A method, device, vehicle and storage medium for identifying lateral takeover actions

Abstract

The present invention discloses a method, device, vehicle and storage medium for identifying lateral takeover actions. The method includes: receiving lateral sensing data from multiple lateral control sensors through a communication link, generating a lateral sensing residual vector based on each lateral sensing data; determining the sensor fault type based on the lateral sensing residual vector; if the sensor fault type is a single point fault, generating a lateral takeover intention vector based on the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data; identifying whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and a set threshold. In the event of a single point failure in the sensor, effective identification of the lateral takeover intention and takeover action can still be achieved, avoiding the potential harm of a single sensor failure causing the system to unexpectedly exit or be unable to exit the automatic driving mode, and improving the safety level of the automatic driving system.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method, device, vehicle, and storage medium for recognizing a lateral takeover action. Background Art

[0002] The safety of autonomous driving systems is a key concern of all sectors, but most of the focus is on the safety of vehicle driving and the driver's passive takeover process, while the safety of the driver's active takeover of the autonomous driving system is rarely mentioned.

[0003] Active takeover actions can be categorized as lateral takeover and longitudinal takeover. Existing lateral takeover recognition methods primarily use steering wheel torque and hands-off status to identify the driver's lateral takeover intention. When the steering wheel torque exceeds a set threshold while the driver is holding the steering wheel, the driver is deemed to be attempting a lateral takeover.

[0004] To improve the safety of active takeover, current efforts are primarily focused on improving the accuracy of identifying the driver's takeover intention, thereby avoiding misidentification of the driver's takeover intention, which can lead to an inability to exit the autonomous driving mode or an unexpected exit from the autonomous driving mode. However, existing methods do not consider the impact of takeover sensor failures and communication link failures on system functionality and safety. When the system has these failures, there is still the potential for misjudging the driver's active takeover, resulting in an unexpected exit from the autonomous driving mode, or missing the driver's active takeover, resulting in an inability to exit the autonomous driving mode. Summary of the Invention

[0005] The present invention provides a method for identifying lateral takeover actions to address the problem that the existing technology lacks a fault diagnosis and processing mechanism, and there is a potential hazard of unexpected exit or inability to exit the autonomous driving mode, thereby avoiding unacceptable harm to the system caused by a single sensor failure and improving the safety level of the autonomous driving system.

[0006] According to one aspect of the present invention, a method for identifying a lateral takeover action is provided, comprising:

[0007] receiving lateral sensing data from a plurality of lateral control sensors via a communication link;

[0008] generating a lateral sensing residual vector according to each of the lateral sensing data;

[0009] determining a sensor fault type according to the lateral sensing residual vector;

[0010] If the sensor fault type is a single point fault, generating a lateral takeover intention vector according to the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data;

[0011] It is determined whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action according to the lateral takeover intention vector and at least one corresponding set threshold.

[0012] According to another aspect of the present invention, a device for recognizing a lateral takeover action is provided, comprising:

[0013] a data receiving module, configured to receive lateral sensing data from a plurality of lateral control sensors via a two-way communication link;

[0014] a residual vector determination module, configured to generate a lateral sensing residual vector according to each of the lateral sensing data;

[0015] a fault type determination module, configured to determine a sensor fault type based on the lateral sensing residual vector;

[0016] an intention vector determination module, configured to generate a lateral takeover intention vector based on the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data if the sensor fault type is a single point fault;

[0017] The takeover action recognition module is used to identify whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and at least one corresponding set threshold.

[0018] According to another aspect of the present invention, there is provided a vehicle, comprising:

[0019] at least one controller; and

[0020] a memory in communication with the at least one controller; wherein,

[0021] The memory stores a computer program that can be executed by the at least one controller, and the computer program is executed by the at least one controller to enable the at least one controller to execute the lateral takeover action identification method described in any embodiment of the present invention.

[0022] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a controller to implement the lateral takeover action recognition method described in any embodiment of the present invention when executed.

[0023] The technical solution of the embodiment of the present invention receives lateral sensing data from multiple lateral control sensors through a communication link, generates a lateral sensing residual vector based on each lateral sensing data; determines the sensor fault type based on the lateral sensing residual vector; if the sensor fault type is a single point fault, generates a lateral takeover intention vector based on the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data; and identifies whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and a set threshold. In the event of a single point failure in the sensor, effective identification of the lateral takeover intention and takeover action can still be achieved, solving the problem of the prior art lacking a fault diagnosis and processing mechanism, and the potential hazards of unexpected exit or inability to exit the autonomous driving mode, avoiding unacceptable hazards to the system caused by a single sensor failure, and improving the safety level of the autonomous driving system.

[0024] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 This is a flow chart of a method for identifying a lateral takeover action provided in Example 1 of the present invention;

[0027] Figure 2 This is a flow chart of a method for identifying a lateral takeover action provided by the second embodiment of the present invention;

[0028] Figure 3 It is a schematic diagram of the principle of generating the first residual value;

[0029] Figure 4 It is a schematic diagram of the principle of generating the second residual value;

[0030] Figure 5 It is a schematic diagram of the principle of generating the third residual value;

[0031] Figure 6 This is a fault tree analysis diagram of a method for identifying a lateral takeover action provided in the second embodiment of the present invention;

[0032] Figure 7 This is a flow chart of a method for identifying a lateral takeover action provided by Example 3 of the present invention;

[0033] Figure 8 This is a diagram of the communication architecture for horizontal takeover of action recognition.

[0034] Figure 9 This is a schematic structural diagram of a device for recognizing a lateral takeover action provided by a third embodiment of the present invention;

[0035] Figure 10 2 is a schematic structural diagram of a vehicle for implementing the method for identifying a lateral takeover action according to an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0037] It should be noted that the terms "first", "second", and "third" in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products, or apparatus.

[0038] Example 1

[0039] Figure 1 This is a flow chart of a lateral takeover action recognition method provided by the first embodiment of the present invention. This embodiment is applicable to the case of identifying whether a lateral takeover action is effective in an autonomous driving scenario. The method can be executed by a lateral takeover action recognition device, which can be implemented in the form of hardware and / or software. The lateral takeover action recognition device can be configured in a vehicle. Figure 1 As shown, the method includes:

[0040] S110 , receiving lateral sensing data from a plurality of lateral control sensors via a communication link.

[0041] The lateral takeover action recognition in the embodiments of the present invention utilizes a single-channel, dual-channel, or multi-channel communication architecture. Exemplarily, the lateral control sensors may include at least a steering wheel angle sensor, a steering gear angle sensor, a steering wheel torque sensor, and a hands-off steering wheel sensor. It is understood that the lateral control sensors may also include sensors related to the vehicle's lateral control.

[0042] Lateral sensing data refers to sensor data collected by multiple lateral control sensors and received by the controller via a communication link. Exemplarily, the lateral sensing data may include at least: steering wheel angle sensor data collected by a steering wheel angle sensor, steering gear angle sensor data collected by a steering gear angle sensor, steering wheel torque sensor data collected by a steering wheel torque sensor, and steering wheel hands-off sensor data collected by a hands-off sensor.

[0043] Specifically, lateral sensing data from a plurality of lateral control sensors are received via a communication link, and the lateral sensing data are pre-processed to obtain valid lateral sensing data.

[0044] S120 , generating a lateral sensing residual vector according to each lateral sensing data.

[0045] The lateral residual vector can be understood as a vector consisting of residual values ​​corresponding to the lateral sensor data collected by the lateral sensor. The residual value can be the residual value between lateral sensor estimation data estimated based on the lateral sensor data and the corresponding lateral sensor data, or it can be the residual value between different lateral sensor data sets. Each residual value in the lateral residual vector can represent the state of a sensor type, which can include a fault state, a takeover state, and the like.

[0046] Specifically, a residual value is determined according to each lateral sensing data, and a lateral sensing residual vector is formed by the residual values ​​corresponding to each lateral sensing data.

[0047] S130 : Determine the sensor fault type according to the lateral sensing residual vector.

[0048] Sensor failure types can include single-point failures or multiple-point failures. A single-point failure indicates that only one sensor is faulty, while a multiple-point failure indicates that at least two sensors are faulty. Single-point failure types can be further categorized as single-point failures of individual sensors. Examples include single-point failures of the steering wheel torque sensor, single-point failures of the steering wheel hands-off sensor, single-point failures of the steering wheel angle sensor, and single-point failures of the steering gear angle sensor.

[0049] Specifically, the sensor fault type may be determined to be a single-point fault or a multi-point fault according to the values ​​of the residual values ​​in the lateral sensing residual vector.

[0050] S140: If the sensor fault type is a single point fault, generate a lateral takeover intention vector according to the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data.

[0051] The lateral takeover intention vector can be understood as a vector composed of parameters representing the lateral takeover intention, and can be used to represent the magnitude of the lateral takeover intention.

[0052] Specifically, when the sensor fault type is a single-point fault, the mask tensor corresponding to the lateral sensing residual vector is determined according to the single-point fault type, and the lateral takeover intention vector is generated according to the lateral sensor and the corresponding mask tensor.

[0053] S150 , identifying whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and at least one corresponding set threshold.

[0054] It should be noted that a lateral takeover intention vector may include two or more intent vector elements, each of which may correspond to one, two, or more preset thresholds. Different preset thresholds can determine whether a lateral takeover intention is valid from different perspectives. Each preset threshold can be set based on actual needs and is not limited in this embodiment of the present invention.

[0055] Specifically, each element in the lateral takeover intention vector is compared with a corresponding set threshold to determine whether the takeover action corresponding to the lateral sensor data represented by the intention vector element is a valid lateral takeover action. If a valid lateral takeover action is determined, the automatic immediate mode can be terminated and the driver takeover mode can be switched.

[0056] The technical solution of the embodiment of the present invention receives lateral sensing data from multiple lateral control sensors through a communication link, generates a lateral sensing residual vector based on each lateral sensing data; determines the sensor fault type based on the lateral sensing residual vector; if the sensor fault type is a single point fault, generates a lateral takeover intention vector based on the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data; and identifies whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and a set threshold. In the event of a single point failure in the sensor, effective identification of the lateral takeover intention and takeover action can still be achieved, solving the problem of the prior art lacking a fault diagnosis and processing mechanism, and the potential hazards of unexpected exit or inability to exit the autonomous driving mode, avoiding unacceptable hazards to the system caused by a single sensor failure, and improving the safety level of the autonomous driving system.

[0057] Example 2

[0058] Figure 2This is a flowchart of a lateral takeover action recognition method provided by Example 2 of the present invention. This embodiment further optimizes the lateral sensing data from multiple lateral control sensors received through the communication link in the above embodiment as follows: the lateral sensing data includes: steering wheel angle sensing data collected by the steering wheel angle sensor, steering gear angle sensing data collected by the steering gear angle sensor, steering wheel torque sensing data collected by the steering wheel torque sensor, and steering wheel hand-off sensing data collected by the steering wheel hand-off sensor.

[0059] Correspondingly, generating a lateral sensing residual vector based on each of the lateral sensing data includes: estimating steering wheel torque estimation data based on the steering wheel angle sensing data and the steering gear angle sensing data; estimating steering wheel hands-off state estimation data based on the steering wheel torque sensing data and the steering wheel angle sensing data; generating a first residual value based on the steering wheel torque sensing data and the steering wheel torque estimation data; generating a second residual value based on the steering wheel hands-off sensing data and the steering wheel hands-off state estimation data; generating a third residual value based on the steering wheel angle sensing data and the steering gear angle sensing data; and generating a lateral sensing residual vector based on the first residual value, the second residual value and the third residual value.

[0060] like Figure 2 As shown, the method includes:

[0061] S201. Receive lateral sensing data from multiple lateral control sensors through a communication link, where the lateral sensing data includes: steering wheel angle sensing data collected by a steering wheel angle sensor, steering gear angle sensing data collected by a steering gear angle sensor, steering wheel torque sensing data collected by a steering wheel torque sensor, and steering wheel hand-off sensing data collected by a steering wheel hand-off sensor.

[0062] Specifically, the steering wheel torque sensor data collected by the steering wheel torque sensor can be obtained by: the steering wheel torque sensor S tq Steering wheel torque signal T q Measure and obtain steering wheel torque sensing data T′ q And upload the steering wheel angle sensor data to the communication link, so that the steering wheel torque sensor data T' can be obtained through the communication link q .

[0063] The method of obtaining the steering wheel angle sensor data collected by the steering wheel angle sensor and the steering gear angle sensor data collected by the steering gear angle sensor can be: q Input the steering column process model G1 to obtain the steering wheel angle signal δ sw and steering gear angle signal δ sc ;Through the steering wheel angle sensor Ssw Measuring the steering wheel angle signal δ sw Get the steering wheel angle sensor data δ′ sw And upload to the communication link, through the steering angle sensor S sc Measure the steering gear angle signal to obtain the steering gear angle sensor data δ′ sc And upload it to the communication link. Thus, the steering wheel angle sensor data δ′ can be obtained through the communication link sw and steering angle sensor data δ′ sc .

[0064] The method of obtaining the steering wheel hand-off sensor data collected by the steering wheel hand-off sensor can be: HOD The steering wheel hand-off signal HOD is measured to obtain steering wheel hand-off sensor data HOD′, and the steering wheel hand-off sensor data HOD′ is uploaded to the communication link, so that the steering wheel hand-off sensor data HOD′ can be obtained through the communication link.

[0065] S202: Estimate steering wheel torque estimation data based on the steering wheel angle sensor data and the steering gear angle sensor data.

[0066] The steering wheel angle sensor data may be understood as steering wheel angle data collected by the steering wheel angle sensor. The steering gear angle sensor data may be understood as steering gear angle data collected by the steering gear sensor. The steering wheel torque estimation data may be understood as estimated steering wheel torque data.

[0067] Specifically, the method of estimating the steering wheel torque estimation data based on the steering wheel angle sensor data and the steering gear angle sensor data may be: transforming the steering wheel angle sensor data δ′ into sw and steering angle sensor data δ′ sc Input steering wheel torque estimation model , obtain steering wheel torque estimation data T″ q The steering wheel torque estimation data can be specifically T″ q =k[δ′ sw -δ′ sc ], k is the steering wheel torque estimation model The correlation coefficient.

[0068] S203 : Estimate steering wheel hands-off state estimation data based on the steering wheel torque sensor data and the steering wheel angle sensor data.

[0069] The steering wheel torque sensing data may be understood as the steering wheel torque data collected by the steering torque sensor, and the steering wheel hands-off state estimation data may be understood as the estimated steering wheel hands-off state data.

[0070] Specifically, the method of estimating the steering wheel hands-off state estimation data based on the steering wheel torque sensor data and the steering wheel angle sensor data may be: q and steering wheel angle sensor data δ′ sw Input steering wheel hands-off state estimation model , obtain the steering wheel hand-off state estimation data HOD″. Among them, the steering wheel hand-off state estimation model The principle is: Among them, J sw is the steering wheel moment of inertia, is the first-order differential of the steering wheel angle sensor data, is the second-order differential of the steering wheel angle sensor data, C sw is the steering column rotation damping coefficient, T thre is the threshold value related to the maximum rotation rate of the steering gear, the maximum rotation acceleration, and the quality parameters of the steering column, T thre It can be obtained through test calibration.

[0071] S204 : Generate a first residual value according to the steering wheel torque sensing data and the steering wheel torque estimation data.

[0072] The first residual value may be understood as a residual value between the sensing data (ie, the measurement data) and the estimated data of the steering wheel torque.

[0073] Specifically, calculate the steering wheel torque sensor data T′ q and steering wheel torque estimation data T″ q The difference between them is taken as the first residual value r1, that is, r1 = T′ q -T″ q .

[0074] For example, Figure 3 This is a schematic diagram of the principle of generating the first residual value. Figure 3 As shown, the steering wheel torque sensor S tq Steering wheel torque signal T q Measure and obtain steering wheel torque sensing data T′ q ; The steering wheel torque signal T q Input the steering column process model G1 to obtain the steering wheel angle signal δ sw and steering gear angle signal δ sc ;Through the steering wheel angle sensor S sw Measuring the steering wheel angle signal δ sw Get the steering wheel angle sensor data δ′ sw ;Through the steering angle sensor S sc Measure the steering gear angle signal to obtain the steering gear angle sensor data δ′sc ; The steering wheel angle sensor data δ′ sw and steering angle sensor data δ′ sc Input steering wheel torque estimation model , obtain steering wheel torque estimation data T″ q , calculate the steering wheel torque sensor data T′ q and steering wheel torque estimation data T″ q The difference between them is taken as the first residual value r1.

[0075] S205 : Generate a second residual value according to the hands-off steering wheel sensing data and the hands-off steering wheel state estimation data.

[0076] The second residual value may be understood as the residual value between the sensing data (ie, the measurement data) and the estimated data of the hands-off state.

[0077] Specifically, the difference between the hands-off steering wheel sensing data HOD′ and the hands-off steering wheel state estimation data HOD″ is calculated as the second residual value r2, that is, r2=HOD′-HOD″.

[0078] For example, Figure 4 This is a schematic diagram of the principle of generating the second residual value. Figure 4 As shown, the steering wheel hand-off sensor S HOD The steering wheel hand-off signal HOD is measured to obtain the steering wheel hand-off sensor data HOD′, and the steering wheel torque sensor data T′ is converted to q and steering wheel angle sensor data δ′ sw Input steering wheel hands-off state estimation model , obtain the steering wheel hand-off state estimation data HOD″, and calculate the difference between the steering wheel hand-off sensor data HOD′ and the steering wheel hand-off state estimation data HOD″ as the second residual value r2.

[0079] S206 : Generate a third residual value according to the steering wheel angle sensor data and the steering gear angle sensor data.

[0080] The second residual value may be understood as the difference between the steering wheel angle and the steering gear angle.

[0081] Specifically, calculate the steering wheel angle sensor data δ′ sw and steering angle sensor data δ′ sc The difference between them is taken as the third residual value r3, that is, r3 = δ′ sw -δ′ sc .

[0082] For example, Figure 5 This is a schematic diagram of the principle of generating the third residual value. Figure 5As shown, through the steering wheel angle sensor S sw Measuring the steering wheel angle signal δ sw Get the steering wheel angle sensor data δ′ sw ;Through the steering angle sensor S sc Measure the steering gear angle signal to obtain the steering gear angle sensor data δ′ sc ; Steering wheel angle sensor data δ′ sw and steering angle sensor data δ′ sc The difference between them is taken as the third residual value r3.

[0083] S207 . Generate a lateral sensing residual vector according to the first residual value, the second residual value, and the third residual value.

[0084] Specifically, the lateral sensing residual vector r=[r1 r2 r3] is generated by sequentially arranging the first residual value r1, the second residual value r2, and the third residual value r3.

[0085] In an optional embodiment, S207, generating a lateral sensing residual vector according to the first residual value, the second residual value, and the third residual value, includes:

[0086] If the residual value is greater than a preset residual threshold, the residual value is set to 1; wherein the residual value includes: a first residual value, a second residual value, or a third residual value;

[0087] If the residual value is less than or equal to the preset residual threshold, the residual value is set to 0;

[0088] Determine a lateral sensing residual vector formed by arranging the residual values ​​in sequence.

[0089] Specifically, the calculated first residual value, second residual value, and third residual value are normalized respectively. The normalization method can be to determine whether the residual value is greater than a preset residual threshold. If so, the residual value is set to 1, indicating that the sensor associated with the residual value is faulty. If not, the residual value is set to 0, indicating that the sensor associated with the residual value is not faulty. Thus, the lateral sensing residual vector composed of the residual values ​​arranged in sequence is obtained. in, i=1,2,3,r i_thr is the preset residual threshold.

[0090] S208 : Determine the sensor fault type according to the lateral sensing residual vector.

[0091] S209: If the sensor fault type is a single point fault, generate a lateral takeover intention vector according to the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data.

[0092] S210. Identify whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and at least one corresponding set threshold.

[0093] The technical solution of the embodiment of the present invention receives steering wheel angle sensor data, steering gear angle sensor data, steering wheel torque sensor data and steering wheel hands-off sensor data through a communication link; and calculates a lateral sensor residual vector composed of a first residual value, a second residual value and a third residual value based on the above sensor data; determines the sensor fault type based on the lateral sensor residual vector; if the sensor fault type is a single point fault, generates a lateral takeover intention vector based on the mask tensor and lateral sensor data corresponding to the lateral sensor residual vector; and identifies whether the takeover action corresponding to the lateral sensor data is a valid lateral takeover action based on the lateral takeover intention vector and the corresponding at least one set threshold. In the event of a single sensor failure, the lateral takeover intention and takeover action can still be effectively identified, which solves the problem that the prior art lacks a fault diagnosis and processing mechanism and has the potential danger of unexpected exit or inability to exit the autonomous driving mode, avoids unacceptable harm to the system caused by a single sensor failure, and improves the safety level of the autonomous driving system.

[0094] Based on the above embodiment, S208, identifying the sensor fault type according to the lateral sensing residual vector includes:

[0095] S2081. If the lateral sensor residual vector is a non-zero vector, determine whether the lateral sensor residual vector is the same as a preset single-point fault residual vector.

[0096] Specifically, if the lateral sensing residual vector is a zero vector, that is, the residual values ​​in the lateral sensing residual vector are all zero, it means that there is no sensor fault in the autonomous driving system, and the takeover intention and whether the takeover action is effective can be directly identified based on the lateral sensing data. If the lateral sensing residual vector is a non-zero vector, that is, the residual values ​​in the lateral sensing residual vector are not all zero, it means that there is at least one sensor fault in the autonomous driving system. Therefore, fault identification is required to determine the type of sensor fault. The fault identification method can be to determine whether the sensor fault type is a single point fault by judging whether the lateral sensor residual vector is the same as the preset vector. The preset vector is the lateral sensing residual vector corresponding to the sensor under different fault conditions.

[0097] S2082: If the lateral sensor residual vector is the same as any preset single-point fault residual vector, determine that the sensor fault type is a single-point fault.

[0098] Specifically, as shown in Table 1, if the lateral sensor residual vector is the same as any preset single-point fault residual vector, that is, the lateral sensor residual vector The sensor fault type is determined to be a single-point fault. Single-point faults can be further categorized as steering wheel torque sensor single-point fault, steering wheel hands-off sensor single-point fault, steering wheel angle sensor single-point fault, and steering gear angle sensor single-point fault. If the lateral sensor residual vector differs from any of the preset single-point fault residual vectors, the sensor fault type is determined to be a multi-point fault. This embodiment of the present invention does not limit the method for identifying takeover actions under multi-point faults.

[0099] Based on the above optional embodiment, S2082, if the lateral sensor residual vector is the same as any preset single-point fault residual vector, determining that the sensor fault type is a single-point fault includes:

[0100] If the lateral sensor residual vector is the same as a first preset single-point fault residual vector, it is determined that the sensor fault type is a steering wheel torque sensor single-point fault; the first preset single-point fault residual vector is [1 0 0];

[0101] If the lateral sensor residual vector is the same as the second single-point fault residual preset vector, it is determined that the sensor fault type is a hands-off steering sensor single-point fault; the second preset single-point fault residual vector is [0 1 0];

[0102] If the lateral sensor residual vector is the same as the third single-point fault residual preset vector, it is determined that the sensor fault type is a steering wheel angle sensor single-point fault; the third preset single-point fault residual vector is [1 1 1];

[0103] If the lateral sensor residual vector is the same as the fourth single-point fault residual preset vector, it is determined that the sensor fault type is a steering gear angle sensor single-point fault; the fourth preset single-point fault residual vector is [1 0 1].

[0104] For example, Table 1 is a table of preset single point fault residual vectors corresponding to single point fault types. j=1,2,3,4.

[0105] Table 1

[0106]

[0107]

[0108] As shown in Table 1, if Indicates that the sensor fault type is a single-point fault of the steering wheel torque sensor; if Indicates that the sensor fault type is a single-point fault of the steering wheel hands-off sensor; if Indicates that the sensor fault type is a single-point fault of the steering wheel angle sensor; if Indicates a single-point failure in the steering angle sensor.

[0109] In an optional embodiment, S209, generating a lateral takeover intention vector according to the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data, includes:

[0110] S2091. Obtain a mask tensor corresponding to the lateral sensing residual vector;

[0111] S2092: Determine a transverse data vector to be processed, which is formed by sequentially arranging the steering wheel torque sensor data, the steering wheel torque estimation data, the steering wheel hands-off sensor data, and the steering wheel hands-off state estimation data;

[0112] S2093. Determine the vector product of the horizontal data vector to be processed and the mask tensor as the horizontal takeover intention vector.

[0113] Specifically, the steering wheel torque sensor data T′ q , steering wheel torque estimation data T″, steering wheel hand-off sensor data HOD′ and steering wheel hand-off state estimation data HOD″ are arranged in sequence to form the lateral data to be processed, D i =[T′ q T″ q HOD′HOD″] T ; Obtain the corresponding mask tensor M according to the fault type represented by the lateral sensing residual vector i , according to D o =M·D i , calculate the lateral takeover intention vector D o .

[0114] Table 2

[0115]

[0116]

[0117] In an optional embodiment, S210, identifying whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and at least one corresponding set threshold, includes:

[0118] S2101: Compare each intention vector element in the lateral takeover intention vector with a first set threshold and a second set threshold respectively.

[0119] Among them, the first set threshold can be a threshold related to the intensity of the takeover action, and the second set threshold can be a threshold related to the duration of the takeover action; the first set threshold and the second set threshold can be set according to actual needs, and the embodiment of the present invention does not limit this.

[0120] For example, the lateral takeover intention vector is D o =[T″ q HOD′] T , the steering wheel torque estimation data T″ is compared with the first set threshold value T1 and the second set threshold value T2 respectively; and the steering wheel hands-off state estimation data HOD″ is compared with the first set threshold value T1 and the second set threshold value T2 respectively.

[0121] S2102: If each of the intention vector elements is greater than the first set threshold and the second set threshold, determine that the takeover action corresponding to the lateral sensing data is a valid lateral takeover action.

[0122] For example, if T″>T1, T″>T2, and HOD″>T1, HOD″>T2; then the takeover action corresponding to the steering wheel torque estimation data T″ and the steering wheel hands-off state estimation data HOD″ is a valid lateral takeover action, that is, the lateral takeover action corresponding to the takeover intention recognized by the automatic driving system is a valid lateral takeover action.

[0123] S2103: If there is at least one intention vector element that is less than or equal to the first set threshold or the second set threshold, determine that the takeover action corresponding to the lateral sensing data is an invalid lateral takeover action.

[0124] For example, if T″≤T1, T″≤T2, HOD″≤T1 or HOD″>T2, the takeover action corresponding to the steering wheel torque estimation data T″ and the steering wheel hands-off state estimation data HOD″ is an invalid lateral takeover action, that is, the lateral takeover action corresponding to the takeover intention recognized by the automatic driving system is an invalid lateral takeover action.

[0125] Example 3

[0126] Figure 6 This is a flowchart of a method for identifying a lateral takeover action provided in Example 3 of the present invention. This embodiment further optimizes the lateral sensing data from multiple lateral control sensors received through a communication link in the above embodiment to: receiving lateral sensing data from multiple lateral control sensors through a two-way communication link.

[0127] At the same time, before generating a lateral sensing residual vector based on each of the lateral sensing data, it also includes: parsing and verifying the lateral sensing data from multiple lateral control sensors received through the dual-path communication link; if the lateral sensing data on one of the communication links fails to be verified, the communication link corresponding to the lateral sensing data is confirmed as a faulty link, and the lateral sensing data is discarded.

[0128] like Figure 6 As shown, the method includes:

[0129] S310 , receiving lateral sensing data from a plurality of lateral control sensors via a two-way communication link.

[0130] Among them, the lateral takeover action recognition used in the embodiment of the present invention adopts a communication architecture of a dual-channel communication link, which has two communication links. Each communication link can be connected to multiple lateral control sensors to obtain lateral sensing data collected by each lateral control sensor.

[0131] For example, Figure 7 This is a diagram of the communication architecture for horizontal takeover action recognition. Figure 7 As shown, the controller establishes communication with communication link 1 and communication link 2, respectively. Communication link 1 and communication link 2 each establish communication with multiple lateral control sensors. Lateral sensing data from the multiple lateral control sensors can be acquired via communication link 1 and communication link 2. The lateral sensing data may include at least steering wheel angle sensing data, steering gear angle sensing data, steering wheel torque sensing data, and hands-off steering wheel sensing data.

[0132] S320: Analyze and verify the lateral sensing data received through the two-way communication link.

[0133] Specifically, lateral sensing data from multiple lateral control sensors acquired via communication link 1 and communication link 2 are parsed and verified to determine whether a link failure has occurred. Parsing methods may include decryption and data message parsing. Verification methods may include performing message checksum verification, numerical analysis verification, etc. on the lateral sensing data acquired and parsed via communication link 1 and communication link 2.

[0134] S330: If the lateral sensing data on one communication link fails to be verified, the lateral sensing data that fails to be verified is discarded.

[0135] Specifically, for lateral sensing data that fails verification, it can be considered that the communication link transmitting the lateral sensing data fails. Therefore, the lateral sensing data that fails verification is discarded, and the data that passes verification is regarded as valid lateral sensing data.

[0136] It should be noted that lateral sensing data that fails to be parsed can also be considered as a case of verification failure.

[0137] S340, generating a lateral sensing residual vector according to the verified lateral sensing data;

[0138] Specifically, the verified lateral sensing data is used as valid lateral sensing data to generate a lateral sensing residual vector. This allows the system to receive lateral sensing data via another communication link if a single communication link fails, thus avoiding the potential hazards of unexpectedly exiting or being unable to exit the autonomous driving mode due to a communication link failure.

[0139] S350: Determine the sensor fault type according to the lateral sensing residual vector.

[0140] S360: If the sensor fault type is a single point fault, generate a lateral takeover intention vector according to the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data.

[0141] S370. Identify whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and at least one corresponding set threshold.

[0142] The technical solution of the embodiment of the present invention is to parse and verify the lateral sensing data from multiple lateral control sensors received through a dual-path communication link; if the lateral sensing data on one communication link fails the verification, the communication link corresponding to the lateral sensing data is confirmed as a faulty link, and the lateral sensing data is discarded; if the sensor failure type is a single point failure, a lateral takeover intention vector is generated based on the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data; based on the lateral takeover intention vector and the corresponding at least one set threshold, it is identified whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action. In the event of a single point failure in the communication link or sensor, the lateral takeover intention and takeover action can still be effectively identified, thereby avoiding the potential harm of a single sensor failure to the system causing unexpected exit or inability to exit the automatic driving mode, and improving the safety level of the automatic driving system.

[0143] Figure 8 This is a fault tree analysis diagram of a method for identifying a lateral takeover action provided by the second embodiment of the present invention. Figure 8As shown, the autonomous driving system's failure to recognize a lateral takeover maneuver or its inability to recognize a takeover maneuver may be due to unreliable hands-off sensor data, unreliable steering torque sensor data, or an unreliable communication link. Unreliable hands-off sensor data further stems from a faulty steering wheel sensor and unreliable hands-off estimation data. Unreliable hands-off estimation data stems from a faulty steering wheel angle sensor or steering wheel torque sensor. Unreliable steering torque sensor data further stems from a faulty steering wheel torque sensor and unreliable steering wheel torque estimation data. Unreliable steering wheel torque estimation data stems from a faulty steering wheel angle sensor or steering gear angle sensor. Unreliable communication link data stems from a faulty communication link 1 or communication link 2.

[0144] Example 4

[0145] Figure 9 This is a schematic diagram of the structure of a device for identifying a lateral takeover action provided by the fourth embodiment of the present invention. Figure 9 As shown, the device includes:

[0146] a data receiving module 410 for receiving lateral sensing data from a plurality of lateral control sensors via a two-way communication link;

[0147] a residual vector determining module 420, configured to generate a lateral sensing residual vector according to each of the lateral sensing data;

[0148] a fault type determination module 430, configured to determine a sensor fault type based on the lateral sensing residual vector;

[0149] an intention vector determination module 440 for generating a lateral takeover intention vector based on the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data if the sensor fault type is a single point fault;

[0150] The takeover action recognition module 450 is used to identify whether the takeover action corresponding to the lateral sensor data is a valid lateral takeover action based on the lateral takeover intention vector and at least one corresponding set threshold.

[0151] Optionally, the lateral sensing data includes:

[0152] Steering wheel angle sensing data collected by the steering wheel angle sensor, steering gear angle sensing data collected by the steering gear angle sensor, steering wheel torque sensing data collected by the steering wheel torque sensor, and steering wheel hand-off sensing data collected by the steering wheel hand-off sensor.

[0153] Optionally, the residual vector determination module 420 includes:

[0154] a first estimating unit, configured to estimate steering wheel torque estimation data based on the steering wheel angle sensing data and the steering gear angle sensing data;

[0155] a second estimating unit, configured to estimate steering wheel hands-off state estimation data based on the steering wheel torque sensing data and the steering wheel angle sensing data;

[0156] a first residual calculation unit, configured to generate a first residual value according to the steering wheel torque sensing data and the steering wheel torque estimation data;

[0157] a second residual calculation unit, configured to generate a second residual value according to the hands-off steering wheel sensing data and the hands-off steering wheel state estimation data;

[0158] a third residual calculation unit, configured to generate a third residual value according to the steering wheel angle sensor data and the steering gear angle sensor data;

[0159] A residual vector generating unit is used to generate a lateral sensing residual vector according to the first residual value, the second residual value and the third residual value.

[0160] Optionally, the residual vector generating unit is specifically configured to:

[0161] If the residual value is greater than a preset residual threshold, the residual value is set to 1; wherein the residual value includes: a first residual value, a second residual value, or a third residual value;

[0162] If the residual value is less than or equal to the preset residual threshold, the residual value is set to 0;

[0163] Determine a lateral sensing residual vector formed by arranging the residual values ​​in sequence.

[0164] Optionally, the fault type determination module 430 includes:

[0165] a judgment unit, configured to judge whether the lateral sensor residual vector is the same as a preset single-point fault residual vector if the lateral sensor residual vector is a non-zero vector;

[0166] A single point fault determination unit is configured to determine that the sensor fault type is a single point fault if the lateral sensor residual vector is the same as any preset single point fault residual vector.

[0167] Optionally, the single point failure determination unit is specifically configured to:

[0168] If the lateral sensor residual vector is the same as a first preset vector, it is determined that the sensor fault type is a steering wheel torque sensor single-point fault; the first preset vector is [1 0 0];

[0169] If the lateral sensor residual vector is the same as a second preset vector, it is determined that the sensor fault type is a single-point fault of the hands-off steering wheel sensor; the second preset vector is [0 1 0];

[0170] If the lateral sensor residual vector is the same as a third preset vector, it is determined that the sensor fault type is a single-point fault of the steering wheel angle sensor; the third preset vector is [1 1 1];

[0171] If the lateral sensor residual vector is the same as a fourth preset vector, it is determined that the sensor fault type is a steering gear angle sensor single-point fault; the fourth preset vector is [1 0 1].

[0172] Optionally, the intention vector determination module 440 is specifically configured to:

[0173] Obtaining a mask tensor corresponding to the lateral sensing residual vector;

[0174] determining a lateral data vector to be processed, which is formed by sequentially arranging the steering wheel torque sensor data, the steering wheel torque estimation data, the steering wheel hands-off sensor data, and the steering wheel hands-off state estimation data;

[0175] The vector product of the horizontal data vector to be processed and the mask tensor is determined as the horizontal takeover intention vector.

[0176] Optionally, the taken-over action recognition module 450 is specifically configured to:

[0177] Comparing each intention vector element in the lateral takeover intention vector with a first set threshold and a second set threshold respectively;

[0178] If each of the intention vector elements is greater than the first set threshold and the second set threshold, determining that the takeover action corresponding to the lateral sensing data is a valid lateral takeover action;

[0179] If there is at least one intention vector element that is less than or equal to the first set threshold or the second set threshold, it is determined that the takeover action corresponding to the lateral sensing data is an invalid lateral takeover action.

[0180] Optionally, the data receiving module 410 is specifically configured to:

[0181] receiving lateral sensing data from a plurality of lateral control sensors via a two-way communication link;

[0182] Optionally, also include:

[0183] an analysis module for analyzing and verifying the lateral sensing data received via the two-way communication link before generating a lateral sensing residual vector based on each of the lateral sensing data;

[0184] The data discarding module is used to discard the lateral sensing data that fails the verification if there is lateral sensing data on one communication link that fails the verification.

[0185] The lateral takeover action recognition device provided in the embodiment of the present invention can execute the lateral takeover action recognition method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0186] Example 5

[0187] Figure 10 A schematic diagram of a vehicle 10 is shown that may be used to implement an embodiment of the present invention. The components shown herein, their connections and relationships, and their functions are provided as examples only and are not intended to limit implementations of the inventions described and / or claimed herein.

[0188] like Figure 10 As shown, the vehicle 10 includes at least one controller 11 and a memory, such as a read-only memory (ROM) 12 and a random access memory (RAM) 13, communicatively connected to the at least one controller 11. The memory stores a computer program executable by the at least one controller, and the controller 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the vehicle 10 can also be stored in the RAM 13. The controller 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0189] Various components in the vehicle 10 are connected to the I / O interface 15 , including an input unit 16 , such as a touchscreen display and buttons; an output unit 17 , such as various types of displays and speakers; a storage unit 18 , such as a magnetic disk or optical disk; and a communication unit 19 , such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the vehicle 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0190] The controller 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the controller 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various controllers running machine learning model algorithms, digital signal controllers (DSPs), and any other suitable controllers, controllers, microcontrollers, etc. The controller 11 executes the various methods and processes described above, such as the lateral takeover action recognition method.

[0191] In some embodiments, the lateral takeover maneuver recognition method can be implemented as a computer program that is tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on vehicle 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by controller 11, one or more steps of the lateral takeover maneuver recognition method described above can be performed. Alternatively, in other embodiments, controller 11 can be configured to perform the lateral takeover maneuver recognition method in any other appropriate manner (e.g., by means of firmware).

[0192] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable controller, which can be a special purpose or general purpose programmable controller that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0193] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0194] In the context of the present invention, computer-readable storage media can be tangible media that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage media can be machine-readable signal media. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0195] To provide interaction with a user, the systems and techniques described herein can be implemented in a vehicle having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the vehicle. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0196] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0197] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0198] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0199] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A method for identifying a lateral takeover action, characterized in that: Applied to a vehicle, the method comprises: receiving lateral sensing data from a plurality of lateral control sensors via a communication link; generating a lateral sensing residual vector according to each of the lateral sensing data; determining a sensor fault type according to the lateral sensing residual vector; If the sensor fault type is a single point fault, generating a lateral takeover intention vector according to the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data; Identifying whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action according to the lateral takeover intention vector and at least one corresponding set threshold; The lateral sensing data includes: Steering wheel angle sensor data collected by a steering wheel angle sensor, steering gear angle sensor data collected by a steering gear angle sensor, steering wheel torque sensor data collected by a steering wheel torque sensor, and steering wheel hands-off sensor data collected by a steering wheel hands-off sensor; Generating a lateral sensing residual vector according to each of the lateral sensing data includes: estimating steering wheel torque estimation data based on the steering wheel angle sensor data and the steering gear angle sensor data; estimating steering wheel hands-off state estimation data based on the steering wheel torque sensor data and the steering wheel angle sensor data; generating a first residual value based on the steering wheel torque sensing data and the steering wheel torque estimation data; generating a second residual value according to the hands-off steering wheel sensing data and the hands-off steering wheel state estimation data; generating a third residual value according to the steering wheel angle sensor data and the steering gear angle sensor data; generating a lateral sensing residual vector according to the first residual value, the second residual value, and the third residual value; Generating a lateral sensing residual vector according to the first residual value, the second residual value, and the third residual value includes: If the residual value is greater than a preset residual threshold, the residual value is set to 1; wherein the residual value includes: a first residual value, a second residual value, or a third residual value; If the residual value is less than or equal to the preset residual threshold, the residual value is set to 0; Determine a lateral sensing residual vector formed by arranging the residual values ​​in sequence.

2. The method according to claim 1, characterized in that Identifying the sensor fault type according to the lateral sensing residual vector includes: If the lateral sensing residual vector is a non-zero vector, comparing the lateral sensing residual vector with a preset single-point fault residual vector to determine whether they are the same; If the lateral sensing residual vector is the same as any preset single-point fault residual vector, it is determined that the sensor fault type is a single-point fault.

3. The method according to claim 2, characterized in that If the lateral sensing residual vector is the same as any preset single-point fault residual vector, determining that the sensor fault type is a single-point fault includes: If the lateral sensing residual vector is the same as the first preset vector, it is determined that the sensor fault type is a single-point fault of the steering wheel torque sensor; the first preset vector is [1, 0, 0]; If the lateral sensor residual vector is the same as a second preset vector, the sensor fault type is determined to be a single-point fault of the hands-off steering wheel sensor; the second preset vector is [0, 1, 0]; If the lateral sensor residual vector is the same as a third preset vector, it is determined that the sensor fault type is a single-point fault of the steering wheel angle sensor; the third preset vector is [1, 1, 1]; If the lateral sensing residual vector is the same as a fourth preset vector, it is determined that the sensor fault type is a steering gear angle sensor single-point fault; the fourth preset vector is [1, 0, 1].

4. The method according to claim 3, characterized in that The generating of the lateral takeover intention vector according to the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data includes: Obtaining a mask tensor corresponding to the lateral sensing residual vector; determining a lateral data vector to be processed, which is formed by sequentially arranging the steering wheel torque sensor data, the steering wheel torque estimation data, the steering wheel hands-off sensor data, and the steering wheel hands-off state estimation data; The vector product of the horizontal data vector to be processed and the mask tensor is determined as the horizontal takeover intention vector.

5. The method according to claim 1, wherein The identifying, based on the lateral takeover intention vector and at least one corresponding set threshold, whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action includes: Comparing each intention vector element in the lateral takeover intention vector with a first set threshold and a second set threshold respectively; If each of the intention vector elements is greater than the first set threshold and the second set threshold, determining that the takeover action corresponding to the lateral sensing data is a valid lateral takeover action; If there is at least one intention vector element that is less than or equal to the first set threshold or the second set threshold, it is determined that the takeover action corresponding to the lateral sensing data is an invalid lateral takeover action.

6. The method according to claim 1, wherein Receives lateral sensing data from multiple lateral control sensors via a communication link, including: receiving lateral sensing data from a plurality of lateral control sensors via a two-way communication link; Accordingly, before generating the lateral sensing residual vector according to each of the lateral sensing data, the method further includes: Parsing and validating lateral sensing data received via the two-way communication link; If the lateral sensing data on one communication link fails to be verified, the lateral sensing data that fails to be verified will be discarded.

7. A lateral takeover action recognition device implementing the lateral takeover action recognition method according to any one of claims 1 to 6, characterized in that: include: a data receiving module, configured to receive lateral sensing data from a plurality of lateral control sensors via a two-way communication link; a residual vector determination module, configured to generate a lateral sensing residual vector according to each of the lateral sensing data; a fault type determination module, configured to determine a sensor fault type based on the lateral sensing residual vector; an intention vector determination module, configured to generate a lateral takeover intention vector based on the mask tensor corresponding to the lateral sensing residual vector and the lateral sensing data if the sensor fault type is a single point fault; The takeover action recognition module is used to identify whether the takeover action corresponding to the lateral sensing data is a valid lateral takeover action based on the lateral takeover intention vector and at least one corresponding set threshold.

8. A vehicle, characterized in that: The vehicle comprises: at least one controller; and a memory in communication with the at least one controller; wherein, The memory stores a computer program that can be executed by the at least one controller. The computer program is executed by the at least one controller to enable the at least one controller to perform the lateral takeover action identification method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a controller to implement the lateral takeover action recognition method according to any one of claims 1 to 6 when executed.

Images