Seat belt use detection based on relative movement

Abstract

The method and system presented in this paper implement seatbelt use detection based on relative movement. Image data of vehicle occupants is received over time. This image data is then fed into a machine learning model or other module to determine whether the relative movement between one or more parts of the occupant and the corresponding part of the seatbelt is less than a threshold amount. If the relative movement is less than the threshold amount, a seatbelt misuse indication (e.g., dummy seatbelt) is output to the vehicle component. By indicating seatbelt misuse based on the relative movement between the occupant and the seatbelt, seatbelt use detection can be improved.

Description

Background Technology

[0001] Seatbelt use detection systems are already implemented in a wide variety of vehicles. For example, many vehicles have detection sensors built into the seatbelt buckles that trigger an alarm when the vehicle is traveling without a seatbelt clip inserted. Occupants placing a fake seatbelt clip (e.g., not attached to the seatbelt) into the buckle could fool such systems.

[0002] Recently, advanced systems for detecting seatbelt misuse have been developed. For example, systems have been developed to determine whether a seatbelt crosses an occupant's chest. These systems can also be fooled by occupants wearing shirts with seatbelt patterns (e.g., by the straps crossing their chests). Summary of the Invention

[0003] This document relates to techniques and hardware for seatbelt use detection based on relative movement. Techniques may include one or more methods for seatbelt use detection based on relative movement. Hardware may include means (e.g., one or more processors) for performing seatbelt use detection based on relative movement. Hardware may also include a computer-readable medium (e.g., a non-transitory storage medium) including instructions that, when executed, cause one or more associated systems to perform seatbelt use detection based on relative movement.

[0004] Some aspects described below include a method. The method includes receiving sensor data from one or more sensors that indicate properties of the interior of a vehicle. The method also includes determining, based on the sensor data, that a seatbelt is crossing the chest of a vehicle occupant. The method further includes, based on the sensor data, tracking over time one or more occupant portions and one or more seatbelt portions. The seatbelt portions correspond to corresponding portions within the occupant portions. The method also includes, based on the tracking, determining relative movement between the occupant portions and their corresponding seatbelt portions. The method further includes, based on the relative movement, establishing that the seatbelt is a spurious seatbelt, and outputting an indication of a spurious seatbelt to the vehicle system of the vehicle.

[0005] Some aspects described below include another method. This other method includes receiving image data over time from one or more cameras having a field of view that includes the occupants of a vehicle. This other method also includes feeding at least a portion of the image data into a machine learning model trained to find relative movement between one or more occupant portions and a corresponding seatbelt portion of a seatbelt corresponding to an occupant portion. This other method further includes receiving from the machine learning model an indication of relative movement that the seatbelt is a dummy seatbelt, and outputting the dummy seatbelt indication to the vehicle system of the vehicle.

[0006] Some aspects described below include one or more components and / or one or more systems, including one or more processors configured to perform the methods described above, parts thereof, and / or other methods, processes, and techniques described below. Some aspects described below also include a computer-readable medium (e.g., a non-transitory storage medium) including instructions that, when executed (e.g., by a processor of the component and / or system), implement the methods described above, parts thereof, and / or other methods, processes, and techniques described below.

[0007] This invention provides a simplified concept for seatbelt usage detection based on relative movement, which is further described in the detailed description and accompanying drawings. This invention is not intended to identify essential features of the claimed subject matter, nor is it intended to define the scope of the claimed subject matter. Attached Figure Description

[0008] The following figures describe a system and method for detecting seatbelt use based on relative movement. Some of the same reference numerals are used throughout the figures to refer to examples of similar objects or features and components.

[0009] Figure 1 An example environment and an example process for seatbelt use detection based on relative movement are shown in this disclosure.

[0010] Figure 2 An example of a vehicle system configured for seatbelt use detection based on relative movement, according to this disclosure, is shown.

[0011] Figure 3 An example flow for generating a seatbelt misuse instruction according to this disclosure is shown.

[0012] Figure 4 Another example process for generating a seatbelt misuse instruction according to this disclosure is shown.

[0013] Figure 5 An example of an environment in which seatbelt usage instructions can be generated according to this disclosure is shown.

[0014] Figure 6 An example of an environment in which a seatbelt misuse indication can be generated according to this disclosure is shown.

[0015] Figure 7 An example method for detecting seatbelt use based on relative movement according to this disclosure is shown.

[0016] Figure 8Another example method for seat belt usage detection based on relative movement according to this disclosure is shown. Detailed Implementation

[0017] Overview

[0018] Many seatbelt detection systems can be fooled into determining that an occupant is properly seatbeltted when they are not. For example, some vision-based systems can be fooled by occupants wearing shirts printed with seatbelt markings. Deceiving such systems (e.g., wearing a seatbelt shirt instead of properly using the seatbelt) is not only illegal in many cases, but also unsafe.

[0019] The method and system presented in this paper implement seatbelt use detection based on relative movement. Image data of vehicle occupants is received over time. This image data is then fed into a machine learning model or other module to determine whether the relative movement between one or more parts of the occupant and the corresponding part of the seatbelt is less than a threshold amount. If the relative movement is less than the threshold amount, a seatbelt misuse indication is output to the vehicle component. By indicating seatbelt misuse based on the relative movement between the occupant and the seatbelt, seatbelt use detection can be improved. Improved seatbelt use detection can enhance safety and legitimacy in many environments.

[0020] Example Environment

[0021] Figure 1 An example environment 100 in which seatbelt usage detection based on relative movement is shown. Figure 1 An example flow 102 for seatbelt use detection based on relative movement is also shown. The example environment 100 is within a vehicle 104 (not shown). The vehicle 104 can be any type of object utilizing seatbelts (taxi, car, truck, motorcycle, electric bicycle, boat, air transport, etc.). The vehicle 104 includes a vehicle system 106 configured to perform seatbelt use detection based on relative movement.

[0022] The vehicle system 106 includes a motion tracking module 108 that receives image data and tracks the position of an occupant 110 and the seat belt 112 associated with the occupant 110 (e.g., corresponding to the seat where the occupant 110 is seated). Specifically, the motion tracking module 108 tracks one or more occupant sections 114 and one or more seat belt sections 116 corresponding to the respective occupant sections 114.

[0023] For example, as shown, one of the occupant portions 114 (e.g., occupant portion 114-1) may correspond to the shoulder of occupant 110 (left or right shoulder, depending on the configuration of the seat belt 112, seating position, whether the vehicle 104 is right-hand drive or left-hand drive, etc.), and the corresponding seat belt portion 116 (e.g., seat belt portion 116-1) may be a portion of the seat belt 112 that passes through that shoulder. Another occupant portion 114 may correspond to the hip of occupant 110 (left or right hip, depending on the configuration of the seat belt 112), and the corresponding seat belt portion 116 may be a portion of the seat belt 112 that passes through that hip.

[0024] Although illustrated as being located on one side of the seatbelt portion 116, the occupant portion 114 may initially be located on either side of the corresponding seatbelt portion 116 (e.g., partially around it) or as the occupant 110 moves. When the occupant 110 moves, the seatbelt portion 116 may move below the corresponding seatbelt portion (if it was not initially configured to be below or around the seatbelt portion 116). Furthermore, the occupant portion 114 may be located on the opposite side of the seatbelt portion 116. Those skilled in the art will recognize that any number of techniques can be used to isolate the occupant portion 114 and its corresponding seatbelt portion 116 from the image data.

[0025] The motion tracking module 108 is configured to track the relative movement or motion between the occupant portion 114 and its corresponding seatbelt portion 116. Any number of known techniques can be used to determine the relative movement (e.g., pixel offset, point tracking, tracking, point tracking, centroid tracking).

[0026] In response to determining that relative movement between one or more occupant / seatbelt portion pairs does not indicate a genuine seatbelt, the movement tracking module 108 can generate a seatbelt misuse indication 118, which indicates that seatbelt 112 is misused or that seatbelt 112 is dummy. For example, if seatbelt 112 is dummy (e.g., it is printed on a shirt), then seatbelt portion 116 may move together with occupant portion 114. However, if seatbelt 112 is genuine, then occupant portions 114 may move independently of their corresponding seatbelt portions 116.

[0027] Any number of techniques can be used to determine when a seatbelt misuse indication 118 is generated. For example, the movement tracking module 108 may determine that the average relative movement between some or all of the occupant / seatbelt sections is less than a threshold, determine that a single relative movement between one of the occupant / seatbelt sections is less than a threshold (or another value), determine that the sum of the relative movements between some or all of the occupant / seatbelt sections is less than a threshold, or any other way of thresholding the amount of movement between occupant 110 and seatbelt 112.

[0028] Seatbelt misuse indication 118 can be received by vehicle component 120, which is configured to issue an alarm (e.g., visual, audible, tactile) to occupant 110 or the driver of vehicle 104 (if occupant 110 is not the driver) indicating that occupant 110 is not properly fastening their seatbelt. Furthermore, vehicle component 120 can send a notification to an external system that can warn a third party (e.g., parent, insurance company, law enforcement) of seatbelt misuse. Vehicle component 120 can also receive seatbelt use indications from motion tracking module 108. Seatbelt use indications may prevent vehicle component 120 from performing any additional actions (e.g., issuing an alarm).

[0029] By tracking the relative movement / motion between the occupant section 114 and its corresponding seatbelt section 116, the vehicle system 106 is able to detect fake seatbelts that may deceive conventional seatbelt detection systems (e.g., those fake seatbelts printed on shirts). Doing so can improve the safety of the occupant (e.g., occupant 110) by increasing the correct use of the seatbelt (e.g., because for occupant 110, an alarm is worse than fastening the seatbelt 112).

[0030] Example System

[0031] Figure 2 An example of a transportation system 106 is shown. Components of the transportation system 106 may be arranged within or on a vehicle 104 at any location. The transportation system 106 may include at least one processor 200, a computer-readable storage medium 202 (e.g., a medium, a media, or multiple media), and a transportation component 120. These components are operatively and / or communicatively coupled via a link 204.

[0032] Processor 200 (e.g., application processor, microprocessor, digital signal processor (DSP), controller) is coupled to computer-readable storage medium 202 via link 204 and executes computer-readable instructions (e.g., code) stored in computer-readable storage medium 202 (e.g., non-transient storage device such as hard disk drive, solid-state drive (SSD), flash memory, read-only memory (ROM)) to implement or otherwise cause motion tracking module 108 (or a portion thereof) to perform the techniques described herein. Although shown within computer-readable storage medium 202, motion tracking module 108 may be a separate component (e.g., a dedicated computer-readable storage medium having instructions that execute on dedicated hardware such as a dedicated processor, pre-programmed field-programmable gate array (FPGA), system-on-a-chip (SOC), etc.). The processor 200 and the computer-readable storage medium 202 can be any number of components, including multiple components distributed throughout the vehicle 104, remote from the vehicle 104, dedicated to or shared with other components, modules or systems of the vehicle 104, and / or configured in a manner different from that shown in the figures without departing from the scope of this disclosure.

[0033] The computer-readable storage medium 202 also contains sensor data 206 generated by one or more sensors or one or more types of sensors (not shown), which may be located locally or remotely from the vehicle system 106. The sensor data 206 indicates or otherwise makes it possible to determine information that can be used to perform the techniques described herein. For example, the sensors may generate sensor data 206 indicating aspects that can be used to determine when to generate the seatbelt misuse indication 118. In some implementations, the sensor data 206 may originate from a remote source (e.g., via link 204). The vehicle system 106 may include a communication system (not shown) for receiving the sensor data 206 from the remote source.

[0034] Vehicle component 120 includes one or more systems or components communicatively coupled to motion tracking module 108 and configured to perform one or more vehicle functions based on a received seatbelt misuse indication 118. For example, vehicle component 120 may include an advanced driver assistance system (ADAS) or autonomous driving system that can alter operation (e.g., disallow autonomous vehicle control) based on seatbelt misuse indication 118. In another example, vehicle component 120 may include an alarm system that provides visual, auditory, and / or tactile alerts to occupants based on seatbelt misuse indication 118. Vehicle component 120 is communicatively coupled to motion tracking module 108 via link 204. Although shown as separate components, motion tracking module 108 can be part of vehicle component 120 and vice versa.

[0035] Example Process

[0036] Figure 3 This is an example process 300 for detecting seatbelt use based on relative movement. Example process 300 can be implemented in any of the previously described environments and by any of the previously described systems or components. For example, example process 300 can be implemented in example environment 100 and / or by vehicle system 106. Example process 300 can also be implemented in other environments by other systems or components and utilizing other processes or technologies. Example process 300 can be implemented by any number of entities. The order of operations shown and / or described is not intended to be construed as limiting and can be rearranged without departing from the scope of this disclosure. Furthermore, any number of operations can be combined with any other number of operations to implement the example process or alternative processes.

[0037] Example process 300 begins by receiving sensor data 206 from motion tracking module 108. Sensor data 206 may be camera or image data (e.g., visible image data, infrared image data) of occupant 110 from one or more image sensors. Seatbelt detection module 302 of motion tracking module 108 identifies and tracks seatbelt portion 116. Seatbelt portion 116 may correspond to any portion of seatbelt 112. For example, seatbelt detection module 302 may identify and track seatbelt portion 116-1 near the shoulder of occupant 110 and another seatbelt portion 116-2 near the hip of occupant 110. In another example, seatbelt portion 116 may include a portion from the shoulder of occupant 110 to the B-pillar of vehicle 104 and / or a portion from the hip of occupant 110 to the seatbelt receiver corresponding to seatbelt 112. Furthermore, different seatbelt portions 116 may be used for different steps of seatbelt use detection based on relative movement.

[0038] At decision 304, the seatbelt detection module 302 determines whether the seatbelt 112 crosses the chest of the occupant 110. For example, if seatbelt portions 116-1 and 116-2 are not found, decision 304 can be "No". The seatbelt detection module 302 may alternatively or additionally search for another seatbelt portion 116-3 located in the middle of the occupant 110's chest. In some implementations, the seatbelt detection module 302 may determine whether the strap is crossing the occupant 110's chest, thus indicating that the seatbelt is crossing it. If seatbelt portion 116-3 is not found, decision 304 can be "No".

[0039] If it is determined that the seatbelt 112 does not cross the chest of the occupant 110 (e.g., the output of decision 304 is "No"), the seatbelt detection module 302 may generate or otherwise cause a seatbelt misuse indication 122. However, if it is determined that the seatbelt 112 crosses the chest of the occupant 110 (e.g., the output of decision 304 is "Yes"), the seatbelt detection module 302 may send or otherwise indicate the position of the seatbelt portion 116 to the relative movement module 306.

[0040] The relative movement module 306 determines the position of the occupant portion 114 corresponding to the seat belt portion 116 and tracks the movement of the occupant portion 114. For example, the relative movement module 306 can track occupant portion 114-1 and / or other occupant portions. The relative movement module 306 also compares the movement of the occupant portions 114 relative to their corresponding seat belt portions 116. For example, the relative movement module 306 can determine the relative movement between occupant portion 114-1 and seat belt portion 116-1.

[0041] At decision 308, relative movement module 306 determines whether the relative movement is less than a threshold. For example, the decision may be based on a single relative movement (e.g., only one seatbelt portion 116 and occupant portion 114) or a combination of relative movements (e.g., average, median, sum). If the relative movement is greater than (or equal to, depending on the implementation) the threshold, decision 308 can be "No," thus indicating the correct seatbelt usage instruction (e.g., seatbelt usage instruction 310). In some implementations, a "No" output from decision 308 may cause relative movement module 306 to avoid outputting an instruction. If the relative movement is less than (or equal to, depending on the implementation) the threshold, decision 308 can be "Yes," thus indicating or identifying seatbelt 112 misuse or that seatbelt 112 is spurious. A "Yes" decision may cause relative movement module 306 to generate seatbelt misuse instruction 122.

[0042] Figure 4This is an example process 400 for detecting seatbelt use based on relative movement. Example process 400 can be implemented in any of the previously described environments and by any of the previously described systems or components. For example, example process 400 can be implemented in example environment 100 and / or by vehicle system 106. Example process 400 can also be implemented in other environments by other systems or components and utilizing other processes or technologies. Example process 400 can be implemented by any number of entities. The order of operations shown and / or described is not intended to be construed as limiting and can be rearranged without departing from the scope of this disclosure. Furthermore, any number of operations can be combined with any other number of operations to implement the example process or alternative processes.

[0043] Example flow 400 begins by receiving sensor data 206 from motion tracking module 108. Sensor data 206 may be camera or image data (e.g., visible image data, infrared image data) of occupant 110 from one or more image sensors. Relative motion machine learning (ML) model 402 tracks the relative movement between seatbelt 112 and occupant 110 and determines whether the relative movement indicates a spurious seatbelt. In some implementations, relative motion ML model 402 may identify and track occupant portion 114 and seatbelt portion 116. In other implementations, relative motion ML model 402 may treat sensor data 206 as a whole. Furthermore, motion tracking module 108 may use portion module 404 to determine which portions of sensor data 206 are input into relative motion ML model 402 (e.g., portions containing occupant portion 114 and seatbelt portion 116).

[0044] The relative movement ML model 402 can be trained using training sensor data on the movement of an occupant in a vehicle seat, both with a properly fastened seatbelt and one with a dummy seatbelt. The relative movement ML model can be configured to receive sensor data 206 and distinguish between situations indicating seatbelt misuse and normal seatbelt use. Seatbelt portion 116 can correspond to any portion of seatbelt 112. For example, seatbelt detection module 302 can identify and track seatbelt portion 116-1 near the shoulder of occupant 110 and another seatbelt portion 116-2 near the hip of occupant 110.

[0045] The relative movement ML model 402 can be implemented and / or trained using any techniques known to those skilled in the art. For example, supervised learning, unsupervised learning, and / or reinforcement learning can be used. Furthermore, the relative movement ML model 402 can be implemented as a classification model (e.g., a binary classification model with seatbelt use / misuse as an outcome), a regression model, and / or a clustering model. The relative movement ML model 402 can also be trained using deep learning techniques and / or include one or more neural networks.

[0046] Regardless of how the relative movement ML model 402 is trained and / or implemented, at decision 406 it determines whether the relative movement between occupant 110 and seatbelt 112 indicates that seatbelt 112 is a genuine seatbelt (relative movement exists or exceeds a threshold, e.g., "No" from decision 406) or a spurious seatbelt (relative movement does not exist or does not exceed a threshold, e.g., "Yes" from decision 406). If decision 406 is "No", the relative movement ML model 402 may generate or otherwise cause seatbelt use at least 310 or avoid performing additional actions. If decision 406 is "Yes", the relative movement ML model 402 may determine that the seatbelt is spurious or misused, and generate or otherwise cause the generation of seatbelt misuse indication 122.

[0047] Figure 5 Example illustration 500 shows a large relative movement (e.g., sufficient to not trigger seatbelt misuse indication 122) between occupant 110 and seatbelt 112 when seatbelt 112 is a real seatbelt and correctly routed (e.g., across the chest of occupant 110). Example illustration 500 shows occupant 110 and seatbelt 112 at times 502-1 and 502-2. It can be seen that from time 502-1 to time 502-2, occupant portion 114-1 moves relative to seatbelt portion 116-1. When seatbelt 112 is real and correctly routed, it moves slightly compared to the movement of occupant 110. In other words, at least in the plane shown, occupant 110 moves under seatbelt 112, while seatbelt 112 does not normally move. Example illustration 500 allows movement tracking module 108 to generate seatbelt use indication 310, avoid generating seatbelt misuse indication 122, or avoid performing any additional operations.

[0048] Figure 6 Example illustration 600 shows that when seatbelt 112 is a dummy seatbelt, there is little or no relative movement between occupant 110 and seatbelt 112 (e.g., low enough to trigger seatbelt misuse indication 122). Example illustration 600 shows occupant 110 and seatbelt 112 at times 602-1 and 602-2. It can be seen that from time 602-1 to time 602-2, occupant portion 114-1 moves together with seatbelt portion 116-1. When seatbelt 112 is a dummy seatbelt, it moves with the movement of occupant 110. Occupant portion 114-1 moves, but seatbelt portion 114-1 also moves. Therefore, the relative movement is small, indicating that the seatbelt is a dummy seatbelt. Example illustration 600 allows movement tracking module 108 to generate seatbelt misuse indication 122.

[0049] Example Method

[0050] Figure 7 This is an example method 700 for detecting seatbelt use based on relative movement. Example method 700 can be implemented in any of the previously described environments, by any of the previously described systems or components, and by utilizing any of the previously described processes, procedures, or techniques. For example, example method 700 can be implemented in example environment 100, by vehicle system 106, and / or by following example process 300. Example method 700 can also be implemented in other environments, by other systems or components, and by utilizing other processes, procedures, or techniques. Example method 700 can be implemented by one or more entities (e.g., motion tracking module 108). The order of operations shown and / or described is not intended to be construed as limiting and may be rearranged without departing from the scope of this disclosure. Furthermore, any number of operations can be combined with any other number of operations to implement the example process or alternative process.

[0051] At 702, sensor data indicating the interior properties of the vehicle is received from one or more sensors. For example, motion tracking module 108 may receive sensor data 206 from one or more cameras.

[0052] At point 704, sensor data is used to determine whether the seatbelt is crossing the chest of the vehicle occupant. For example, at point 304, the seatbelt detection module 302 can determine that the seatbelt 112 is crossing the chest of the occupant 110.

[0053] At 706, one or more occupant portions are tracked over time based on sensor data. For example, relative motion module 306 can track occupant portion 114 of occupant 110.

[0054] At point 708, one or more seatbelt sections are tracked over time. Each seatbelt section corresponds to a corresponding occupant section. For example, seatbelt detection module 302 can track seatbelt section 116 corresponding to occupant section 114.

[0055] At 710, the relative movement between the occupant portions and their corresponding seatbelt portions is determined based on tracking. For example, the relative movement module 306 can compare the movement of the occupant portions 114 with their corresponding seatbelt portions 116.

[0056] At point 712, the seatbelt is identified as a spurious seatbelt based on relative movement. For example, at decision 308, the relative movement module 306 can determine that the relative movement is less than a threshold amount.

[0057] At point 714, a false seatbelt instruction is output to the vehicle system of the vehicle. For example, relative movement module 306 can generate a seatbelt misuse instruction 122 for reception by vehicle component 120.

[0058] Figure 8 This is an example method 800 for detecting seatbelt use based on relative movement. Example method 800 can be implemented in any of the previously described environments, by any of the previously described systems or components, and by utilizing any of the previously described processes, procedures, or techniques. For example, example method 800 can be implemented in example environment 100, by vehicle system 106, and / or by following example process 400. Example method 800 can also be implemented in other environments, by other systems or components, and by utilizing other processes, procedures, or techniques. Example method 800 can be implemented by one or more entities (e.g., motion tracking module 108). The order of operations shown and / or described is not intended to be construed as limiting and may be rearranged without departing from the scope of this disclosure. Furthermore, any number of operations can be combined with any other number of operations to implement the example process or alternative process.

[0059] At 802, image data is received over time from one or more cameras having a field of view that includes the vehicle occupants. For example, motion tracking module 108 may receive sensor data 206 from one or more cameras within vehicle 104 having a field of view that includes the occupants 110.

[0060] At 804, at least a portion of the image data is input into a machine learning model trained to find the relative motion between one or more occupant portions and the corresponding seatbelt portions of the seatbelts corresponding to those occupant portions. For example, motion tracking module 108 may input sensor data 206, or a portion thereof, into relative motion ML model 402.

[0061] At point 806, an indication that the relative movement indicator seatbelt is a spurious seatbelt is received from the machine learning model. For example, at decision 406, the relative movement ML model 402 can determine that the relative movement indicator is a spurious seatbelt and provide the indication of this decision to the movement tracking module 108.

[0062] At point 808, a false seatbelt indication is output to the vehicle system of the vehicle. For example, motion tracking module 108 may generate a seatbelt misuse indication 122 for reception by vehicle component 120.

[0063] Further examples

[0064] Below are some additional examples of seatbelt usage detection based on relative movement.

[0065] Example 1: A method comprising: receiving sensor data from one or more sensors indicating properties of the interior of a vehicle; determining, based on the sensor data, that a seatbelt is crossing the chest of a vehicle occupant; tracking, over time, one or more occupant portions of the occupant based on the sensor data; tracking, over time, one or more seatbelt portions of the seatbelt corresponding to corresponding occupant portions; determining, based on the tracking, relative movement between the occupant portions and their corresponding seatbelt portions; identifying, based on the relative movement, that the seatbelt is a spurious seatbelt; and outputting an indication of a spurious seatbelt to a vehicle system of the vehicle.

[0066] Example 2: The method of Example 1, wherein the sensor data includes visible image data or infrared image data.

[0067] Example 3: The method of Example 1 or 2, wherein the occupant is the driver of the vehicle.

[0068] Example 4: The method of Example 1, 2 or 3, wherein determining that the seat belt is crossing the occupant's chest includes determining that the strip runs from the occupant's left shoulder to the right hip, or from the occupant's right shoulder to the left hip.

[0069] Example 5: The method of any of the preceding examples, wherein the seat belt portion includes a first seat belt portion near the occupant's shoulder and a second seat belt portion near the occupant's hip.

[0070] Example 6: The method of Example 5, wherein the occupant portion includes the occupant's shoulder and hip.

[0071] Example 7: The method of Example 6, wherein the occupant portions at least partially surround their corresponding seat belt portions.

[0072] Example 8: A method of any of the preceding examples, wherein: the relative movement includes at least one of relative movement between the occupant's shoulder and a first seatbelt portion or relative movement between the occupant's hip and a second seatbelt portion; the relative movement is below a threshold; and the seatbelt is identified as a spurious seatbelt based on the relative movement being below the threshold.

[0073] Example 9: The method of Example 8 further identifies the seat belt as a spurious seat belt based on: at least one of the relative movements being below the threshold; or the average of the relative movements being below the threshold.

[0074] Example 10: A method comprising: receiving image data over time from one or more cameras having a field of view, the field of view including an occupant of a vehicle; inputting at least a portion of the image data into a machine learning model trained to find relative movement between one or more occupant portions and a corresponding seatbelt portion of a seatbelt corresponding to the occupant portion; receiving from the machine learning model an indication of the relative movement that the seatbelt is a dummy seatbelt; and outputting the dummy seatbelt indication to a vehicle system of the vehicle.

[0075] Example 11: The method of Example 10, wherein the portion of the image data corresponds to the occupant portion.

[0076] Example 12: A system comprising: one or more image sensors; and at least one processor configured to: receive image data of an occupant of a vehicle from the image sensors; determine, based on the image data, that a seatbelt is crossing the chest of the occupant; in response to determining that the seatbelt is crossing the chest of the occupant: track one or more occupant portions of the occupant over time based on the image data; track one or more seatbelt portions of the seatbelt over time based on the image data, each seatbelt portion corresponding to one occupant portion; determine, based on the tracking, relative movement between the occupant portions and their corresponding seatbelt portions; identify, based on the relative movement, that the seatbelt is a spurious seatbelt; and output an indication of the spurious seatbelt.

[0077] Example 13: The system of Example 12, wherein the tracking of the seat belt portion and the occupant portion, the determination of the relative movement, and the determination that the seat belt is a spurious seat belt are performed via a machine learning model.

[0078] Example 14: The system of Example 12 or 13, characterized in that the image data is infrared image data.

[0079] Example 15: The system of Example 12, 13 or 14, wherein the occupant is the driver of the vehicle.

[0080] Example 16: A system of any one of Examples 12 to 15, wherein determining that the seat belt is crossing the occupant's chest includes determining that the strip runs from the occupant's left shoulder to the right hip, or from the occupant's right shoulder to the left hip.

[0081] Example 17: A system of any one of Examples 12 to 16, wherein the seat belt portion includes a first seat belt portion near the occupant's shoulder and a second seat belt portion near the occupant's hip.

[0082] Example 18: A system of any of Examples 12 to 17, wherein the occupant portions at least partially surround their corresponding seat belt portions.

[0083] Example 19: A system of any one of Examples 12 to 18, wherein determining that a seatbelt is a false seatbelt includes: determining that at least one of the relative movements is below a threshold; or determining that the average of the relative movements is below a threshold.

[0084] Example 20: A system of any one of Examples 12 to 19, wherein the indication is configured to issue a visual alarm, an auditory alarm, or a tactile alarm to the occupant.

[0085] Example 21: A system comprising: at least one processor configured to perform a method of any one of Examples 1-13.

[0086] Example 22: A computer-readable storage medium including instructions that, when executed, cause at least one processor to perform any of the methods in Examples 1-13.

[0087] Example 23: A system including means for performing the method of any one of Examples 1-13.

[0088] Example 24: A method executed by the system from any of Examples 14-19.

[0089] Example 25: Methods included by the instructions in Example 20.

[0090] Conclusion

[0091] While various embodiments of the present disclosure have been described in the foregoing description and illustrated in the accompanying drawings, it should be understood that the present disclosure is not limited thereto, but can be practiced in various ways within the scope of the following claims. It will be apparent from the foregoing description that various modifications can be made without departing from the spirit and scope of the present disclosure as defined by the following claims.

[0092] Unless the context clearly specifies otherwise, the use of "or" and grammatically related terms indicates an unrestricted, non-exclusive alternative. As used herein, the phrase referring to "at least one" of a list of items means any combination of those items, including a single member. For example, "at least one of a, b, or c" is intended to cover: a, b, c, ab, ac, bc, and abc, as well as any combination with multiple identical elements (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc, and ccc, or any other ordering of a, b, and c).

Claims

1. A method comprising: Receive sensor data from one or more sensors, the sensor data indicating the internal properties of the vehicle; Based on the sensor data, it is determined that the seatbelt is crossing the chest of the occupant of the vehicle; Based on the sensor data, one or more occupant portions of the occupant are tracked over time; Based on the sensor data, one or more seat belt portions of the seat belt are tracked over time, the seat belt portions corresponding to corresponding occupant portions; Based on the tracking, the relative movement between the occupant portions and their corresponding seat belt portions is determined; Based on the fact that the relative movement is below the threshold, the seat belt is determined to be a fake seat belt; as well as The dummy seatbelt instruction is output to the vehicle system of the vehicle.

2. The method as described in claim 1, characterized in that, The sensor data includes visible image data or infrared image data.

3. The method as described in claim 1, characterized in that, The occupant is the driver of the vehicle.

4. The method as described in claim 1, characterized in that, Determining that the seatbelt is crossing the occupant's chest includes determining that the strip runs from the occupant's left shoulder to their right hip, or from their right shoulder to their left hip.

5. The method as described in claim 1, characterized in that, The seat belt includes a first seat belt portion and a second seat belt portion, the first seat belt portion being close to the occupant's shoulder and the second seat belt portion being close to the occupant's hip.

6. The method as described in claim 5, characterized in that, The occupant portion includes the occupant's shoulder and hip.

7. The method as described in claim 6, characterized in that, The occupant portions at least partially surround their corresponding seatbelt portions.

8. The method as described in claim 1, characterized in that: The relative movement includes at least one of relative movement between the occupant's shoulder and the first seatbelt portion or relative movement between the occupant's hip and the second seatbelt portion.

9. The method as described in claim 8, characterized in that, The determination that the seat belt is a counterfeit seat belt is further based on: At least one of the relative movements is below the threshold; or The average value of the relative movement is lower than the threshold.

10. A method comprising: Image data is received over time from one or more cameras having a field of view, which includes the occupants of the vehicle; At least a portion of the image data is input into a machine learning model, which is trained to find the relative movement between one or more occupant portions and corresponding seat belt portions of the seat belts corresponding to the occupant portions; Receive from the machine learning model an indication that the relative movement indicates the seatbelt is a sham when the relative movement is below a threshold; and The dummy seatbelt instruction is output to the vehicle system of the vehicle.

11. The method as described in claim 10, characterized in that, The portion of the image data corresponds to the occupant portion.

12. A system comprising: One or more image sensors; as well as At least one processor, said at least one processor being configured to: Image data of the vehicle's occupants are received from the image sensor; Based on the image data, it was determined that the seatbelt was crossing the occupant's chest; In response to determining that the seatbelt is crossing the occupant's chest: Based on the image data, one or more occupant portions of the occupant are tracked over time; Based on the image data, one or more seat belt portions of the seat belt are tracked over time, each of the seat belt portions corresponding to one of the occupant portions; Based on the tracking, the relative movement between the occupant portions and their corresponding seat belt portions is determined; Based on the fact that the relative movement is below the threshold, the seat belt is determined to be a fake seat belt; as well as Output the indication of the dummy seatbelt.

13. The system as described in claim 12, characterized in that, The tracking of the seat belt portion and the occupant portion, the determination of the relative movement, and the identification of the seat belt as a spurious seat belt are performed via a machine learning model.

14. The system as described in claim 12, characterized in that, The image data is infrared image data.

15. The system as described in claim 12, characterized in that, The occupant is the driver of the vehicle.

16. The system as claimed in claim 12, characterized in that, Determining that the seatbelt is crossing the occupant's chest includes determining that the strip runs from the occupant's left shoulder to their right hip, or from their right shoulder to their left hip.

17. The system as claimed in claim 12, characterized in that, The seat belt includes a first seat belt portion and a second seat belt portion, the first seat belt portion being close to the occupant's shoulder and the second seat belt portion being close to the occupant's hip.

18. The system as claimed in claim 12, characterized in that, The occupant portions at least partially surround their corresponding seatbelt portions.

19. The system as claimed in claim 12, characterized in that, The determination that the seat belt is a dummy seat belt includes: Determine that at least one of the relative movements is below a threshold; or The average value of the relative movement is determined to be below a threshold.

20. The system as claimed in claim 12, characterized in that, The indication is configured to issue visual, auditory, or tactile alarms to the occupant.

Images