Abnormality detection method and device and unmanned vehicle

Abstract

The embodiment of the invention provides an anomaly detection method and device and an unmanned vehicle. The method comprises the steps of receiving sensing data collected by sensing equipment; for each detection period, performing anomaly detection on the sensing data received in the detection period; under the condition that the detection result of the detection period represents that abnormity exists, whether the number of continuously detected abnormity times reaches a preset value or not is judged according to a historical detection result, and if the number of continuously detected abnormity times reaches the preset value, it is determined that the sensing equipment is abnormal; and when it is determined that the sensing device is abnormal, the abnormal condition is reported. The probability of false detection is reduced, and the accuracy of abnormal detection of the sensing equipment is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of autonomous driving technology, and in particular to an abnormality detection method, device and unmanned vehicle. Background Art

[0002] With the rapid development of artificial intelligence and Internet of Things technologies, autonomous vehicles have become a major trend in the field of transportation. However, for cars to drive autonomously, it is crucial to perceive the surrounding environment. This requires relying on advanced environmental perception systems. The environmental perception system is the core of the autonomous driving system, which directly affects the safety and reliability of the system. At the same time, the application and popularization of ADAS (Advanced Driving Assistance System) and autonomous driving technology have further stimulated the unprecedented demand for perception devices. At present, cameras, millimeter-wave radars, ultrasonic radars, and lidars are all commonly used perception devices in the field of autonomous driving. The development and progress of these perception devices and environmental perception systems are crucial to improving the perception capabilities of vehicles in complex driving environments. However, when the perception devices collect and / or transmit perception data, not only will the perception devices themselves fail, but the perception data may also be affected by various factors, resulting in data anomalies. Existing anomaly detection technologies still have certain limitations when dealing with complex environments and dynamic changes. Summary of the invention

[0003] The embodiments of the present disclosure provide an anomaly detection method, device and unmanned vehicle to solve the related problems existing in the existing solutions.

[0004] Based on the above problems, in a first aspect, an anomaly detection method is provided, comprising:

[0005] Receiving perception data collected by the perception device;

[0006] For each detection cycle, anomaly detection is performed on the sensing data received in this detection cycle;

[0007] If the detection result of the current detection cycle indicates that an abnormality exists, it is determined whether the number of continuously detected abnormalities reaches a preset value based on the historical detection results. If the preset value is reached, it is determined that the sensing device has an abnormality.

[0008] When it is determined that the sensing device has an abnormality, the abnormality is reported.

[0009] In combination with the first aspect, in a possible implementation manner, the anomaly detection includes: data transmission anomaly detection;

[0010] For each detection cycle, anomaly detection is performed on the sensing data received in this detection cycle, including:

[0011] For each detection cycle, comparing a first data volume of the sensing data received in the current detection cycle with a preset data volume;

[0012] In the case where the detection result of the current detection cycle indicates that an abnormality exists, judging whether the number of continuously detected abnormalities reaches a preset value according to the historical detection results, and if the preset value is reached, determining that an abnormality occurs in the sensing device, includes:

[0013] When the ratio of the first data amount to the preset data amount satisfies a preset range, determining that a data transmission abnormality occurs;

[0014] Determine whether the number of consecutive data transmission anomalies represented by the historical detection results and the detection results of the current detection cycle since the last time the sensing device was determined to have an anomaly has reached a first preset value;

[0015] When the first preset value is reached, it is determined that an abnormality occurs in the sensing device.

[0016] In combination with the first aspect, in a possible implementation manner, when a ratio of the first data amount to the preset data amount satisfies a preset range, determining that a data transmission abnormality occurs includes:

[0017] When the ratio of the first data amount to the preset data amount satisfies a first preset range, determining that a small amount of packet loss occurs;

[0018] When the ratio of the first data amount to the preset data amount satisfies a second preset range, determining that a large amount of packet loss occurs;

[0019] When it is determined that the sensing device is abnormal, reporting the abnormal situation includes:

[0020] When it is determined that the sensing device has an abnormality, each abnormality of the continuous data transmission abnormalities is reported; the abnormality includes: a small amount of packet loss and / or a large amount of packet loss.

[0021] In combination with the first aspect, in a possible implementation, the anomaly detection includes: sensing distance anomaly detection;

[0022] For each detection cycle, anomaly detection is performed on the sensing data received in this detection cycle, including:

[0023] For each detection cycle, determining the perception distances corresponding to the perception data obtained in the detection cycle; comparing the determined perception distances with a preset perception distance range;

[0024] When the detection result of the current detection cycle indicates that an abnormality exists, judging whether the number of continuously detected abnormalities reaches a preset value according to the historical detection results, and if the preset value is reached, determining that an abnormality occurs in the sensing device, including:

[0025] When the amount of the sensing data whose sensing distance exceeds the preset sensing distance range accounts for a proportion of the amount of the sensing data acquired in the current detection period that reaches a preset proportion, it is determined that a sensing distance abnormality occurs;

[0026] Determine whether the number of consecutive occurrences of perception distance anomalies represented by the detection results of the historical detection cycle and the current detection cycle since the last time the perception device was determined to have an abnormality has reached a second preset value;

[0027] When the second preset value is reached, determining that an abnormality occurs in the sensing device;

[0028] When it is determined that the sensing device is abnormal, reporting the abnormal situation includes:

[0029] When it is determined that the sensing device is abnormal, the sensing distance abnormality is reported.

[0030] In combination with the first aspect, in a possible implementation manner, before receiving the perception data collected by the perception device, the method further includes:

[0031] Detecting whether the sensing device has sensing data output;

[0032] In the absence of sensor data output, determining that data transmission anomaly occurs;

[0033] Determining whether an abnormality occurs in the sensing device according to a historical abnormality detection result of the sensing device includes:

[0034] Determine whether the number of consecutive data transmission anomalies represented by the historical detection results and the current detection results since the last time the sensing device was determined to have an anomaly has reached a first preset value;

[0035] When the first preset value is reached, determining that an abnormality occurs in the sensing device;

[0036] When it is determined that the sensing device is abnormal, reporting the abnormal situation includes:

[0037] When it is determined that the sensing device has an abnormality, each abnormality of the consecutive data transmission abnormalities is reported; the abnormality includes: the sensing device has no sensing data output.

[0038] In combination with the first aspect, in a possible implementation manner, the method further includes:

[0039] In the case of sensory data output, it also includes:

[0040] Determine whether the time difference between the time of entering the previous detection cycle and the current time reaches a preset time difference;

[0041] When the preset time difference is reached, the next detection cycle is entered, and the time of entering the next detection cycle is recorded.

[0042] In combination with the first aspect, in a possible implementation manner, when it is determined that the sensing device is abnormal, a detection result of the sensing device is monitored;

[0043] In the case where it is monitored that the sensing device has returned to normal, determining whether the number of detection cycles in which the detection result is continuously normal since the sensing device has returned to normal has reached a third preset value;

[0044] When the third preset value is reached, determining that the sensing device has returned to normal and stopping monitoring;

[0045] When the third preset value is not reached, the detection result of the sensing device continues to be monitored.

[0046] In combination with the first aspect, in a possible implementation, when monitoring that the sensing device has returned to normal, determining whether the number of detection cycles in which the detection result is continuously normal since the sensing device has returned to normal has reached a third preset value; when the third preset value is reached, determining that the sensing device has returned to normal and stopping monitoring includes:

[0047] For each detection cycle monitored, if the detection result of the current detection cycle is normal, the counter is counted cumulatively; if the detection result of the current detection cycle is abnormal, the counter is cleared;

[0048] Determining the cumulative counting result of the counter;

[0049] When the accumulated count of the counter reaches a third preset value, and / or the accumulated count result of the counter indicates that the sensing device has been working normally for a preset time, it is determined that the sensing device has returned to normal and monitoring is stopped.

[0050] In a second aspect, an abnormality detection device is provided, comprising:

[0051] A data receiving module, used to receive the sensing data collected by the sensing device;

[0052] The data anomaly detection module is used to perform anomaly detection on the sensed data received in each detection cycle;

[0053] The device anomaly detection module is used to determine whether the number of consecutively detected anomalies reaches a preset value based on historical detection results when the detection results of the current detection cycle indicate that an anomaly exists, and if so, determine that an anomaly occurs in the sensing device;

[0054] The abnormality reporting module is used to report the abnormal situation when it is determined that the sensing device has an abnormality.

[0055] In a third aspect, an unmanned vehicle is provided, comprising the second aspect, or an abnormality detection device described in any possible implementation manner of the second aspect.

[0056] The beneficial effects of the embodiments of the present disclosure include:

[0057] The disclosed embodiment provides an abnormality detection method, device and unmanned vehicle, which are used to detect whether the perception data collected by the perception device installed on the autonomous driving unmanned vehicle is abnormal, and then determine whether the corresponding perception device has a fault and report the fault situation, so as to facilitate the subsequent processing of the fault. The abnormality detection method provided by the disclosed embodiment first receives the perception data collected by the perception device; then for each detection cycle, determines whether the perception data received in the cycle is abnormal; then, in the case of an abnormality in this cycle, combined with the previous historical detection results, determines whether the number of consecutive abnormalities reaches a preset value, and in the case of the number of consecutive abnormalities reaching the preset value, determines whether the perception device is abnormal, and determines whether the device is abnormal by detecting whether the number of consecutive abnormalities reaches the preset value, thereby avoiding the influence of occasional data abnormalities on device abnormality detection and improving the accuracy of detection; finally, in the case of determining that the perception device is abnormal, the specific abnormal situation is reported. The probability of false detection is reduced and the robustness and reliability of the detection algorithm are enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flowchart of an anomaly detection method provided by an embodiment of the present disclosure;

[0059] Figure 2 One of the flowcharts of data transmission anomaly detection provided by the embodiment of the present disclosure;

[0060] Figure 3 A flow chart of data distance anomaly detection provided by an embodiment of the present disclosure;

[0061] Figure 4 The second flowchart of data transmission anomaly detection provided by the embodiment of the present disclosure;

[0062] Figure 5 A flowchart of monitoring and determining that a sensing device has returned to normal according to an embodiment of the present disclosure;

[0063] Figure 6 One of the structural schematic diagrams of an abnormality detection device provided in an embodiment of the present disclosure;

[0064] Figure 7 The second structural schematic diagram of an abnormality detection device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0065] The embodiments of the present disclosure provide an abnormality detection method, device and unmanned vehicle. The preferred embodiments of the present disclosure are described below in conjunction with the drawings of the specification. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present disclosure, and are not used to limit the present disclosure. In addition, the embodiments in the present application and the features in the embodiments can be combined with each other if there is no conflict.

[0066] The present disclosure provides an abnormality detection method, such as Figure 1 As shown, including:

[0067] S101, receiving perception data collected by a perception device;

[0068] S102: for each detection period, perform anomaly detection on the sensing data received in the current detection period;

[0069] S103, when the detection result of the current detection cycle indicates that an abnormality exists, judging whether the number of consecutively detected abnormalities reaches a preset value according to the historical detection results, and if so, determining that an abnormality occurs in the sensing device;

[0070] S104: When it is determined that an abnormality occurs in the sensing device, the abnormality is reported.

[0071] In the disclosed embodiment, during the driving process of the autonomous driving vehicle, multiple sensing devices of different types (e.g., laser radar, millimeter wave radar, camera, etc.) installed on the vehicle will collect the surrounding environmental information to obtain sensing data, and determine whether the corresponding sensing device is abnormal by analyzing the sensing data, and report the abnormal situation after determining that the device is abnormal, so as to facilitate the subsequent corresponding processing of the abnormal situation of the sensing device. Since the traditional method only detects abnormalities in the data and uses the detection results as the basis for judging whether the device is abnormal, occasional data anomalies are very common in complex environments, which often leads to false detection of sensing devices. Therefore, the anomaly detection method provided by the present invention first obtains the perception data collected by the perception device; secondly, performs anomaly detection on the perception data obtained in each cycle according to the cycle to determine whether the perception data in the current cycle is abnormal; then, after determining that the anomaly detection result for the perception data in the current cycle is anomaly, it will not blindly report that the perception device has an abnormality, but will combine historical detection results to determine whether a preset value of consecutive detection results including the current cycle are all abnormal, that is, whether the number of consecutively detected abnormalities reaches a preset value. If at least one of the preset value of detection results is no abnormality, it is not determined that the corresponding perception device has an abnormality; if it reaches the preset value, it is determined that the corresponding perception device has an abnormality; by determining whether the perception device has an abnormality when the number of consecutive abnormalities reaches the preset value, it is avoided to detect the abnormality of the perception device due to occasional data anomalies, thereby improving the accuracy of anomaly detection; finally, when it is determined that the perception device has an abnormality, the specific abnormal situation is reported.

[0072] The execution subject of the abnormality detection method provided in the present disclosure can be a perception device fault detection system in the field of autonomous driving. In one possible implementation, the fault detection system is mainly composed of a data collection unit, a data processing unit, a data storage unit and a data communication unit, wherein the data collection unit is responsible for collecting data from multiple perception devices of the autonomous driving vehicle; the data processing unit includes a central processing unit (CPU) and a memory for executing data analysis and fault detection algorithms; the data storage unit is used to store historical data and real-time data, as well as fault detection models; the data communication unit is responsible for data exchange between the system and external devices (such as maintenance stations or other vehicles). In one possible implementation, the data processing unit. In one possible implementation, the perception device fault detection system can be a domain controller.

[0073] In yet another embodiment provided by the present disclosure, anomaly detection includes: data transmission anomaly detection;

[0074] In the above S102, for each detection period, anomaly detection is performed on the sensed data received in the current detection period, including:

[0075] Step 1: for each detection period, compare the first data volume of the sensed data received in the detection period with the preset data volume;

[0076] In the above S103, when the detection result of the current detection cycle indicates that an abnormality exists, judging whether the number of continuously detected abnormalities reaches a preset value according to the historical detection results, and if the preset value is reached, determining that the sensing device has an abnormality, includes (continuing from the above step 1):

[0077] Step 2: When the ratio of the first data volume to the preset data volume satisfies a preset range, determining that a data transmission abnormality occurs;

[0078] Step 3: determine whether the number of consecutive data transmission anomalies represented by the historical detection results and the detection results of the current detection cycle since the last time the sensing device was determined to have an abnormality has reached a first preset value;

[0079] Step 4: When the first preset value is reached, determine that an abnormality occurs in the sensing device.

[0080] In the embodiment of the present disclosure, taking the domain controller as an example, after the sensing device collects the sensing data, it needs to transmit the sensing data to the domain controller so that the data processing unit of the domain controller processes the sensing data through a preset algorithm to realize the corresponding function (for example, obstacle recognition, etc.). Therefore, it is necessary to ensure that the data transmission between the sensing device and the domain controller is normal, and it is necessary to perform data transmission anomaly detection on the sensing device.

[0081] In order to make the detection more reasonable and orderly, and to avoid false detection of abnormalities of the sensing device due to occasional data anomalies, data transmission anomaly detection can be performed on the sensing data received in each period according to the data collection time. The first data volume of the sensing data received in this detection period can be compared with the preset data volume; if the ratio of the first data volume to the preset data volume meets the preset range, it can be determined that a data transmission anomaly has occurred, and the abnormal detection result of the sensing data in this detection period is that an anomaly exists; then, based on step S103 in the above-mentioned anomaly detection method combined with the historical detection results, it is determined whether the number of consecutive anomalies reaches the first preset value, and then it is determined whether the corresponding sensing device has an abnormality. If the number of consecutive anomalies reaches the first preset value, it is determined that the corresponding sensing device has an abnormality.

[0082] In another embodiment provided by the present disclosure, the above step 2, when the ratio of the first data amount to the preset data amount satisfies a preset range, determining that a data transmission abnormality occurs, includes:

[0083] When the ratio of the first data amount to the preset data amount satisfies a first preset range, determining that a small amount of packet loss occurs;

[0084] When the ratio of the first data amount to the preset data amount satisfies a second preset range, determining that a large amount of packet loss occurs;

[0085] The above S104, when it is determined that the sensing device is abnormal, reporting the abnormal situation, includes:

[0086] When it is determined that an abnormality occurs in the sensing device, each abnormality of the continuous data transmission abnormality is reported; the abnormality includes: a small amount of packet loss and / or a large amount of packet loss.

[0087] In the disclosed embodiment, the detection result of the data transmission anomaly detection is the situation that an anomaly exists, which can be divided into multiple sub-situations. One of the sub-situations is the occurrence of a small amount of packet loss: if the ratio of the first data volume to the preset data volume in the current detection cycle satisfies the first preset range, it is determined that a small amount of packet loss occurs; another sub-situation is the occurrence of a large amount of packet loss: if the ratio of the first data volume to the preset data volume in the current detection cycle satisfies the second preset range, it is determined that a large amount of packet loss occurs. Therefore, when the above-mentioned step S104 is implemented, in the case of determining that an abnormality has occurred in the sensing device, for the first preset value of continuous detection results detected including the current cycle, if a small amount of packet loss and / or a large amount of packet loss are included, the specific abnormal situation including a small amount of packet loss and / or a large amount of packet loss will be reported. For example: the first preset value is 3, and the abnormal situations detected three times are: a large amount of packet loss, a small amount of packet loss and a large amount of packet loss, respectively, then the corresponding abnormal situations of these three times will all be reported.

[0088] In a possible implementation, if a small amount of packet loss occurs, the ratio between the amount of received perception data and the theoretical amount of data is large, that is, the ratio of the first data amount to the preset data amount is large, and when a large amount of packet loss occurs, the ratio between the amount of received perception data and the theoretical amount of data is small, that is, the ratio of the first data amount to the preset data amount is small, so when setting, the ratio within the above-mentioned first preset range is greater than the ratio within the second preset range (for example, the first preset range is set to 70%-90%, and the second preset range is set to 0%-70%). In the above-mentioned step 2, when the ratio of the first data amount to the preset data amount satisfies the preset range, it is determined that the preset range in which the data transmission anomaly occurs includes the first preset range and the second preset range (for example, the preset range is set to 0%-90%).

[0089] like Figure 2As shown, after entering the abnormality detection cycle, when data transmission abnormality detection is performed on the sensing data, in a possible implementation manner, in order to facilitate the accumulation of the number of consecutive data transmission abnormalities, a first counter may be set, and the process of detecting an abnormality in a sensing device may include the following steps:

[0090] S201. When a detection period arrives, compare a first data volume of the sensing data received in the detection period with a preset data volume;

[0091] S202, determining whether the ratio of the first data volume to the preset data volume satisfies a first preset range, if so, determining that a small amount of packet loss occurs, and proceeding to step S205, if not, proceeding to step S203;

[0092] S203, determining whether the ratio of the first data volume to the preset data volume meets the second preset range, if so, determining that a large number of packets are lost, and proceeding to step S205; if not, proceeding to step S204;

[0093] S204, clear the first counter and proceed to step S201;

[0094] S205, determining that the abnormality detection result is abnormal; accumulating the first counter, and recording the abnormality corresponding to this accumulation;

[0095] S206, determining whether the current accumulated number of abnormal data transmission times of the first counter has reached a first preset value; if not, proceeding to step S201; if reached, proceeding to step S207;

[0096] S207, determine that an abnormality occurs in the sensing device, and report the accumulated first preset value of consecutive transmission abnormalities respectively corresponding to the abnormal situation; clear the first counter.

[0097] In another embodiment provided by the present disclosure, the abnormality detection in the above S102 includes: sensing distance abnormality detection;

[0098] In the above S102, for each detection period, anomaly detection is performed on the sensed data received in the current detection period, including:

[0099] For each detection cycle, determining the perception distances corresponding to the perception data obtained in the detection cycle; comparing the determined perception distances with a preset perception distance range;

[0100] In the above S103, when the detection result of the current detection cycle indicates that an abnormality exists, judging whether the number of continuously detected abnormalities reaches a preset value according to the historical detection results, and if so, determining that an abnormality occurs in the sensing device, including:

[0101] Step 1: When the amount of sensing data whose sensing distance exceeds the preset sensing distance range accounts for a proportion of the amount of sensing data acquired in the current detection period that reaches a preset proportion, it is determined that a sensing distance abnormality occurs;

[0102] Step 2: determine whether the number of consecutive occurrences of perception distance anomalies represented by the detection results of the historical detection cycle and the current detection cycle since the last time the perception device was determined to be abnormal has reached a second preset value;

[0103] Step 3: When the second preset value is reached, determining that an abnormality occurs in the sensing device;

[0104] The above S104, when it is determined that the sensing device is abnormal, reporting the abnormal situation, includes:

[0105] When it is determined that the sensing device has an abnormality, the sensing distance abnormality is reported.

[0106] In the embodiments of the present disclosure, if there is point cloud occlusion (taking lidar as an example) or data anomaly during the perception data collection process of the perception device, the perception distance represented by the perception data will be too long or too short, that is, it will exceed the preset threshold and cannot normally reflect the perceived environmental information. Therefore, the anomaly detection provided in the embodiments of the present disclosure also includes perception distance anomaly detection.

[0107] Similarly, the perception distance anomaly detection can also be performed on the perception data received in each cycle according to the data collection time. For each detection cycle, the perception distance corresponding to the perception data obtained in this detection cycle is determined; the determined perception distance is compared with the preset perception distance range. If the data volume of the perception data whose perception distance exceeds the preset perception distance range accounts for the data volume of the perception data obtained in this detection cycle, it is determined that the perception distance anomaly occurs. At this time, the perception device is not blindly reported to have an abnormality, but based on the detection results of this detection cycle and the detection results corresponding to the historical detection cycles before this detection cycle, it is determined whether the number of consecutive abnormalities reaches the second preset value. If it reaches the second preset value, it is determined that the perception device has a perception distance anomaly. Therefore, when step S104 is implemented, in the case of determining that the perception device has an abnormality, for each detection result of the second preset value of consecutive cycles including the current cycle, if the detection results include a perception distance anomaly, the specific abnormality of the perception distance anomaly is reported.

[0108] like Figure 3 As shown, after entering the abnormality detection cycle, when performing data distance abnormality detection on the sensing data, in a possible implementation, in order to facilitate the accumulation of the number of consecutive data distance abnormalities, a second counter may be set, and the process of detecting an abnormality in the sensing device may include the following steps:

[0109] S301: When a detection period is reached, determining the sensing distances corresponding to the sensing data acquired in the current detection period;

[0110] S302, comparing the determined perception distance with a preset perception distance range;

[0111] S303, determining whether the proportion of the amount of the sensing data whose sensing distance exceeds the preset sensing distance range to the amount of the sensing data acquired in the current detection period reaches a preset proportion, if so, executing step S305, if not, executing step S304;

[0112] S304: Determine that there is no data sensing distance anomaly in the current detection cycle, clear the second counter, and proceed to step S301;

[0113] S305, determining that there is a data perception distance anomaly in the current detection period, accumulating the second counter, and recording the anomaly corresponding to the current accumulation;

[0114] S306, determining whether the current accumulated number of abnormal data transmission times of the second counter has reached a second preset value; if not, proceeding to step S301; if reached, proceeding to step S307;

[0115] S307: Determine that an abnormality occurs in the sensing device, and report the accumulated second consecutive preset value of distance abnormalities; and clear the second counter.

[0116] In another embodiment provided by the present disclosure, before receiving the perception data collected by the perception device, the method further includes:

[0117] Step 1: Detect whether the sensing device has sensing data output;

[0118] Step 2: In the absence of sensor data output, determine that data transmission abnormality occurs;

[0119] In the above S103, determining whether an abnormality occurs in the sensing device according to the historical abnormality detection result of the sensing device includes:

[0120] Step 1: determine whether the number of consecutive data transmission anomalies represented by the historical detection results and the current detection results since the last time the sensing device was determined to have an abnormality has reached a first preset value;

[0121] Step 2: When the first preset value is reached, determining that an abnormality occurs in the sensing device;

[0122] The above S104, when it is determined that the sensing device is abnormal, reporting the abnormal situation, includes:

[0123] When it is determined that an abnormality occurs in the sensing device, each abnormality of the continuous data transmission abnormalities is reported; the abnormality includes: the sensing device has no sensing data output.

[0124] In the disclosed embodiment, before entering the detection cycle, a pre-detection step may also be included. In the pre-detection, it can be detected whether the sensing device has a sensing data output. If the sensing device does not have a sensing data output, it can also be determined that the sensing device may have a data transmission anomaly. It can be seen that the abnormal situation of data transmission anomaly includes the sensing device having no sensing data output in addition to the above-mentioned small amount of packet loss and large amount of packet loss. When it is determined in the pre-detection that the sensing device has no sensing data output, it can be accumulated together with the detection results of the above-mentioned small amount of packet loss and large amount of packet loss, and it is determined whether the number of continuous data transmission anomalies reaches the first preset value. When the first preset value is reached, it is determined that the sensing device has an abnormality. If an abnormality occurs in which there is no sensing data, when executing the above-mentioned step S104, the specific abnormal situation that the corresponding sensing device has no sensing data output is reported. For example: the first preset value is 3, and the abnormal situations detected three times are: large amount of packet loss, small amount of packet loss and no sensing data output, then the corresponding abnormal situations of these three times are all reported.

[0125] In another embodiment provided by the present disclosure, before receiving the perception data collected by the perception device, the method further includes:

[0126] Step 3: When there is a sensed data output, determine whether the time difference between the time of entering the previous detection cycle and the current time reaches a preset time difference;

[0127] Step 4: When the preset time difference is reached, enter the next detection cycle and record the time of entering the next detection cycle.

[0128] In the embodiment of the present disclosure, before executing step S101 and receiving the perception data collected by the perception device, the above-mentioned step one is carried out to detect whether the perception device has perception data output. If it is determined that there is perception data output, a preset time difference is set according to the cycle length during implementation, and it is determined whether the time difference between the time of entering the previous detection cycle and the current time reaches the preset time difference. If the preset time difference is reached, it is determined to enter the next detection cycle, and the time of entering the next detection cycle is recorded to facilitate determination of the entry time of the subsequent cycle.

[0129] In the embodiment of the present disclosure, before executing the above step 3 and determining whether the time difference between the time of entering the previous detection cycle and the current time reaches the preset time difference when there is perception data output, the present disclosure may further include checking the packet loss detection switch for data transmission anomaly detection, which may be implemented as follows:

[0130] Check whether the packet loss detection switch is on or off. If it is on, execute the above step three to detect whether the next detection cycle has arrived. If it is off, control the packet loss detection switch to be turned on and enter the above step three, or return a notification to turn on the packet loss detection switch.

[0131] like Figure 4 As shown, in a possible implementation, the data transmission anomaly detection may include the following pre-detection: for the sensed data, including whether there is sensed data output, whether the packet loss detection switch is turned on, and whether the time difference between the time of entering the previous detection cycle and the current time reaches a preset time difference, the data transmission anomaly detection may include the following steps:

[0132] S401, detecting whether the sensing device has sensing data output, if no sensing data is output, executing step S407, if there is sensing data output, executing step S402;

[0133] S402, check whether the packet loss detection switch is turned on, if it is turned on, execute step S403, if it is not turned on, turn on the packet loss detection switch, and then go to step S403;

[0134] S403, determine whether the time difference between the time of entering the previous detection cycle and the current time reaches a preset time difference, if so, execute step S404, if not, execute step S403;

[0135] S404: When the detection period arrives, compare the first data volume of the sensing data received in the detection period with the preset data volume;

[0136] S405, determining whether the ratio of the first data volume to the preset data volume satisfies a first preset range, if so, determining that a small amount of packet loss occurs, and executing step S407, if not, executing step S406;

[0137] S406, determining whether the ratio of the first data volume to the preset data volume satisfies a second preset range, if so, determining that a large number of packet losses occur, and executing step S407, if not, clearing the first counter, and executing step S401;

[0138] S407, determining that the abnormality detection result is abnormal; accumulating the first counter, and recording the abnormality corresponding to this accumulation;

[0139] S408, determining whether the current accumulated number of abnormal data transmission times of the first counter has reached a first preset value; if not, proceeding to step S401; if reached, proceeding to step S409;

[0140] S409, determine that an abnormality occurs in the sensing device, and report the accumulated first preset value of consecutive transmission abnormalities respectively corresponding to the abnormal situation; clear the first counter.

[0141] In another embodiment provided by the present disclosure, the above method further includes:

[0142] Step 1: When it is determined that the sensing device is abnormal, the detection result of the sensing device is monitored;

[0143] Step 2: when the sensing device is detected to be back to normal, determine whether the number of detection cycles in which the detection result is continuously normal since the sensing device is detected to be back to normal has reached a third preset value;

[0144] Step 3: When the third preset value is reached, determine that the sensing device returns to normal and stop monitoring;

[0145] Step 4: When the third preset value is not reached, continue to monitor the detection results of the sensing device.

[0146] In the disclosed embodiment, when it is determined that an abnormality has occurred in the perception device, the detection results of the perception device are monitored and it is determined whether the perception data has returned to normal. By confirming that the fault has been resolved through continuous and stable data detection, false alarms that the perception device has recovered due to accidental recovery are avoided. Before the number of detection cycles in which the detection results have been normal since the perception device has returned to normal has reached a third preset value, once the detection results of the perception device indicate an abnormality, the number of detection cycles in which the perception device has returned to normal is counted again. This not only reduces false alarms that the fault has been recovered, but also ensures the continuity and stability of the unmanned vehicle's operating process, greatly improving the efficiency of the on-site operations of the autonomous driving vehicle and the overall operating efficiency of the perception system.

[0147] In another embodiment provided by the present disclosure, the above step 2, when monitoring that the sensing device has returned to normal, determining whether the number of detection cycles in which the detection result has been continuously normal since the sensing device has returned to normal has reached a third preset value, and the above step 3, when the third preset value is reached, determining that the sensing device has returned to normal and stopping monitoring, includes:

[0148] Step 1: For each detection cycle monitored, if the detection result of the current detection cycle is normal, the counter is counted cumulatively;

[0149] Step 2: If the detection result of this detection cycle is abnormal, the counter is cleared;

[0150] Step 3, judging the cumulative counting result of the counter;

[0151] Step 4: When the cumulative count of the counter reaches a third preset value, and / or the cumulative count result of the counter indicates that the sensing device has been working normally for a preset time, determine that the sensing device has returned to normal and stop monitoring.

[0152] In the disclosed embodiment, whether the sensing device has returned to normal can be determined by the number of detection cycles in which the device has been continuously detected to be back to normal, or by the continuous working time in which the device has been detected to be back to normal. Taking the duration of each cycle as 1 second and the third preset value as 5 as an example, when the cumulative count of the counter reaches the third preset value 5, and the sensing device has been monitored to be back to normal for 5 consecutive detection cycles, it can be determined that the sensing device has returned to normal; and the duration of each cycle is 1 second, it can also be determined that the sensing device has returned to normal by the continuous normal working time of the sensing device reaching the preset time of 5 seconds. There is no restriction on which monitoring method is used.

[0153] like Figure 5 As shown, the above monitoring and determination of whether the sensing device has returned to normal ensures that the system can automatically recover after the data returns to normal and lasts for a certain period of time, thereby avoiding unnecessary manual intervention and maintenance costs. In a possible implementation, the following steps are included:

[0154] S501. When it is determined that an abnormality occurs in the sensing device, monitoring the detection result of the sensing device;

[0155] S502, when the sensing device is monitored to be restored to normal, for each detection cycle after the sensing device is restored to normal, determine whether the detection result is normal, if so, execute step S503, if not, execute step S504;

[0156] S503, count the counter cumulatively, and execute step S505;

[0157] S504, clear the counter and execute step S501;

[0158] S505, determining whether the accumulated count of the counter reaches a third preset value, if so, executing step S506, if not, executing step S502;

[0159] S506: Determine that the sensing device has returned to normal and stop monitoring, and this process ends.

[0160] In another embodiment provided by the present disclosure, the method further comprises:

[0161] Step 1: Establish a big data model based on the historical data of reported abnormal results of sensing devices;

[0162] Step 2: for each sensing device installed on the unmanned vehicle, obtain the abnormal frequencies corresponding to different abnormal types of the sensing device;

[0163] Step 3: Determine a high-risk sensing device whose abnormal frequency is higher than a first threshold, and an abnormal type corresponding to the abnormal frequency;

[0164] Step 4: Replace or maintain the high-risk sensing device according to the abnormality type;

[0165] and / or

[0166] Step 1: Establish a big data model based on the historical data of reported abnormal results of sensing devices;

[0167] Step 2: For each unmanned vehicle, obtain the abnormal frequency of all sensing devices installed on the unmanned vehicle;

[0168] Step 3: Screen the unmanned vehicles to obtain high-risk unmanned vehicles whose abnormal frequency is higher than a second threshold;

[0169] Step 4: Replace or maintain the high-risk unmanned vehicle;

[0170] The anomaly detection method provided by the disclosed embodiment can be applied to different sensing devices installed at different positions of unmanned vehicles. Based on the reported results of the anomalies of each sensing device, a big data model can be established (for example: a statistical database can be established), and historical data can be backtracked according to the fault detection and reporting strategy (i.e., monitoring of the frequency of abnormalities of sensing devices and monitoring of the frequency of abnormalities of unmanned vehicles). The failure frequency and type of the hardware of different sensing devices on different vehicles, as well as unmanned vehicles with frequent failures, can be obtained. The purpose of early prediction and preventive maintenance can be achieved through the above steps. By analyzing historical abnormal data and operating modes, the model can predict and identify potential faults, thereby achieving early detection and timely repair of faults. This not only reduces the frequency of faults in unmanned driving systems, but also reduces the workload of the maintenance team, and improves maintenance efficiency and equipment availability.

[0171] Based on the same disclosed concept, the embodiments of the present disclosure also provide an abnormality detection device and an unmanned vehicle. Since the principles of the problems solved by these devices and unmanned vehicles are similar to those of the aforementioned abnormality detection methods, the implementation of the device and the unmanned vehicle can refer to the implementation of the aforementioned methods, and the repeated parts will not be repeated.

[0172] With the above Figure 1 Corresponding to the method shown, the embodiment of the present disclosure also provides an abnormality detection device, such as Figure 6 As shown, including:

[0173] The data receiving module 601 is used to receive the sensing data collected by the sensing device;

[0174] The data anomaly detection module 602 is used to perform anomaly detection on the sensed data received in each detection period;

[0175] The device anomaly detection module 603 is used to determine whether the number of consecutively detected anomalies reaches a preset value based on historical detection results when the detection result of the current detection cycle indicates that an anomaly exists, and if so, determine that an anomaly occurs in the sensing device;

[0176] The abnormality reporting module 604 is used to report the abnormality when it is determined that the sensing device has an abnormality.

[0177] In another embodiment provided by the present disclosure, the data anomaly detection module 702 is used for data transmission anomaly detection; for each detection cycle, the first data volume of the sensed data received in the detection cycle is compared with the preset data volume;

[0178] The above-mentioned device abnormality detection module 603 is used to determine that a data transmission abnormality has occurred when the ratio of the first data volume to the preset data volume satisfies a preset range; determine whether the number of consecutive data transmission abnormalities represented by the historical detection results and the detection results of the current detection cycle since the last time the perception device was determined to have an abnormality has reached a first preset value; and when the first preset value is reached, determine that the perception device has an abnormality.

[0179] In another embodiment provided by the present disclosure, the device abnormality detection module 603 is used to determine that a small amount of packet loss occurs when the ratio of the first data amount to the preset data amount meets a first preset range; and to determine that a large amount of packet loss occurs when the ratio of the first data amount to the preset data amount meets a second preset range;

[0180] The above-mentioned abnormality reporting module 604 is used to report the abnormal situation of each abnormality in the continuous data transmission abnormality when it is determined that the sensing device has an abnormality; the abnormal situation includes: a small amount of packet loss and / or a large amount of packet loss.

[0181] In another embodiment provided by the present disclosure, the data anomaly detection module 602 is used for sensing distance anomaly detection; for each detection cycle, anomaly detection is performed on the sensing data received in the current detection cycle, including: for each detection cycle, determining the sensing distance corresponding to the sensing data obtained in the current detection cycle; comparing the determined sensing distance with a preset sensing distance range;

[0182] The device abnormality detection module 603 is used to determine that a perception distance abnormality occurs when the proportion of the amount of perception data whose perception distance exceeds the preset perception distance range to the amount of perception data acquired in the current detection period reaches a preset proportion; determine whether the number of consecutive occurrences of perception distance abnormalities represented by the detection results of the historical detection period and the current detection period since the last time the perception device was determined to have an abnormality reaches a second preset value; and determine that the perception device has an abnormality when the second preset value is reached;

[0183] The above-mentioned abnormality reporting module 604 is used to report the sensing distance abnormality when it is determined that the sensing device has an abnormality.

[0184] In another embodiment provided by the present disclosure, the data anomaly detection module 602 is further used to detect whether the sensing device has sensed data output before receiving the sensed data collected by the sensing device; if there is no sensed data output, determine that a data transmission anomaly occurs;

[0185] The device anomaly detection module 603 is used to determine whether the number of consecutive data transmission anomalies represented by the historical detection results and the current detection results since the last time the sensing device was determined to be abnormal has reached a first preset value; if the first preset value is reached, it is determined that the sensing device has an abnormality;

[0186] The above-mentioned abnormality reporting module 604 is used to report the abnormal situation of each abnormality in the continuous data transmission abnormality when it is determined that the sensing device has an abnormality; the abnormal situation includes: the sensing device has no sensing data output.

[0187] In another embodiment provided by the present disclosure, the above-mentioned data receiving module 601 is used to determine whether the time difference between the time of entering the previous detection cycle and the current time reaches a preset time difference when there is perception data output; when the preset time difference is reached, enter the next detection cycle and record the time of entering the next detection cycle.

[0188] In another embodiment provided by the present disclosure, Figure 7 As shown, the above-mentioned device also includes: an abnormal recovery detection module 605, which is used to monitor the detection results of the sensing device when it is determined that an abnormality has occurred in the sensing device; when it is monitored that the sensing device has returned to normal, determine whether the number of detection cycles in which the detection results have been continuously normal since the sensing device has returned to normal has reached a third preset value; when the third preset value is reached, determine that the sensing device has returned to normal and stop monitoring; when the third preset value is not reached, continue to monitor the detection results of the sensing device.

[0189] In another embodiment provided by the present disclosure, the above-mentioned device abnormality detection module 603 is used to, for each detection cycle of monitoring, if the detection result of the current detection cycle is normal, then accumulate the count of the counter; if the detection result of the current detection cycle is abnormal, then clear the counter; judge the accumulated counting result of the counter; when the accumulated count of the counter reaches a third preset value, and / or the accumulated counting result of the counter indicates that the continuous normal working time of the sensing device reaches a preset time, determine that the sensing device has returned to normal and stop monitoring.

[0190] With the above Figure 1 Corresponding to the method shown, an embodiment of the present disclosure also provides an unmanned vehicle, including an abnormality detection device as provided in any of the above embodiments.

[0191] Through the description of the above implementation methods, those skilled in the art can clearly understand that the embodiments of the present disclosure can be implemented by hardware, or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present disclosure.

[0192] Those skilled in the art will appreciate that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes in the accompanying drawings are not necessarily required for implementing the present disclosure.

[0193] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the present embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.

[0194] The serial numbers of the above-mentioned embodiments of the present disclosure are only for description and do not represent the advantages or disadvantages of the embodiments.

[0195] Obviously, those skilled in the art can make various changes and modifications to the present disclosure without departing from the spirit and scope of the present disclosure. Thus, if these modifications and variations of the present disclosure fall within the scope of the claims of the present disclosure and their equivalents, the present disclosure is also intended to include these modifications and variations.

Claims

1. An anomaly detection method, characterized in that: include: Receiving perception data collected by the perception device; For each detection cycle, anomaly detection is performed on the sensing data received in this detection cycle; If the detection result of the current detection cycle indicates that an abnormality exists, it is determined whether the number of continuously detected abnormalities reaches a preset value based on the historical detection results. If the preset value is reached, it is determined that the sensing device has an abnormality. When it is determined that the sensing device has an abnormality, the abnormality is reported.

2. The method according to claim 1, characterized in that The anomaly detection includes: data transmission anomaly detection; For each detection cycle, anomaly detection is performed on the sensing data received in this detection cycle, including: For each detection cycle, comparing a first data volume of the sensing data received in the current detection cycle with a preset data volume; In the case where the detection result of the current detection cycle indicates that an abnormality exists, judging whether the number of continuously detected abnormalities reaches a preset value according to the historical detection results, and if the preset value is reached, determining that an abnormality occurs in the sensing device, includes: When the ratio of the first data amount to the preset data amount satisfies a preset range, determining that a data transmission abnormality occurs; Determine whether the number of consecutive data transmission anomalies represented by the historical detection results and the detection results of the current detection cycle since the last time the sensing device was determined to have an anomaly has reached a first preset value; When the first preset value is reached, it is determined that an abnormality occurs in the sensing device.

3. The method according to claim 2, characterized in that When the ratio of the first data amount to the preset data amount satisfies a preset range, determining that a data transmission abnormality occurs includes: When the ratio of the first data amount to the preset data amount satisfies a first preset range, determining that a small amount of packet loss occurs; When the ratio of the first data amount to the preset data amount satisfies a second preset range, determining that a large amount of packet loss occurs; When it is determined that the sensing device is abnormal, reporting the abnormal situation includes: When it is determined that the sensing device has an abnormality, each abnormality of the continuous data transmission abnormalities is reported; the abnormality includes: a small amount of packet loss and / or a large amount of packet loss.

4. The method according to claim 1, characterized in that The anomaly detection includes: sensing distance anomaly detection; For each detection cycle, anomaly detection is performed on the sensing data received in this detection cycle, including: For each detection cycle, determining the perception distances corresponding to the perception data obtained in the detection cycle; comparing the determined perception distances with a preset perception distance range; When the detection result of the current detection cycle indicates that an abnormality exists, judging whether the number of continuously detected abnormalities reaches a preset value according to the historical detection results, and if the preset value is reached, determining that an abnormality occurs in the sensing device, including: When the amount of the sensing data whose sensing distance exceeds the preset sensing distance range accounts for a proportion of the amount of the sensing data acquired in the current detection period that reaches a preset proportion, it is determined that a sensing distance abnormality occurs; Determine whether the number of consecutive occurrences of perception distance anomalies represented by the detection results of the historical detection cycle and the current detection cycle since the last time the perception device was determined to have an abnormality has reached a second preset value; When the second preset value is reached, determining that an abnormality occurs in the sensing device; When it is determined that the sensing device is abnormal, reporting the abnormal situation includes: When it is determined that the sensing device is abnormal, the sensing distance abnormality is reported.

5. The method according to any one of claims 1 to 4, characterized in that: Before receiving the sensing data collected by the sensing device, it also includes: Detecting whether the sensing device has sensing data output; In the absence of sensor data output, determining that data transmission anomaly occurs; Determining whether an abnormality occurs in the sensing device according to a historical abnormality detection result of the sensing device includes: Determine whether the number of consecutive data transmission anomalies represented by the historical detection results and the current detection results since the last time the sensing device was determined to have an anomaly has reached a first preset value; When the first preset value is reached, determining that an abnormality occurs in the sensing device; When it is determined that the sensing device is abnormal, reporting the abnormal situation includes: When it is determined that the sensing device has an abnormality, each abnormality of the consecutive data transmission abnormalities is reported; the abnormality includes: the sensing device has no sensing data output.

6. The method according to claim 5, characterized in that In the case of sensory data output, it also includes: Determine whether the time difference between the time of entering the previous detection cycle and the current time reaches a preset time difference; When the preset time difference is reached, the next detection cycle is entered, and the time of entering the next detection cycle is recorded.

7. The method according to claim 1, characterized in that Also includes: When it is determined that the sensing device is abnormal, monitoring the detection result of the sensing device; In the case where it is monitored that the sensing device has returned to normal, determining whether the number of detection cycles in which the detection result is continuously normal since the sensing device has returned to normal has reached a third preset value; When the third preset value is reached, determining that the sensing device has returned to normal and stopping monitoring; When the third preset value is not reached, the detection result of the sensing device continues to be monitored.

8. The method according to claim 7, characterized in that In the case where it is monitored that the sensing device has returned to normal, determining whether the number of detection cycles in which the detection result is continuously normal since the sensing device has returned to normal has reached a third preset value; When the third preset value is reached, determining that the sensing device returns to normal and stops monitoring includes: For each detection cycle monitored, if the detection result of the current detection cycle is normal, the counter is counted cumulatively; if the detection result of the current detection cycle is abnormal, the counter is cleared; Determining the cumulative counting result of the counter; When the accumulated count of the counter reaches a third preset value, and / or the accumulated count result of the counter indicates that the sensing device has been working normally for a preset time, it is determined that the sensing device has returned to normal and monitoring is stopped.

9. An abnormality detection device, characterized in that: include: A data receiving module, used to receive the sensing data collected by the sensing device; The data anomaly detection module is used to perform anomaly detection on the sensed data received in each detection cycle; The device anomaly detection module is used to determine whether the number of consecutively detected anomalies reaches a preset value based on historical detection results when the detection results of the current detection cycle indicate that an anomaly exists, and if so, determine that an anomaly occurs in the sensing device; The abnormality reporting module is used to report the abnormal situation when it is determined that the sensing device has an abnormality.

10. An unmanned vehicle, characterized in that: Comprising an abnormality detection device as claimed in claim 9.

Images