unit, whether the body posture corresponds to a predetermined body posture, and, if the body posture corresponds to the predetermined body posture: o Read out, using an occupant weight sensor device, sensor measurements, and Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0020] 301438- WO- PCT 21 October 2025 o Determine, by means of the processing device, the occupant weight from the sensor readings, tmd-whereby in particular the reading of sensor readings and the determination of the occupant weight shall not take place if the body posture does not correspond to the predetermined body posture.

[0021] In some embodiments of the present invention, a body posture refers to the position of the body, i.e., the orientation of the various limbs of the body relative to one another, in particular of an occupant's upper body, wherein the body posture is parameterized. For example, the body posture can be parameterized in a set of support points and may also include connecting axes linking the support points or a subset of the support points. Various models or specifications exist for the number of support points and connecting axes. In particular, the body posture can be designed in two spatial dimensions or in three spatial dimensions; it is then also called a 2D body posture or a 3D body posture.

[0022] In the context of the present invention, a reference point, also called a key point, is a point on an occupant's body by means of which the occupant's posture can be parameterized. Typically, the various joints of a person are used for this purpose, as well as, in particular, prominent points, especially prominent points on the face, such as the eyes, nose, mouth, and ears. These points are assigned coordinates, for example, by means of an analysis of image data, which can then be used to further process the posture of the respective occupant.

[0023] For the purposes of the present invention, a vehicle is a passenger car, a truck, or a bus. In particular, a vehicle is referred to as a motor vehicle. The two terms are used synonymously here.

[0024] In some embodiments of the present invention, an occupant is a person who is in the vehicle, in particular sitting in the vehicle.

[0025] In some embodiments of the present invention, an occupant weight is defined as the weight of an occupant of a vehicle, in particular the mass of the occupant. The terms weight and mass are used synonymously hereafter. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT October 21, 2025

[0026] In some embodiments of the present invention, a body posture sensor device is in particular a camera, a depth camera or a radar.

[0027] Known methods for determining occupant weight from the prior art exhibit significant inaccuracies due to the lack of consideration for the occupant's posture. Simulations and measurements have shown that occupant weight must be taken into account when controlling certain vehicle safety functions, particularly occupant restraint systems, to ensure optimal occupant protection. It has been shown that occupants whose weight deviates significantly from a standard weight are at particularly high risk of not being optimally protected by safety functions during an accident, as these functions are controlled based on assumptions regarding the occupant's weight.In particular, the inventors of the present invention have recognized that the accuracy of determining occupant weight using an occupant weight sensor device is highly dependent on the occupant's posture. Therefore, the inventors propose determining the occupant weight only when the occupant is seated in a predetermined posture on the vehicle seat. By first comparing the occupant's posture with this predetermined posture, the accuracy of the occupant weight determination can be significantly improved. Based on this more accurately determined occupant weight, the safety functions can be optimally adapted to the individual occupant, thereby increasing the occupant's safety in the vehicle.

[0028] Another advantageous embodiment relates to the above method, wherein the sensor measurements comprise one or more of the following: sensor measurements from a vehicle seat position control device and sensor measurements from one or more weight sensors forming the occupant weight sensor device, which detect the weight of the vehicle seat and the occupant sitting on the vehicle seat.

[0029] Modern vehicles are equipped with a variety of sensors that can be used to determine occupant weight. The present method, Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT 21 October 2025, can be applied variably to obtain the most accurate result possible in each case.

[0030] A further advantageous embodiment relates to the above method, wherein the one or more weight sensors comprise several weight sensors, in particular at least four weight sensors, and the reading of sensor measurements comprises reading the sensor measurements of the several weight sensors, and the determination of the occupant weight comprises determining a total weight across all sensor measurements and determining a weight distribution based on the sensor measurements, wherein, in particular, anthropometric data are incorporated into the determination of the occupant weight from the weight distribution. These weight sensors can, in particular, be strain gauges arranged between the vehicle seat and a vehicle-fixed structure, for example, the seat rail.

[0031] Using weight sensors to determine occupant weight allows for further improvement in weight determination. In particular, using weight distribution and considering anthropometric data can further enhance accuracy, for example, by appropriately combining anthropometrically determined weight proportions of different body parts of an occupant based on posture.

[0032] Another advantageous embodiment relates to the above method, wherein determining the occupant weight includes determining the support state of the thighs and correcting the determined total weight based on the determined support state, thereby significantly increasing the accuracy in determining the occupant weight.

[0033] Alternatively or additionally to determining the occupant's weight using the weight sensors described above, capacitive measurement methods can also be used, which in particular employ sensors in the seat surface. Preferably, at least one mat serving as a capacitor surface is used beneath the seat cover, which, together with the occupant and the vehicle floor, forms an alternating current circuit whose impedance depends on the body shape and consequently correlates with the body volume and mass. A capacitive sensor can also be used to detect changes in impedance. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT October 21, 2025

[0034] A measurement method is used in which, similar to the well-known hands-on detection in the steering wheel, the changed conductivity of a seat mat (e.g., the seat heating) is used to determine the weight.

[0035] Furthermore, the seatbelt extension or a seatbelt extension sensor can also be used to estimate the occupant's weight. This approach utilizes the principle that a large seatbelt extension correlates with a large body volume and therefore a high weight. To obtain meaningful measurement results, the attachment points of the seatbelt system should be fixed relative to the occupant, as is the case, for example, with a seat-integrated seatbelt system.

[0036] These measurement methods can each be used individually in conjunction with the posture sensor device, whereby the combination of several of the aforementioned measurement methods in combination with the posture sensor device leads to a further increase in accuracy.

[0037] Another advantageous embodiment relates to the above method, wherein correcting the determined total weight includes one or more of the following:

[0038] • Determining the occupant weight based on a sum of the total weight and anthropomorphic weight components of the lower extremities, whereby the weight component of the lower extremity in a non-supported state is added to the determined total weight to determine the occupant weight,

[0039] • Determining a linear occupant seat position correction factor based on a quotient of the sum of the weights detected by the front weight sensors and the determined total sum, and determining the occupant weight from the product of the total weight and the determined occupant seat position correction factor.

[0040] It is known from anthropometric data that the lower extremities account for up to 20 percent of the total weight. Accordingly, when determining the occupant's weight, variations in thigh position, particularly between a standing position and a position resting on the seat, can already lead to an error of approximately 20 percent, as the weight is not adequately measured depending on the thigh position (Simi Reality Motion Systems GmbH, ZF Automotive Germany GmbH 301438-WO-PCT, October 21, 2025). The inventors of the present invention have recognized that there is a linear relationship between the position of the thighs and the weight distribution between the weight detected by the front weight sensor and the total weight, based on which the position of the thighs can be determined.Accordingly, the occupant weight determined by the weight sensors can be easily corrected according to the support condition, thus significantly improving the accuracy in determining the occupant weight.

[0041] Another advantageous embodiment concerns the above procedure, whereby the determination of the occupant weight takes place as soon as possible after the start of a journey with the vehicle.

[0042] According to some embodiments of the present invention, "promptly after the start of a journey" here and in the following means within the first five, within the first ten or also within the first 15 minutes after the start of the journey.

[0043] It has been shown that the posture of the occupants changes considerably during the journey, which in turn can significantly affect the accuracy of determining the occupant weight. In particular, the inventors have recognized that the occupant weight can be determined most accurately when the occupants are in an upright posture. This is especially true in the period immediately after the journey begins. Therefore, the occupant weight is preferably determined directly after the journey starts, and this determined weight is then used throughout the journey.

[0044] Another advantageous embodiment relates to the above method, additionally encompassing: determining a position of the vehicle seat, and determining a vehicle seat position correction for determining the occupant weight based on the determined position of the vehicle seat.

[0045] The inventors also determined that, depending on the seating position, particularly along the direction of travel, the vehicle seat position correction must be determined before the weight distribution can be determined to establish the support state. Since the weight sensors are typically attached to seat rails for adjusting the vehicle seat position, and the seating position accordingly influences the Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT 21 October 2025

[0046] Since the weight distribution is variable, the seating position must be factored in accordingly. Specifically, based on an existing seating position, the weight distribution is converted to a nominal seating position, and the weight distribution and thus the contact state are determined based on this nominal seating position, which in turn increases the accuracy of determining the occupant's weight.

[0047] Another advantageous embodiment relates to the above method, wherein determining the occupant weight includes determining an average value, wherein the average value is formed from occupant weights determined over a period of time in a predetermined body posture, and in particular sensor measurements taken when the predetermined body posture is not present are not included in the determination of the occupant weight.

[0048] By calculating an average value, the accuracy in determining the occupant weight can be further improved.

[0049] A further advantageous embodiment relates to the above method, wherein determining whether the body posture corresponds to the predetermined body posture further comprises: taking an image of the occupant, determining reference points of the body posture in the image, and classifying the body posture resulting from the determined reference points, wherein the classification particularly comprises one or more of the following:

[0050] • Classifying body posture using a specially trained classification model,

[0051] • Comparing the posture with the predetermined posture and classifying the posture based on the comparison, wherein comparing the posture with the predetermined posture includes in particular calculating a measure of agreement or a measure of distance between the support points of the posture and support points of the predetermined posture.

[0052] A comparison of the occupant's posture with a predetermined posture can be performed in different ways according to the embodiments. This allows, in particular, for a flexible determination of suitable times for determining the occupant's weight.

[0053] Another aspect concerns an occupant weight determination device for determining the weight of an occupant of a vehicle, in particular for Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0054] 301438-WO-PCT October 21, 2025

[0055] Use in one of the above-described methods, comprising a posture sensor device for detecting the posture of the occupant, an occupant weight sensor device for reading sensor measurements, a processing device for determining the occupant weight from the sensor measurements, characterized in that the occupant's posture at the time of weight measurement corresponds to a predetermined posture.

[0056] Another aspect concerns a processing unit for processing sensor measurements, comprising a processor designed to perform the procedure described above.

[0057] Another aspect concerns a computer program product, comprising instructions which, when executed by a processing unit, cause it to perform the procedure described above.

[0058] Another aspect concerns a computer-readable storage medium, comprising instructions which, when executed by a processing unit, cause it to carry out the procedure described above.

[0059] SUMMARY OF THE CHARACTERS

[0060] The invention is explained in more detail below with reference to the examples shown in the drawings. The drawings show:

[0061] FIG. 1 schematically shows a vehicle for use in a method according to one aspect according to at least one embodiment.

[0062] FIG. 2 schematically shows a processing device for use in a method according to an aspect according to at least one embodiment,

[0063] FIG. 3 schematically shows a method according to one aspect according to at least one embodiment.

[0064] FIG. 4A schematically shows a body posture of a vehicle occupant for use in a method according to one aspect according to at least one embodiment, Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT 21 October 2025

[0065] FIG. 4B schematically shows another body posture of an occupant of a vehicle for use in a method according to an aspect according to at least one embodiment,

[0066] FIG. 5 schematically shows the body postures of an occupant of a vehicle for use in a method according to one aspect according to at least one embodiment.

[0067] FIG. 6 schematically shows a method according to one aspect according to at least one embodiment.

[0068] DETAILED DESCRIPTION OF THE EXECUTION FORMS

[0069] One embodiment of a vehicle 100 comprises a posture sensor device 110 and a processing unit 200 (see FIG. 2). The posture sensor device 110 is communicatively coupled to the processing unit 200 (for example, via a wired or wireless communication link). According to FIG. 1, the vehicle 100 also comprises an interior 104, in which, according to this embodiment, an occupant 102 sits.

[0070] The posture sensor device 110 is configured to detect the posture 400 of the occupant 102 in the interior 104 of the vehicle 100. According to the first embodiment, the posture sensor device 110 comprises at least one camera configured to capture image data. Alternatively, a depth sensor can be used that emits short light pulses and determines a transit time for each pixel of the depth sensor, the time required for a reflected light pulse to return to the depth sensor. In addition to the transit time, the depth sensor also determines the intensity of the reflected light pulse. Based on the determined transit time, the depth map is generated, and the depth map then forms the image data for further processing by the processing device 200.

[0071] According to the present embodiment, the recorded image data is passed on to the processing unit 200 for further processing. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT October 21, 2025

[0072] The processing unit 200 comprises an evaluation module 202, a storage module 204, and a control module 206. The modules can be designed as either software or hardware modules and are interconnected via channels 208 (see FIG. 2). The channels 208 are logical data connections between the individual modules. Alternatively, the modules can also be interconnected via a bus system.

[0073] The storage module 204 stores the image data recorded by posture sensor device 110 and manages the data to be evaluated in the processing unit 200.

[0074] The evaluation module 202 is designed to process the image data acquired by the depth sensor 110. In particular, the evaluation module 202 comprises one or more processing models, especially machine learning models, which are used, firstly, to evaluate the image data in order to determine the posture of an occupant and compare it with a predetermined posture; secondly, a processing model is used to determine the body weight of the occupant 102 from sensor readings of an occupant weight sensor device.

[0075] According to the present invention, the control module 206 is configured to control the processing unit 200. The control module 206 reads data from the storage module 204 and forwards the data to the evaluation module 202. Furthermore, the control module 206 forwards the control information for controlling occupant functions, in particular safety functions, to a respective control unit for controlling the respective occupant function.

[0076] The processing models can include, in particular, machine learning models, but also classical processing models. Specifically, machine learning models trained to determine the support points of occupants captured in the image data are usually used to determine body posture. Different processing models can also be implemented in the processing unit for the various steps involved in processing the image data and the sensor measurements.

[0077] The following describes a method for determining the weight of an occupant on a vehicle seat 500 with reference to FIGS. 3 to FIGS. 6. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0078] 301438-WO-PCT October 21, 2025

[0079] According to the procedure, in step 310, a body posture is first detected using the posture sensor device 110. Detecting the body posture includes, in particular, capturing image data of the interior 104 of the vehicle 100. After capturing the image data, the image data is analyzed to identify image areas in which occupants 102 were detected and to determine the body posture of each occupant 102 for each image area with a detected occupant 102.

[0080] The body posture 400 comprises support points 401 and optionally also connecting axes 404 between certain support points 401, as exemplified in FIG. 4A. For each image area containing an occupant 102, the support points of the occupant 102 are determined. A field of view 111 of the body posture sensor device 110 typically captures the head, upper extremities, and torso of the occupant 102 (see especially FIG. 5). The body posture sensor device 102 can also capture the pelvis of the respective occupant. If, for example, the occupant's seating position is chosen too far forward, the pelvis slips out of the field of view 111 of the body posture sensor device 110. The legs are typically only partially captured at most, and the occupant's feet are typically not captured at all. Accordingly, the body posture 400 typically includes support points 401 extending at most to the pelvis. For clarity, FIG. 4A and FIG. 5 show the support points 401 in the image area 400.4B only a part of the support points 401 shown as hatched circles are provided with a reference sign.

[0081] To determine the support points, a machine learning model is used that has been trained to identify the support points 401 of the body posture 400. For example, the machine learning model outputs a list of image coordinates of the respective support points 401 and the corresponding type of support point. For example, the body posture 400 includes a support point 401 for the head, an upper central support point 402, which corresponds approximately to the position of the sternum and is located centrally between the shoulders of the respective occupant 102, one support point 401 each for the right and left shoulders, and optionally, support points 401 can also be output for the elbows and hands (not shown here). Furthermore, a lower central support point 403 and one support point 401 each for the right and left hips are determined. The lower central support point 403 is located approximately midway between the support points 401 of the hips.If the pelvis is no longer captured in the image data, the machine learning model determines the lower Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH.

[0082] 301438- WO- PCT October 21, 2025. The central support point 403 is positioned approximately in the center of the body. The lower central support point 403 should be located approximately at a position on the spine of the occupant 102 and correspond approximately to the position of the lowest point of the torso captured in the image data. In addition to the support points 401, the machine learning model can also determine connecting axes 404 that connect certain of the adjacent support points 401. According to this embodiment, a central connecting axis 408 that connects the upper central support point 402 with the lower central support point 403 is particularly important, since, according to this embodiment, the orientation 406 of the body posture is determined based on the central connecting axis 408.

[0083] According to the first embodiment, step 310 of detecting one or more body postures 400 of the occupant 102 is followed by step 320 of determining, by means of the processing device 200, whether the body posture 400 corresponds to a predetermined body posture. The inventors of the present invention have recognized that the more upright the occupant 102 sits in the vehicle 100, the more reliably the occupant's weight can be determined. This is particularly due to the fact that occupants 102 sitting in the vehicle seat 500 with a body posture 400 tilted in any direction, as shown, for example, in FIG. 4B, often lean on an armrest or support, for example. Therefore, an occupant weight detected by an occupant weight sensor device is distorted by the respective support force.

[0084] The body posture 400 shown in FIG. 4A corresponds precisely to an upright posture and thus to the predetermined body posture. To determine whether the body posture 400 of an occupant 102 corresponds precisely to the predetermined body posture, the orientation 406 of the body posture 400 can be determined, for example. The orientation 406 corresponds precisely to the direction of the central connecting axis 408, which connects an upper central support point 402, located midway between the support points 401 corresponding to the shoulders of the occupant 102, and a lower central support point 403, located midway between the support points 401 corresponding to the hips of the occupant 102. If this central connecting axis 408 is vertically oriented, this means that the occupant 102 is sitting upright and thus the occupant weight can be reliably detected. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0085] 301438-WO-PCT October 21, 2025

[0086] The body posture 400 shown in FIG. 4B has a tilted orientation 406 compared to the body posture 400 shown in FIG. 4A. Depending on how much the orientation 406 deviates from the orientation 406 in the predetermined body posture, the processing device 200 would determine, when the body posture 400 according to FIG. 4B with the tilted orientation 406 compared to the orientation 406 from FIG. 4A is present, that the body posture 400 does not correspond to the predetermined body posture. If the processing device 200 were to determine that the body posture 400 does not correspond to the predetermined body posture, the process would accordingly terminate at this point and not proceed to the following steps 330, 340.

[0087] If the processing unit 200 determines that the current body posture 400 corresponds to the predetermined body posture, the next step of the method, step 330, reading sensor measurements using an occupant weight sensor device, is executed. According to this embodiment, the occupant weight sensor device comprises several weight sensors 501, 502, see in particular FIG. 5. Only two weight sensors 501, 502 are shown in FIG. 5. For example, one weight sensor 501, 502 can be provided on each side of the vehicle seat, front and rear. Assuming, for example, an approximately right-left symmetrical load on the weight sensors 501, 502, a total weight is determined from the two measured values. According to an alternative of this embodiment, however, the sensor measurements from four weight sensors 501, 502 can also be read.The weight sensors are attached to the front and rear ends of two seat rails 503, so that the weight sensors detect the weight of the vehicle seat 500 and the seat rails 503 together with the occupant weight.

[0088] According to one embodiment, the weight sensors 501, 502 can each be individually zeroed, i.e., for each of the weight sensors 501, 502 there is a tare weight that is subtracted from each measured value. In the nominal forwardmost vehicle seat position, the tare weight corresponds precisely to the proportion of the vehicle seat's weight that is detected by the respective weight sensor 501, 502. Alternatively, a tare weight can be specified for each vehicle seat position for each weight sensor 501, 502. According to another embodiment, the weight of the vehicle seat is subtracted from the sum of the measured values ​​detected by the weight sensors 501, 502 in order to determine the weight of the occupant 102 sitting in the vehicle seat. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0089] 301438-WO-PCT October 21, 2025

[0090] Step 330 is followed by step 340, which involves determining the occupant weight from the sensor readings using the processing unit 200. According to this embodiment, determining the occupant weight includes determining a total weight based on the sensor readings of the individual weight sensors 501 and 502. In this embodiment, the weight sensors 501 and 502 are assigned the weight of the vehicle seat 500 and the weight of the occupant 102. Accordingly, the sum of the sensor readings is proportional to the sum of the weight of the vehicle seat 500 and the weight of the occupant 102. Therefore, the weight of the vehicle seat 500 is subtracted from the total weight to determine the overall weight. Here and in the following, the total weight refers to the weight of the occupant 102 as detected by the weight sensors 501, 502. The total weight does not include the weight components of body parts that are not supported or resting on the vehicle seat 500.For example, if one leg is bent, part of the weight of the lower extremities rests on the vehicle floor; this weight component is not included in the total weight recorded by the weight sensors 501, 502.

[0091] The inventors of the present invention have thus recognized that the total weight detected by the weight sensors 501, 502 includes or excludes the weight of the lower extremities, depending on the position of the occupant's lower extremities. Of the two body postures 400 shown in FIG. 5, one corresponds to the lying position, in which the occupant's thighs rest on the vehicle seat 500, while the second depicted body posture corresponds to an upright position. In the upright position, the knee is bent, so that the thigh does not rest on the vehicle seat and the front weight sensors 502 do not detect a portion of the weight of the lower extremities, or at least only detect it partially.Since, according to anthropometric data, the weight proportion of the lower extremities to a person's weight, in this case the occupant 102, is approximately 20 percent, an error of up to 20 percent in determining the occupant's weight can occur simply by failing to consider these two conditions. To prevent this, the inventors of the present invention propose to determine the leg position of the thighs based on a so-called anterior weight proportion, which corresponds to a quotient of the weight detected by the anterior weight sensors 502 and the total weight. The inventors have recognized that the greater the weight proportion on the anterior Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT 21 October 2025.

[0092] The more weight sensors 502 detect, the more likely it is that the thighs are resting on the vehicle seat 500. Accordingly, according to the present invention, a correction is made based on the front weight component, particularly if the front weight component is less than a limit value, by adding the typical weight component of the lower extremities to the determined total weight. This correction is omitted as soon as the weight component of the front weight sensors 502 exceeds the limit value.

[0093] Thus, the present invention makes it possible to determine the occupant weight particularly precisely and, in particular, to control the safety functions of the vehicle 100 based on the determined occupant weight in a way that is particularly well adapted to the occupant weight.

[0094] According to a further embodiment, the seating position of the vehicle seat 500 on the seat rail 503 can also be taken into account when determining the occupant weight. The inventors have recognized that, by means of, for example, a linear vehicle seat position correction, the sensor measurements can be determined relative to a nominal forwardmost seating position of the vehicle seat 500 on the seat rail 503. FIG. 6 above shows, for various test subjects in a body posture with bent knees, i.e., without the thighs resting on the vehicle seat 500, the front weight component (FS / SUM) for different seating positions (SeatX) of the vehicle seat 500 along the seat rail 503, whereby different markers, for example crosses, diamonds, circles, triangles, and others, are used for different test subjects. Here, a linear relationship between the front weight component and the seating position is shown to a first approximation.Accordingly, according to this embodiment, the front weight component can be used to calculate back to a nominal seating position by means of the linear vehicle seat position correction, which is determined, for example, by means of a linear regression, a straight line equation of the line 601 shown in FIG. 6 above, as shown in FIG. 6 above, for example to a frontmost seating position which, according to the illustration, corresponds exactly to position X = 0.

[0095] According to a second embodiment, sensor readings from a vehicle seat position control unit can be used instead of the sensor readings from the weight sensors 501, 502. The inventors have found that, for example, when adjusting the seat position of the vehicle seat 500 on the seat rail 503 using a motor, Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0096] 301438- WO- PCT October 21, 2025, the current drawn by the motor during adjustment depends on the occupant weight of the respective occupant 102. Accordingly, step 330 can comprise adjusting the seat position on the seat rail 503, wherein the adjustment preferably comprises a short forward and backward movement, the current drawn by the motor during adjustment being proportional to the occupant weight.In particular, the adjustment is carried out in such a way that an occupant 102 sitting in the vehicle seat 500 does not notice this adjustment or at least does not find it disturbing; experience has shown that this can be achieved, for example, by a very short forward and backward movement; it has been shown that it is sufficient to move the vehicle seat 500 only briefly enough to start moving; such short movements are practically imperceptible to the occupant 102, since the car is already in motion during this time and the movement is therefore only perceived as a short jerk.

[0097] Alternatively, the adjustment can also be initiated by the occupant 102, in which case the processing unit 200 additionally executes step 320 during the adjustment to determine whether the respective occupant 102 assumes a posture that corresponds to the predetermined posture. Since such an adjustment typically occurs after boarding, the sensor values ​​measured during this process could be used directly to determine the occupant's weight.

[0098] According to a further embodiment, the occupant weight can be determined from the weight distribution, for example, using a linear relationship. While, according to the embodiment described above, the weight contribution of the lower extremities is either added to the total weight or not, depending on the position of the thigh, a corrected occupant weight can instead be determined depending on the weight contribution from the front.

[0099] Figure 6 below shows, as an example, measured values ​​for different test subjects and different body positions, in particular different support conditions or proportions of the thighs on the vehicle seat. The abscissa represents the front weight fraction (FS / SUM) for different test subjects and different support conditions, and the ordinate represents the quotient of the total weight (SUM) to the actual weight of the respective test subject, hereinafter referred to as the total weight fraction (SUM / weight). A correspondence between the front weight fraction and the total weight fraction can be determined for the recorded measured values. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT 21 October 2025As an example, a straight line 602, determined using linear regression, is plotted on the measurement data. This line appears to describe the data relatively well, suggesting a linear relationship between the front weight fraction and the total weight fraction. However, other approximations can be used instead of a linear one. Based on a measured front weight fraction (FS / SUM), the total weight fraction (SUM / weight) can be determined from the equation of the line determined by linear regression, and from this, a corrected occupant weight can finally be calculated.

[0100] Alternatively, instead of linear regression, another fit function can be used; for example, a quadratic function or any other function can be approximated.

[0101] According to a further embodiment, the occupant weight can also be determined based on the sensor readings from weight sensors 501 and 502 using a specially trained machine learning model. This requires a training dataset containing, for example, the front weight component as training input and the occupant weight as target output. This dataset can then be used to train the machine learning model to determine the occupant weight from the sensor readings.

[0102] According to a further embodiment, step 320, "Determine whether the posture corresponds to the predetermined posture," can also be performed using a classification model trained to classify image data or postures. For this purpose, the classification model is trained with a training dataset that includes postures corresponding to the predetermined posture, postures that do not correspond to the predetermined posture, and associated annotations, also called target outputs. Based on the training data, the classification model then learns to identify postures that correspond to the predetermined posture.

[0103] According to a further embodiment, not only is the central connecting axis 408 used to compare the body posture 400 with the predetermined body posture, but for all detected or identified support points 401 of the body posture, Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT 21 October 2025

[0104] 400 uses a distance measure, also called a distance measure, to determine the distance between the detected body posture 400 and the predetermined body posture. If the distance is greater than a limit value, the body posture 400 does not correspond to the predetermined body posture; if the distance is less than or equal to the limit value, the body posture corresponds to the predetermined body posture, and the occupant weight can be detected using the occupant weight sensor device. Instead of a distance measure, a similarity measure, also called a conformity measure, can also be used. Ultimately, any suitable distance or similarity measure can be used here.

[0105] According to a further embodiment, the posture sensor device 110 can be a simple camera, but alternatively a depth camera or radar can also be used. In particular, any sensor device can be used that makes it possible to determine that the occupants are sitting in an upright posture 400 on the vehicle seat 500.

[0106] According to a further embodiment, the present invention also comprises a device for carrying out the method described above. In particular, a processing unit 200. The processing unit 200 can in particular be an ECU (electronic control unit).

[0107] According to a further embodiment, the present invention provides a computer program product comprising instructions that cause a computer to execute the method described above.

[0108] The variations and designs described for the different figures can be combined with one another. The designs shown and described are purely illustrative, and modifications are possible within the scope of the attached claims.

[0109] Another embodiment relates to an occupant weight determination device for determining the weight of an occupant 102 on a vehicle seat 500, which is configured to perform the above method. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH

[0110] 301438-WO-PCT October 21, 2025

[0111] REFERENCE MARK LIST

[0112] 100 vehicles

[0113] 102 inmates

[0114] 104 Interior

[0115] 110 posture sensor device

[0116] 111 field of vision

[0117] 200 processing facilities

[0118] 202 Evaluation module

[0119] 204 memory module

[0120] 206 Control module

[0121] 208 channels

[0122] 400 Body posture

[0123] 401 Base

[0124] 402 upper central base

[0125] 403 lower central base

[0126] 404 Connecting axis

[0127] 406 Orientation

[0128] 408 central connecting axis

[0129] 500 vehicle seats

[0130] 501 Weight sensor

[0131] 502 Weight sensor

[0132] 503 Seat rail

[0133] 601 Straight

[0134] 602 Straight

Claims

Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT October 21, 2025 REQUIREMENTS 1. A method, in particular a computer-implemented method, for determining the weight of an occupant of a vehicle, comprising: - Detecting, using a posture sensor device, one or more body postures of the occupant, - Determine, using a processing device, whether the body posture corresponds to a predetermined body posture, and, if the body posture corresponds to the predetermined body posture: - Reading sensor readings using an occupant weight sensor device; and - Determining, by means of the processing device, the occupant weight from the sensor readings, and wherein in particular the reading of sensor readings and the determination of the occupant weight does not take place if the body posture does not correspond to the predetermined body posture.

2. The method according to claim 1, wherein the sensor measurements comprise one or more of the following: - Sensor readings from a vehicle seat position control unit, - Sensor readings from one or more weight sensors forming the occupant weight sensor device, which detect the weight of the vehicle seat and the occupant sitting on the vehicle seat.

3. Method according to claim 2, wherein the occupant weight sensor device is based on a capacitive, an inductive or a resistance measurement method.

4. Method according to claim 2 or 3, wherein the occupant weight sensor device comprises a seatbelt extension sensor.

5. A method according to any one of claims 2 to 4, wherein the one or more weight sensors comprise several weight sensors, in particular at least four weight sensors, and the reading of sensor measurements comprises reading the sensor measurements of the several weight sensors and determining the Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT October 21, 2025 The occupant weight calculation includes determining a total weight across all sensor measurements and determining a weight distribution based on the sensor measurements, with particular consideration given to anthropometric data in determining the occupant weight from the weight distribution.

6. Method according to any one of the preceding claims 1 to 5, wherein determining the occupant weight comprises determining a support state of the thighs and correcting the determined total weight based on the determined support state.

7. The method according to claim 6, wherein correcting the determined total weight comprises one or more of the following: - Determining the occupant weight based on a sum of the total weight and anthropomorphic weight components of the lower extremities, whereby a weight of the lower extremity in a non-supporting state is added to the determined total weight to determine the occupant weight, - Correcting the determined total weight based on a correspondence between a weight fraction of front weight sensors in the detected total weight and a weight fraction of the detected total weight in the actual weight, wherein the correspondence is in particular a linear relationship between the weight fraction of the front weight sensors in the total weight and the weight fraction of the detected total weight in the actual weight.

8. Method according to one of the preceding claims, wherein the determination of the occupant weight is carried out as soon as possible after the start of a journey with the vehicle.

9. Method according to any one of the preceding claims 1 to 8, further comprising: - Determining the position of the vehicle seat, and - Determining corrected sensor readings using a vehicle seat position correction, based on the determined position of the vehicle seat, so that the corrected sensor readings are used to determine the occupant weight.

10. Method according to any of the preceding claims, wherein determining the occupant weight comprises determining an average value, wherein the average value is obtained from occupant weights determined over a period of time in a predetermined body position. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438- WO- PCT 21 October 2025 is formed and in particular sensor measurements taken when the predetermined body position is not present are not included in the determination of the occupant weight.

11. Method according to any of the preceding claims, wherein determining whether the body posture corresponds to the predetermined body posture further comprises: - Taking a picture of the occupant, - Determining support points of body posture in the image, and - Classifying the body posture resulting from the specified support points, the classification including in particular one or more of the following: - Classifying body posture using a specially trained classification model, - Comparing the posture with the predetermined posture and classifying the posture based on the comparison, wherein comparing the posture with the predetermined posture includes in particular calculating a measure of agreement or a measure of distance between the support points of the posture and support points of the predetermined posture.

12. An occupant weight determination device for determining the occupant weight of a vehicle occupant, in particular for use in a method according to one of the preceding claims, comprising: - a posture sensor device for detecting the posture of the occupant, - an occupant weight sensor device for reading sensor measurements, - a processing device for determining the occupant's weight from the sensor measurements, characterized in that the occupant's posture at the time of weight measurement corresponds to a predetermined posture.

13. Processing device for processing sensor measurement values, comprising a processor configured to perform the method according to any one of claims 1 to 11.

14. Computer program product comprising instructions which, when executed by a processing device, cause the processing device to execute the method according to any one of claims 1 to 11. Simi Reality Motion Systems GmbH ZF Automotive Germany GmbH 301438-WO-PCT October 21, 2025 15. Computer-readable storage medium comprising instructions which, when executed by a processing device, cause it to execute the method according to any one of claims 1 to 11.

Images