Method and device for detecting handheld position on steering wheel

Abstract

The present invention provides a method and apparatus for identifying a gripping position on a steering wheel, where the gripping position may include a gripping gesture such as gripping the steering wheel with a single hand, gripping the steering wheel with two hands, or releasing the steering wheel from the hand, where a sensor array comprising a plurality of antennas is provided on the steering wheel. The method comprises the following steps: detecting the electromagnetic radiation absorption amount of the plurality of antennas influenced by a hand by using the sensor array; and executing mode recognition on the electromagnetic radiation absorption amount, wherein the mode recognition result indicates the handheld position.

Description

Technical Field

[0001] This disclosure relates to the field of autonomous driving technology, and in particular to hand grip position detection technology for detecting hand grip posture on a steering wheel. Background Technology

[0002] In modern automotive design, the steering wheel is not only a basic driving tool but also a core component for controlling the vehicle's direction. Many drivers often take their hands off the steering wheel or use only one hand to control the vehicle. Clearly, driving with one hand on the steering wheel reduces vehicle control safety and reliability compared to using both hands. Furthermore, if the driver takes both hands off the steering wheel, they may not be able to take control in time in a dangerous situation; therefore, driving with the steering wheel off the wheel is far more dangerous. Thus, it is necessary to identify the driver's hand posture or manner of holding the steering wheel and implement corresponding subsequent controls, such as enhancing the assisted control functions of autonomous driving systems.

[0003] In existing technologies, capacitive sensors are typically used to detect the driver's hand position on the steering wheel to determine whether the steering wheel is off-hand or being held with both hands. However, the detection results are inaccurate in some special cases. Furthermore, existing technologies have also proposed using in-vehicle cameras to detect the driver's hand posture (single-hand or two-handed grip on the steering wheel), but this method still has some drawbacks, including: limitations due to lighting conditions (the performance of the in-vehicle camera may be affected by lighting conditions, such as at night or when driving against the light, the camera may have difficulty accurately recognizing the driver's hand movements); occlusion issues (if the steering wheel or the driver's hands are obstructed by other objects (such as clothing, gloves, or other obstructions), the camera may not be able to accurately detect the hand position); and cost issues (installing and maintaining an in-vehicle camera system may increase the cost of the vehicle). In addition, processing and analyzing video data requires significant computing power, which may further increase costs. Therefore, there is a need to propose a more efficient, reliable, and cost-effective hand posture recognition technology. Summary of the Invention

[0004] This invention proposes a method to identify the driver's hand position on the steering wheel—for example, whether it is gripping the steering wheel with one hand, without gripping, or with both hands—by monitoring the impact of different hand grip positions on the electromagnetic wave emission energy of a sensor array integrated on the steering wheel; furthermore, it can also identify whether the driver is gripping the steering wheel with their left or right hand. The vehicle's infotainment system can then perform corresponding actions based on the identification results. Attached Figure Description

[0005] Figure 1A-1D A schematic diagram of a steering wheel with a sensor array and a hand grip posture is shown;

[0006] Figure 2 A schematic diagram of a hand grip detection device according to the present invention is shown;

[0007] Figure 3 A schematic diagram of the grip arbitration used in this invention is shown.

[0008] Figure 4 The flowchart of a method for recognizing the hand grip posture on a steering wheel according to an embodiment of the present invention is shown.

[0009] Figure 5 The hand grip arbitration process according to an example of the present invention is shown. Detailed Implementation

[0010] Before explaining any embodiments of the invention in detail, it should be understood that the application of the invention is not limited to the construction details set forth in the following description or shown in the accompanying drawings. The invention can have other embodiments and can be practiced or implemented in various ways.

[0011] According to embodiments of the present invention, sensors integrated into the steering wheel are used to identify how the driver grips the steering wheel. By performing pattern recognition processing on the electromagnetic wave absorption data collected by the sensor array, the driver's hand position on the steering wheel can be accurately distinguished, such as whether the driver is using one or both hands, and further, whether the driver is using their left or right hand. In other embodiments of the present invention, the specific position of the hand on the steering wheel can be further identified, such as the left hand at the 10 o'clock position.

[0012] Figure 1A A schematic diagram of a steering wheel according to an example of the present invention is shown. Figure 1A As shown, a sensor array is provided on the steering wheel. The sensor array consists of multiple sensors S. According to one example of the invention, the sensors can be electromagnetic energy absorption rate (SAR) sensors, wherein each sensor includes an antenna and a processor communicating with the antenna. Figure 1AIn the illustrated example, the sensor array includes five SAR sensors S1 to S5. Sensor S1 includes an antenna or antenna patch ANT1 arranged around the circumference of the steering wheel. For ease of antenna placement, ANT1 can be placed on the surface of the steering wheel facing the driver. However, the invention is not limited to this; for example, antenna ANT1 can also be placed on the back of the steering wheel away from the driver, or on the outermost circumference of the steering wheel. To better detect the driver's hand contact with the steering wheel, the invention preferably places antenna ANT1 on the front or outermost circumference of the steering wheel. The antennas ANT2 to ANT5 of the other four sensors S2 to S5 are arranged at intervals on the steering wheel frame, for example, at 90-degree intervals at the 12 o'clock, 3 o'clock, 6 o'clock, and 9 o'clock positions. Each of antennas ANT2 to ANT5 can be arranged to wrap around the steering wheel frame half a circle or a full circle, so that the arrangement direction of antennas ANT2 to ANT5 is perpendicular to the arrangement direction of antenna ANT1. It should be noted that the arrangement direction of antennas ANT2 to ANT5 is not limited to being perpendicular to the arrangement direction of antenna ANT1, as long as they are not in the same direction. Furthermore, although four sensors S2 to S5 are shown in the direction of the steering wheel winding, the invention is not limited to this, and three or more sensors can also be used.

[0013] Each sensor's antennas ANT1 to ANT5 communicate with their respective processors (not shown in the figure). According to the SAR sensor design, the processor (also called the SAR chip) is configured to monitor and output the absorption rate R of electromagnetic wave energy absorbed or consumed by external sources, such as the human body, by reading the electromagnetic radiation capacitance value of its antennas; this is also referred to herein as the electromagnetic radiation quantity R. This invention can be implemented using existing SAR sensors, and the SAR sensor can be integrated into the steering wheel. For example, in one implementation, after arranging antennas ANT1 to ANT5 on the steering wheel as described above, the SAR chip of the SAR sensor can be placed in the inner aperture of the steering wheel frame. The SAR chip controls the antennas to emit signals with a predetermined radiation power and calculates the electromagnetic radiation quantity R based on the antenna's electromagnetic radiation capacitance value. Here, the predetermined radiation power of antennas ANT1 and ANT2 to ANT5 can be the same or different.

[0014] According to the design of SAR sensors, when a human body, such as a hand, approaches or touches the antenna, it changes the electromagnetic radiation capacitance value of the antenna. The SAR chip is designed to determine the amount of electromagnetic radiation R absorbed by the outside world, such as a hand, by monitoring the electromagnetic radiation capacitance value of the antenna, and outputs this amount of radiation R as a measurement of the human body's radiation absorption rate. This invention uses the amount of electromagnetic radiation absorbed R when the driver's hand is gripping the steering wheel to detect the driver's hand grip posture. Figure 2A schematic diagram of a hand grip detection device according to the present invention is shown.

[0015] like Figure 2 As shown, the hand grip posture detection device 100 includes a sensor array formed by multiple sensors, such as Figure 2 The sensors S1 to S5 and the controller CTR are shown. As previously disclosed, the antenna ANT1 of sensor S1 is arranged along the circumference of the steering wheel, while the antennas ANT2 to ANT5 of the other sensors are arranged at certain intervals along the diameter of the steering wheel's frame. The controller CTR is communicatively coupled to the sensor arrays S1 to S5, for example, via an in-vehicle bus (BUS), to receive measurement signals R1 to R5 of radiation absorption from the SAR chips of these sensors. The controller CTR performs pattern recognition processing on the measurement signals R1 to R5 to generate an indication signal of the hand grip posture on the steering wheel.

[0016] According to one embodiment of the present invention, such as Figure 2 As shown, the controller CTR calls the trained grip posture decision model GM to process the electromagnetic radiation absorption R1 detected by sensor S1 and the electromagnetic radiation absorption R2-R5 of sensors S2-S5, thereby generating a prediction of the hand grip posture. Due to the specific arrangement of sensors S1-S5 in this example, when the driver's hands grip the steering wheel, it will have different effects on the radiation capacitance values ​​of the antennas ANT2-ANT5 of sensors S2-S5 located at different positions. By using machine learning models such as neural networks (NNs) to learn the pattern features composed of the measured human radiation absorption R2-R5 and the output R1 from sensor S1, a prediction of the hand grip posture on the steering wheel can be generated. Here, the grip posture decision model GM can be implemented using neural network architectures known in the art, such as fully connected networks or convolutional neural networks; or in other implementations, the grip posture decision model GM can also be implemented using machine learning algorithms such as the K-means algorithm or the Gaussian mixture model algorithm.

[0017] During vehicle driving, whether the driver holds the steering wheel with their left or right hand has a significant impact on driving performance. For example, when using the left hand to hold the steering wheel in a right-hand drive vehicle, the driver can more easily operate the center console and other controls, and may feel more natural and have better control when making left turns. In a left-hand drive vehicle, holding the steering wheel with the right hand makes it easier for the driver to operate other vehicle controls, and if the driver's right hand is stronger, they may feel more comfortable and confident when controlling the steering wheel. Therefore, detecting whether the driver is holding the steering wheel with one hand, and specifically which hand is holding it, becomes very important. According to the above embodiments of the present invention, during the training and learning process of the grip posture decision model GM, a large number of data samples can be used. These data samples can come from different vehicles and different users to fully reflect the characteristics of different steering wheels and different users (reflecting hand size, grip strength, etc.). Furthermore, the collected data samples include electromagnetic radiation absorption values ​​R1-R5 collected by sensors S1-S5 under various possible scenarios, including both hands gripping the steering wheel, left hand gripping the steering wheel, right hand gripping the steering wheel, and hands off the steering wheel. A machine learning model is trained to learn from and make decisions on these sample datasets to obtain a grip posture decision model GM. The output of the grip posture decision model GM can be used to indicate one of the following: both hands gripping the steering wheel, left hand gripping the steering wheel, right hand gripping the steering wheel, or hands off the steering wheel. In another embodiment of the invention, by training the grip posture decision model GM, the specific hand grip position on the steering wheel can be further predicted, such as whether the left hand is gripping at the 9 o'clock position, or whether both hands are gripping at the 9 o'clock and 3 o'clock positions. Therefore, by processing the electromagnetic radiation absorption values ​​R1-R5 collected by sensors S1-S5 using the trained grip posture decision model GM, a wider range of hand grip position information can be predicted, including not only hand grip postures such as both hands gripping, left hand gripping, right hand gripping, and hands off the steering wheel, but also the specific hand grip position prediction on the steering wheel.

[0018] In another embodiment of the present invention, the controller CTR may also make a judgment on the current grip posture by executing the grip posture arbitration rule GA based on the electromagnetic radiation absorption amount R1 to R5 collected by the sensors S1 to S5. Figure 3 A schematic diagram for hand grip arbitration is shown. Since, during vehicle driving, the hands must grip the steering wheel to a certain degree to be considered as effectively controlling the steering wheel, i.e., truly gripping it, the arbitration rules of this invention set a minimum electromagnetic radiation threshold absorption T for monitoring hand grip posture. MIN The minimum radiation threshold absorption T MINThis corresponds to the minimum level of contact or grip required for effective one-handed control of the steering wheel. Setting this minimum threshold absorption also avoids misjudgments caused by the influence of the hand on the sensor antenna before it contacts the steering wheel. It should be noted that in hand grip arbitration, the minimum threshold absorption T used for sensor S1... MIN1 And the minimum threshold absorption T of sensors S2-S5 MIN2 It can be set to be the same or different, for example, T MIN1 >T MIN2 This is because the arrangement of sensor antenna ANT1 is different from that of sensor antennas ANT2-ANT5. Therefore, the influence of a hand on antenna ANT1 is always greater than that on ANT2-ANT5. Thus, T can be appropriately set. MIN1 Greater than T MIN2 In addition, a maximum magnetic radiation threshold absorption value T was set to monitor hand grip posture. MAX The maximum radiation T MAX The minimum level of contact or grip required for effective control of the steering wheel with both hands. Threshold absorption amount T MIN1 T MIN2 T MAX The value can be determined based on the actual test; for example, it can be set as the average of multiple sets of measurement data. As an example, such as... Figure 3 As shown, the threshold absorption amount T MAX Set to 15, and T MIN1 With T MIN2 They are the same, both being 5.

[0019] As mentioned above Figure 1A As described, since the antenna of the SAR sensor S1 is arranged along the circumference of the steering wheel, it can collect information from all directions about the driver's contact or grip on the steering wheel with one or both hands at any position. Therefore, if the sensor S1 does not detect the absorbed radiation signal R1, or if... Figure 3 The detected radiation amount R1 is lower than the minimum threshold absorption amount T. MIN1 This indicates that the driver is not currently holding the steering wheel. Since the driver's hands are on the steering wheel, this will inevitably affect one or more of the antennas ANT2-ANT5 of sensors S2-S5. Therefore, any one of sensors S2-S5 will output a radiation absorption signal R. Thus, as shown in the figure, even if the value of R1 is greater than or equal to T... MIN1 But less than the maximum threshold absorption T MAX However, as long as none of the output signals R2-R5 of sensors S2 to S5 are greater than T MIN2 If the value of R1 is greater than the maximum threshold absorption amount T, it will still be determined that there are no hands on the steering wheel, i.e., the hands have been released. MAXHowever, none of the output signals R2-R5 can exceed T. MIN2 If this is the case, then a system malfunction can be identified, making it impossible to determine the hand grip. If so... Figure 1B As shown, when the user grips the steering wheel with both hands, the value of the radiation absorption signal R1 detected by sensor S1 will necessarily be greater than the maximum threshold absorption T. MAX Furthermore, there must be one or more output signals R greater than T among signals R2-R5. MIN2 Therefore, in this situation, it indicates that both hands are on the steering wheel.

[0020] Furthermore, when the R1 value is greater than or equal to T MIN1 And less than or equal to the threshold absorption amount T MAX Under these circumstances, if any one of the output signals R2-R5 is greater than or equal to T... MIN2 This indicates that at least one of sensors S1 and S2-S5 has detected the presence of a hand gripping the steering wheel, but because the amplitude of the output signal R1 is insufficient, it indicates that the steering wheel is currently being gripped by one hand. In another example of the present invention, the amplitude characteristics of the output signals of sensors S2-S5 can be further used to determine whether the steering wheel is gripped by the left or right hand.

[0021] As described in the previous embodiments, the four sensor antennas ANT2 to ANT5 are arranged at intervals on the steering wheel, for example, as shown in the previous embodiments. Figure 1A As shown, the sensors are evenly spaced at 90-degree intervals at the 12 o'clock, 3 o'clock, 6 o'clock, and 9 o'clock positions. It's easy to understand that the left hand is more comfortable and more accustomed to gripping the left half of the steering wheel, while the right hand is more comfortable and accustomed to gripping the right half. Therefore, for sensor S3 and the steering wheel area around it, the amplitude of the signal R3 output when gripped by the left hand will be greater than the amplitude of the signal output when gripped by the right hand. Furthermore, regardless of whether the left or right hand is gripping sensor S3, the influence on sensor S3 is the greatest, so the output R3 of sensor S3 will also be greater than the outputs of other sensors S2, S4, and S5. Similarly, for sensor S4 and the steering wheel area around it, the amplitude of the signal R4 output when gripped by the left hand will be less than the amplitude of the signal output when gripped by the right hand. Furthermore, regardless of whether the left or right hand is gripping sensor S4, the influence on sensor S4 is the greatest, so the output R4 of sensor S4 will also be greater than the outputs of other sensors S2, S3, and S5. Even for sensors S2 and S5, located in the center of the steering wheel, the signal amplitude will differ depending on whether the steering wheel is held by the left or right hand, due to user hand habits. Figure 1C and 1DSchematic diagrams of the steering wheel with the left and right hands at the grip sensor S2 are respectively shown. Generally, for left-handed users, the output signal R2 generated when holding the steering wheel in the area of the grip sensor S2 and its vicinity is greater than the signal generated when the right hand holds the steering wheel in the same area, and the amplitude of the signal R2 output by the sensor S2 is also greater than the outputs R3, R4, and R5 of other sensors; this signal characteristic also applies to the sensor S5.

[0022] Due to the different effects on the sensor antenna caused by holding different areas of the steering wheel with a single hand and different grasping degrees, in order to accurately detect the left and right hands, the present invention uses a detection reference line RL during grip arbitration. For example, as Figure 3 shown, for the sensor, the applicable reference line RL will be dynamically adjusted according to the output change of the sensor S1, for example, showing a monotonic increase or decrease. Thus, according to this example of the present invention, when the value of the current output R1 of the sensor S1 is between T MIN1 and T MAX For example, when it is 12, the output signals R2 - R5 provided by the sensors S2 - S5 are synchronously read, for example, they are 9, 4, 5, 1 respectively. It can be determined that the sensor S2 has the maximum output value of 9, so it can be preliminarily determined that the hand is held near the sensor S2. In addition, using the reference curve RL2 of the sensor S2, the corresponding reference value R0 can be determined based on the value of R1, for example, it is 8. Since R2 > R0, it can be determined that the left hand is holding the steering wheel currently; conversely, if R2 < R0, it is determined that the right hand is holding the steering wheel currently.

[0023] It should be noted here that the reference line RL used in judging the left and right hand grips is related to the position of the sensor S on the steering wheel, that is, corresponding calibration is carried out for sensors at different positions. For example, for Figure 1A the sensor 4 shown, when it is determined that the sensor output signal, for example, R4 is greater than the reference value R0 calibrated on the reference line RL4 of the sensor S4, it is determined that the right hand is holding the steering wheel currently. According to different embodiments of the present invention, corresponding reference lines RL2 - RL5 can be respectively formulated in advance for the sensors S2 - S5 and stored in the system library for the controller CTR to use when judging the hand grip. In a further embodiment of the present invention, the reference line RL can also require the user to perform calibration updates when initially using the vehicle. For example, it is required that the user perform left hand holding the steering wheel practice and right hand holding the steering wheel practice respectively at or near the sensors S2 - S5, so as to further calibrate or replace the reference lines RL2 - RL5 pre-stored in the system library. In this way, a reference curve suitable for the grip force and user habit (left or right hand) of a specific user can be generated for later arbitration use.

[0024] Figure 4The present invention illustrates a method flow for identifying hand grip posture on a steering wheel according to an embodiment of the present invention. As previously described, multiple sensors, such as S1 to S5, are arranged on the steering wheel. As an example, the antenna ANT1 of sensor S1 is arranged along the circumference of the steering wheel, while the antennas ANT2 to ANT5 of the multiple sensors S2 are arranged at certain intervals along the diameter of the steering wheel frame. In step 401, a radiation absorption measurement signal R1 is received from sensor S1, and electromagnetic radiation absorption measurement signals R2 to R5 are received from sensors S2-S5. Subsequently, in step 403, pattern recognition processing is performed on the measurement signals R1 to R5 to generate an indication signal of the hand grip position on the steering wheel. Here, the hand grip position can include both the hand grip posture on the steering wheel and the specific application position of the hand on the steering wheel. According to one embodiment of the present invention, a trained grip posture decision model GM is invoked to process the electromagnetic radiation quantity R1 and the electromagnetic radiation quantities R2-R5 of sensors S2-S5 to generate a prediction of the hand grip posture and the specific application position. The predicted hand grip posture includes two-hand grip, left-hand grip, right-hand grip, or release. In another example of the present invention, an arbitration method can also be used to identify the hand grip posture. Figure 5 The identification process flow based on this example is shown.

[0025] like Figure 5 As shown, after receiving the measurement signal R1 from sensor S1 and the measurement signals R2 to R5 from sensors S2 to S5 in step 501, the process proceeds to step 503. In step 503, it is determined whether the value of R1 is greater than or equal to the maximum threshold absorption amount T. MAX If the result is 'yes', proceed to step 505 to further determine whether there exists any output signal R′ (where R′ can represent any one of R2-R5) that is greater than or equal to T. MIN2 If such an output signal R′ exists, proceed to step 507 to determine if the steering wheel is currently being held with both hands. If step 505 determines that there is no signal greater than or equal to T... MIN2 If the output signal R′ is not found, then proceed to step 509, where a system alarm signal is generated, indicating a possible system fault and suggesting inspection, etc. If, in step 503, it is determined that the value of R1 is less than the maximum threshold absorption amount T... MAX Then proceed to step 511 to further determine whether the value of R1 is greater than or equal to the minimum threshold absorption amount T. MIN1 If it is determined that the value of R1 is still less than the minimum threshold absorption amount T MIN1 If so, proceed to step 513 to determine if the current situation is a sell-off.

[0026] If in step 511 it is determined that the value of R1 is greater than or equal to the minimum threshold absorption amount T MIN1Then proceed to step 515 to further determine whether any of the signals R2 to R5 has an output signal R′ that is greater than or equal to T. MIN2 If it is determined in step 515 that there is no value greater than or equal to T MIN2 If the output signal R′ is found, then proceed to step 513 to determine if the current action is to release the item. Conversely, if step 515 determines that there exists a signal greater than or equal to T, then proceed to step 513 to determine if the current action is to release the item. MIN2 If the output signal R′ is received, proceed to step 517. In step 517, the maximum value among signals R2 to R5 is further determined. For example, if R2>R3>R4>R5 is determined, then R2 can be determined to have the maximum output value. Therefore, in step 519, it can be determined that the sensor corresponding to the maximum value R2 is S2, and further in step 521, it is determined that the reference line RL2 corresponding to the sensor S2 is used as the reference line for the current grip posture arbitration. Therefore, in step 523, it is compared whether the maximum value R2 is above the reference line RL2. As described above, the reference line RL2 is a monotonic function of the output R1 of sensor S1. Therefore, based on the currently received R1 value, the corresponding reference point R0 can be marked in the reference line RL2. Thus, when R2>R0, it indicates that R2 is above the reference line RL2, and therefore it is determined that the current grip posture is left-handed. If R2<R0, it is determined that the current grip posture is right-handed. It should be noted that if R2 = R0, the sensor outputs R2 to R5 will continue to be collected, and the above process will be repeated until R2 is not equal to R0 in order to perform the left-hand or right-hand determination.

[0027] According to a further application of the present invention, the hand grip recognition results can be reported to the vehicle's central control system, such as the vehicle infotainment system, in real time. The vehicle infotainment system then performs corresponding operations based on the recognition results, including but not limited to: the vehicle system performing a series of response operations based on the received information. These operations may include issuing warnings of driver inattention, adjusting the operating mode of the driver assistance system, or automatically taking safety measures when an emergency is detected. Furthermore, these recognition results can also be recorded and analyzed as user data. By continuously recording the driver's steering wheel grip habits and performing data analysis, personalized driving suggestions can be provided to the driver, or the vehicle's safety system can be further optimized.

[0028] Although as mentioned above Figures 2 to 5 The embodiments according to this disclosure have been described, but the invention is not limited to the specific embodiments disclosed herein. According to another embodiment of the invention, a machine-readable medium may also be provided having machine-readable instructions stored thereon, which, when executed by a controller, cause the controller to perform any of the foregoing methods disclosed herein. Therefore, the scope of protection of this invention is defined by the appended claims.

Claims

1. A method of identifying a hand grip position on a steering wheel, wherein a sensor array comprising a plurality of antennas is provided on the steering wheel, the method comprising: detecting, with the sensor array, an amount of electromagnetic radiation absorbed by the plurality of antennas affected by a hand; performing pattern recognition on the amount of electromagnetic radiation absorbed, wherein the result of the pattern recognition indicates the hand grip position. 2.The method of claim 1, wherein the sensor array comprises at least: a first sensor comprising a first antenna arranged along a first direction; and a plurality of second sensors, each comprising a second antenna arranged along a second direction, wherein the plurality of second antennas are arranged at different locations of the steering wheel, respectively, wherein the method further comprises: obtaining a first amount of electromagnetic radiation absorbed by the first sensor in relation to the first antenna; obtaining a plurality of second amounts of electromagnetic radiation absorbed by the plurality of second sensors in relation to the plurality of second antennas. 3.The method of claim 2, wherein the first antenna is arranged along a circumferential direction of the steering wheel; and wherein the second direction is perpendicular to the circumferential direction. a single hand grip, a double hand grip, or a hand off the steering wheel.

4. The method of claim 3, wherein the hand position comprises a hand grip pose of the hand on the steering wheel, comprising: 5.The method of any one of claims 2-4, wherein performing pattern recognition comprises: processing the first amount of electromagnetic radiation absorbed and the plurality of second amounts of electromagnetic radiation absorbed with a trained machine learning model to generate a prediction of the hand grip position. 6.The method of claim 4, wherein performing pattern recognition comprises: determining a double hand grip when the first amount of electromagnetic radiation absorbed is greater than or equal to a first radiation threshold and any one of the plurality of second amounts of electromagnetic radiation absorbed is greater than a second threshold; and determining a single hand grip when the first amount of electromagnetic radiation absorbed is less than the first radiation threshold and greater than or equal to a third radiation threshold, and any one of the plurality of second amounts of electromagnetic radiation absorbed is greater than the second threshold. 7.The method of claim 6, wherein performing pattern recognition comprises: determining a second sensor corresponding to a second amount of electromagnetic radiation absorbed having a maximum value among the plurality of second amounts of electromagnetic radiation absorbed when the first amount of electromagnetic radiation absorbed is less than the first radiation threshold and greater than or equal to the third radiation threshold; selecting a reference line associated with the corresponding second sensor; and comparing the second amount of electromagnetic radiation absorbed having the maximum value with the reference line to determine whether the current hand grip is a left hand grip or a right hand grip. 8.The method of any one of claims 2-4, wherein the plurality of second sensors comprises four sensors spaced apart by 90 degrees. 9.The method of claim 7, wherein performing pattern recognition comprises: generating an alarm signal if the first amount of electromagnetic radiation absorbed is greater than the first radiation threshold and any one of the plurality of second amounts of electromagnetic radiation absorbed is less than the second threshold. 10.A device for detecting a hand grip position on a steering wheel, comprising: ​ a sensor array on a steering wheel, the antenna array comprising a plurality of antennas for emitting electromagnetic radiation, at least one controller electrically coupled to the sensor array, the controller configured to: detect, using the sensor array, an amount of absorption of electromagnetic radiation by the plurality of antennas due to a hand; perform pattern recognition on the amount of absorption of electromagnetic radiation, wherein the pattern recognition result indicates the hand grip position.

11. The apparatus of claim 10, wherein the sensor array comprises at least: a first sensor comprising a first antenna arranged along a first direction; and a plurality of second sensors, wherein each second sensor comprises a second antenna arranged along a second direction, wherein the plurality of second antennas are arranged at different locations of the steering wheel, wherein the amount of absorption of electromagnetic radiation by the plurality of antennas comprises: a first amount of absorption of electromagnetic radiation associated with the first antenna collected by the first sensor, and a plurality of second amounts of absorption of electromagnetic radiation associated with the plurality of second antennas collected by the plurality of second sensors.

12. The apparatus of claim 11, wherein the first antenna is arranged along a circumferential direction of the steering wheel; and wherein the second direction is perpendicular to the circumferential direction. a single hand grip, a double hand grip, or a hand off the steering wheel.

13. The apparatus of claim 12, wherein said hand grip position comprises a hand grip pose of said hand on a steering wheel, comprising:

14. The apparatus of any one of claims 11-13, wherein the controller is further configured to: process the first amount of absorption of electromagnetic radiation and the plurality of second amounts of absorption of electromagnetic radiation using a trained machine learning model to generate a prediction of the hand grip position.

15. The apparatus of claim 13, wherein the controller is further configured to: determine the grip position as a double hand grip when the first amount of absorption of electromagnetic radiation is greater than or equal to a first radiation threshold and any one of the plurality of second amounts of absorption of electromagnetic radiation is greater than a second threshold; and determine the grip position as a single hand grip when the first amount of absorption of electromagnetic radiation is less than the first radiation threshold and greater than or equal to a third radiation threshold, and any one of the plurality of second amounts of absorption of electromagnetic radiation is greater than the second threshold.

16. The apparatus of claim 15, wherein the controller is further configured to: determine, when the first amount of absorption of electromagnetic radiation is less than the first radiation threshold and greater than or equal to the third radiation threshold, a second amount of absorption of electromagnetic radiation having a maximum value among the plurality of second amounts of absorption of electromagnetic radiation; select a reference line associated with the corresponding sensor; compare the second amount of absorption of electromagnetic radiation having the maximum value with the reference line to determine whether the current hand grip position is a left hand grip or a right hand grip.

17. The apparatus of any one of claims 11-13, wherein the plurality of second sensors comprises four sensors spaced apart by 90 degrees.

18. The apparatus of claim 16, wherein the controller is further configured to: generate an alert signal if the first amount of absorption of electromagnetic radiation is greater than the first radiation threshold and any one of the plurality of second amounts of absorption of electromagnetic radiation is less than the second threshold. ​ 19. The apparatus of any one of claims 10-13, wherein the controller is located in the head unit system.

20. A computer program product comprising machine readable programs, wherein the machine readable programs, when executed by a controller, implement the method of any one of claims 1-9.

21. A computer readable storage medium having stored thereon machine readable programs, wherein the machine readable programs, when executed by at least one controller, implement the method of any one of claims 1-9.

Images