Method, apparatus, device and storage medium for collision detection

Abstract

According to embodiments of this disclosure, a method, apparatus, electronic device, and storage medium for collision detection are provided. The method includes: acquiring position information of an obstacle, the position information indicating a circumscribed rectangular region and a polygonal contour region of the obstacle; determining first overlap information between the circumscribed rectangular region and a target region of an autonomous vehicle, and second overlap information between the polygonal contour region and the target region; determining a first weight corresponding to the first overlap information and a second weight corresponding to the second overlap information based on acceleration information of the autonomous vehicle, the acceleration information indicating a target acceleration independent of control signals of the autonomous vehicle; determining weighted information of the first overlap information and the second overlap information based on the first weight and the second weight; and determining a target confidence level of a collision between the autonomous vehicle and the obstacle based on the weighted information.

Description

Technical Field

[0001] The exemplary embodiments disclosed herein generally relate to the field of computers, and particularly to methods, apparatus, devices, and computer-readable storage media for collision detection. Background Technology

[0002] With the rapid development of transportation and the improvement of people's living standards, more and more people are choosing vehicle-based travel. Vehicle safety has always been considered the most important issue, and how to efficiently and accurately detect whether a collision has occurred is a key concern. Summary of the Invention

[0003] In a first aspect of this disclosure, a method for collision detection is provided. The method includes: acquiring position information of an obstacle, the position information indicating a circumscribed rectangular region and a polygonal contour region of the obstacle; determining first overlap information between the circumscribed rectangular region and a target region of an autonomous vehicle, and second overlap information between the polygonal contour region and the target region; determining a first weight corresponding to the first overlap information and a second weight corresponding to the second overlap information based on acceleration information of the autonomous vehicle, the acceleration information indicating a target acceleration independent of a control signal of the autonomous vehicle; determining weighted information of the first overlap information and the second overlap information based on the first weight and the second weight; and determining a target confidence level of a collision between the autonomous vehicle and the obstacle based on the weighted information.

[0004] In a second aspect of this disclosure, an apparatus for collision detection is provided. The apparatus includes: an information acquisition module configured to acquire position information of an obstacle, the position information indicating a bounding rectangular region and a polygonal contour region of the obstacle; a first determination module configured to determine first overlap information between the bounding rectangular region and a target region of an autonomous vehicle, and second overlap information between the polygonal contour region and the target region; a second determination module configured to determine a first weight corresponding to the first overlap information and a second weight corresponding to the second overlap information based on acceleration information of the autonomous vehicle, the acceleration information indicating a target acceleration independent of a control signal of the autonomous vehicle; a third determination module configured to determine weighted information of the first overlap information and the second overlap information based on the first weight and the second weight; and a fourth determination module configured to determine a target confidence level of a collision between the autonomous vehicle and the obstacle based on the weighted information.

[0005] In a third aspect of this disclosure, an electronic device is provided. The device includes at least one processing unit; and at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit. When executed by the at least one processing unit, the instructions cause the device to perform the method of the first aspect.

[0006] In a fourth aspect of this disclosure, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program that can be executed by a processor to implement the method of the first aspect.

[0007] It should be understood that the content described in this content section is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0008] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0009] Figure 1 A schematic diagram of an example environment in which embodiments of the present disclosure can be implemented is shown;

[0010] Figure 2 A flowchart of a collision detection process according to some embodiments of the present disclosure is shown;

[0011] Figure 3 This disclosure provides a schematic diagram of collision detection according to some embodiments of the present disclosure;

[0012] Figure 4 A schematic structural block diagram of an apparatus for collision detection according to certain embodiments of the present disclosure is shown;

[0013] Figure 5 A block diagram of an electronic device capable of implementing several embodiments of the present disclosure is shown. Detailed Implementation

[0014] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0015] It should be noted that the headings of any section / subsection provided herein are not limiting. Various embodiments are described throughout this document, and embodiments of any type may be included under any section / subsection. Furthermore, embodiments described in any section / subsection may be combined in any way with any other embodiments described in the same section / subsection and / or different sections / subsections.

[0016] In the description of embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may also be included below. The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0017] The embodiments of this disclosure may involve user data, data acquisition, and / or use. All of these aspects comply with applicable laws, regulations, and relevant provisions. In the embodiments of this disclosure, all data collection, acquisition, processing, manipulation, forwarding, and use are conducted with the user's knowledge and confirmation. Accordingly, in implementing the embodiments of this disclosure, the type, scope of use, and usage scenarios of any data or information that may be involved should be communicated to the user and their authorization obtained in accordance with relevant laws and regulations through appropriate means. The specific methods of notification and / or authorization may vary depending on the actual situation and application scenario, and the scope of this disclosure is not limited in this respect.

[0018] In this specification and the embodiments, any processing of personal information will be carried out only under the premise of legality (such as obtaining the consent of the personal information subject, or being necessary for the performance of a contract), and will only be carried out within the scope stipulated or agreed upon. A user's refusal to process personal information other than that necessary for basic functions will not affect the user's use of basic functions.

[0019] Example Environment

[0020] Figure 1 A schematic diagram of an example environment 100 in which several embodiments of the present disclosure can be implemented is shown. Typical objects are schematically shown in this example environment 100, including an autonomous vehicle 110 and obstacles 120. Figure 1 In the example, autonomous vehicle 110 can be any type of vehicle capable of carrying people and / or goods and moving via a power system such as an engine. Examples of autonomous vehicle 110 include, but are not limited to, cars, trucks, buses, electric vehicles, motorcycles, RVs, trains, etc. One or more autonomous vehicles 110 in environment 100 are vehicles with some degree of assisted driving capability or autonomous driving capability; such vehicles are also referred to as intelligent driving vehicles.

[0021] The autonomous vehicle 110 can be communicatively coupled to the computing device 130. Although shown as a separate entity, the computing device 130 may also be embedded within the autonomous vehicle 110. Alternatively, the computing device 130 may be an entity external to the autonomous vehicle 110 and may communicate with the autonomous vehicle 110 via a wireless network. For example, the computing device 130 may be deployed on a roadside or as a remote server. The computing device 130 may be implemented as one or more computing devices, which at least include a processor, memory, and other components typically found in general-purpose computers to perform functions such as computing, storage, communication, and control.

[0022] Obstacle 120 can be any type of vehicle capable of carrying people and / or objects and moving via a power system such as an engine. Examples of obstacle 120 include, but are not limited to, cars, trucks, buses, electric vehicles, motorcycles, motorhomes, trains, etc. Obstacle 120 can also be other movable or fixed objects that are not vehicles, such as railings, walls, roadblocks, etc.

[0023] It should be understood that the structure and function of the various elements in environment 100 are described for illustrative purposes only and do not imply any limitation on the scope of this disclosure.

[0024] In related technologies, the autonomous vehicle 110 calculates the overlap area between its own outline and the bounding boxes perceived by surrounding objects (obstacles 120), and combines this with changes in the vehicle's IMU sensors to comprehensively determine whether a collision has occurred. However, this method has a drawback: it is highly sensitive to false detections and false negatives. False detections of object type or shape can cause temporary overlap of bounding boxes, especially when the autonomous vehicle 110 and the obstacle 120 briefly intersect or change rapidly, leading to false collision detections and increasing the false alarm rate. False negatives, on the other hand, occur when environmental information is complex, noise levels are high, or there are obstructions, preventing the system from accurately detecting all objects and affecting the overall reliability of the detection.

[0025] Embodiments of this disclosure propose a collision detection scheme, which acquires the location information of an obstacle, the location information indicating the circumscribed rectangular region and the polygonal contour region of the obstacle; determines first overlap information between the circumscribed rectangular region and the target region of an autonomous vehicle, and second overlap information between the polygonal contour region and the target region; based on the acceleration information of the autonomous vehicle, determines a first weight corresponding to the first overlap information and a second weight corresponding to the second overlap information, the acceleration information indicating the target acceleration independent of the control signal of the autonomous vehicle; and based on the first weight and the second weight, determines weighted information of the first overlap information and the second overlap information; and based on the weighted information, determines the target confidence level of a collision between the autonomous vehicle and the obstacle.

[0026] According to embodiments of this disclosure, the autonomous vehicle 110 can, on the one hand, perform weighted calculations based on the contour detection of the obstacle 120 and the acceleration of the autonomous vehicle 110, thereby improving the calculation accuracy of the overlapping portion between the autonomous vehicle 110 and the obstacle 120 and thus enhancing the robustness of the collision detection calculation. On the other hand, when performing acceleration calculations, the autonomous vehicle 110 can isolate the influence of its own control signals, improving the correlation between the calculated target acceleration and the collision, further enhancing the reliability and accuracy of collision detection.

[0027] Example process

[0028] Figure 2 A flowchart of a collision detection process 200 according to some embodiments of the present disclosure is shown. Process 200 can be implemented at an autonomous vehicle 110 and / or a computing device 130. Reference is made below. Figure 1 Describe the process 200.

[0029] At frame 210, the autonomous vehicle 110 acquires the position information of the obstacle 120, which indicates the bounding rectangular region and polygonal outline region of the obstacle 120.

[0030] In some embodiments, the autonomous vehicle 110 can be an autonomous vehicle used to provide mobility services, such as a car, truck, or bus, which has a certain degree of assisted driving or autonomous driving capabilities. The obstacle 120 can also be any type of vehicle, or the obstacle 120 can be other movable or fixed objects that are not vehicles, such as railings, walls, roadblocks, etc.

[0031] In some embodiments, the location information of obstacle 120 may include, but is not limited to, the location information of obstacle 120, the relative position and distance information between obstacle 120 and autonomous vehicle 110, the motion tracking information of obstacle 120 such as motion direction, motion speed, and motion acceleration, and the attribute information of obstacle 120 itself, such as the shape and size of obstacle 120. The location information of obstacle 120 can be detected by sensors or other detection devices of autonomous vehicle 110; it can also be transmitted by devices such as computing device 130 that are communicatively connected to autonomous vehicle 110, and there are no limitations on this.

[0032] Figure 3 This disclosure provides a schematic diagram of collision detection according to some embodiments of the present disclosure.

[0033] like Figure 3As shown, in some embodiments, based on the tracking information of obstacle 120, the autonomous vehicle 110 can determine the circumscribed rectangular region 320 of obstacle 120; and / or based on the point cloud data associated with obstacle 120, the autonomous vehicle 110 can determine the polygonal outline region 325 of obstacle 120.

[0034] In some embodiments, the approximate position and size of obstacle 120 can be determined based on the tracking information of obstacle 120, and the autonomous vehicle 110 can then depict the approximate outline information of obstacle 120 as shown in the circumscribed rectangular region 320. The calculation of obtaining the circumscribed rectangular region 320 of obstacle 120 is simple and efficient.

[0035] In some embodiments, the autonomous vehicle 110 senses and detects the obstacle 120, removes noise from the semantic point cloud of the obstacle 120, and clusters it to generate polygons based on the semantic point cloud. This allows for the depiction of relatively accurate contour information of the obstacle 120, as shown in the polygon contour region 325. The polygon contour region 325 of the obstacle 120 can more accurately depict the position and contour information of the obstacle 120.

[0036] When the autonomous vehicle 110 detects the obstacle 120, it simultaneously retains the circumscribed rectangular region 320 and the polygonal outline region 325 for subsequent analysis and calculation, thereby balancing the accuracy and efficiency of collision detection calculation.

[0037] Similarly, in some embodiments, the target area 310 of the autonomous vehicle 110 may be a rectangular area 310-1 that is easy to calculate, or it may be a polygonal area 310-2 with higher precision, without limitation.

[0038] Continue to refer to Figure 2 At frame 220, the first overlap information between the circumscribed rectangular region 320 and the target region 310 of the autonomous vehicle 110 and the second overlap information between the polygonal outline region 325 and the target region 310 are determined.

[0039] In some embodiments, the first overlap information or the second overlap information indicates at least one of the following: the distance from the circumscribed rectangular region 320 or the polygonal outline region 325 to the target region 310; the overlap ratio between the circumscribed rectangular region 320 or the polygonal outline region 325 and the target region 310.

[0040] Specifically, in some embodiments, when the circumscribed rectangular region 320 or polygonal outline region 325 does not overlap with the target region 310, the first overlap information or the second overlap information can be used to determine the distance from the circumscribed rectangular region 320 or polygonal outline region 325 to the target region 310; when the circumscribed rectangular region 320 or polygonal outline region 325 overlaps with the target region 310, the first overlap information or the second overlap information can indicate the overlap ratio between the circumscribed rectangular region 320 or polygonal outline region 325 and the target region 310.

[0041] In some embodiments, the circumscribed rectangular region 320 or the polygonal outline region 325 corresponds to the overhead view of the obstacle 120. In other embodiments, the circumscribed rectangular region 320 or the polygonal outline region 325 may also correspond to multi-angle or stereoscopic detection graphics, and this is not limited thereto.

[0042] In some embodiments, the shape and position of the circumscribed rectangular region 320 or the polygonal outline region 325 are different, and their overlap area with the target region 310 is also different. The overlap ratios of the circumscribed rectangular region 320 or the polygonal outline region 325 with the target region 310 are set to S1 and S2, respectively, that is, the first overlap information between the obstacle 120 and the autonomous vehicle 110 is S1, and the second overlap information is S2.

[0043] At box 230, based on the acceleration information of the autonomous vehicle 110, a first weight corresponding to the first overlapping information and a second weight corresponding to the second overlapping information are determined. The acceleration information indicates the target acceleration independent of the control signal of the autonomous vehicle 110.

[0044] In the existing technology, the acceleration of an autonomous vehicle 110 is usually determined based on the detection data of the sensing and navigation equipment installed on the autonomous vehicle 110. However, during the actual driving, starting, and stopping process of the autonomous vehicle 110, the acceleration is significantly interfered with by the high-frequency noise inside the sensor, which greatly affects the accurate estimation of acceleration changes.

[0045] In some embodiments, the autonomous vehicle 110 acquires a first acceleration collected by an inertial navigation device, and performs low-pass filtering on the acceleration data collected by the inertial navigation device to determine the first acceleration a. A This reduces noise interference and improves the accuracy of acceleration detection for autonomous vehicles.

[0046] In some embodiments, the location where the inertial navigation device is installed on the autonomous vehicle 110 is uncertain, and its installation location may not coincide with the center of mass of the autonomous vehicle 110 itself. Based on the installation location of the inertial navigation device, the autonomous vehicle 110 can convert the first acceleration into a second acceleration corresponding to the center of mass of the autonomous vehicle 110; and based on the second acceleration, determine the acceleration information of the autonomous vehicle 110.

[0047] In some embodiments, the autonomous vehicle 110 can use the following formula to calculate the first acceleration a A Converted to second acceleration a B :

[0048]

[0049] in, The rotation transformation matrix represents the coordinate system of the inertial navigation device and the centroid coordinate system of the autonomous vehicle 110; This represents the translational changes between the coordinate system of the inertial navigation device and the coordinate system of the centroid of the autonomous vehicle 110; a A and a B Let ω represent the acceleration of the autonomous vehicle 110 in the coordinate system of the inertial navigation device and the coordinate system of the center of mass of the autonomous vehicle 110, respectively; A ] ∧ and [Ω] A ] ∧ These represent the antisymmetric matrix of angular velocity and the antisymmetric matrix of angular acceleration in the coordinate system of the inertial navigation device, respectively.

[0050] Using the above formula, the first acceleration a of the inertial navigation device A The second acceleration a is converted into the center-of-mass coordinate system of the autonomous vehicle 110. B This reduces the impact of the vehicle's own stiffness and the installation position of the inertial navigation equipment on acceleration detection, thereby further improving the acceleration detection accuracy of the autonomous vehicle 110.

[0051] In some embodiments, determining the acceleration information of the autonomous vehicle 110 based on the second acceleration further includes: determining a third acceleration based on the target control signal of the autonomous vehicle 110; and determining the acceleration information of the autonomous vehicle 110 based on the second acceleration and the third acceleration.

[0052] During the start-stop process of autonomous vehicle 110, the target control signal provides it with active acceleration, namely the third acceleration a. C To more accurately capture the passive acceleration caused by the collision, the second acceleration 'a' obtained after noise removal and coordinate correction is used. B Remove the third acceleration a CTo obtain more accurate passive acceleration.

[0053] During the actual start-stop process of the autonomous vehicle 110, the target control signal provides it with active acceleration. There is a time difference ΔT between the time T1 when the target control signal issues commands such as acceleration or braking and the time T2 when the autonomous vehicle 110 actually performs the acceleration or braking action. Therefore, when detecting the passive acceleration at the current moment, it is necessary to calculate it based on the active acceleration corresponding to the control signal issued before time ΔT.

[0054] In some embodiments, the third acceleration a is determined based on the target control signal of the autonomous vehicle 110. C This includes: determining the response delay ΔT corresponding to the autonomous vehicle 110; and determining the target control signal from the historical control signals of the autonomous vehicle 110 based on the response delay ΔT.

[0055] In some embodiments, the response delay ΔT is determined based on a comparison between a set of expected accelerations corresponding to a set of historical control signals and a set of historical accelerations. For example, a set of historical control signals may be a ctrl1 a ctrl2 a ctrl3 and a ctrl4 Their emission times are t 01 t 02 t 03 and t 04 And their respective expected accelerations are a 01 a 02 a 03 and a 04 The actual acceleration detected by the autonomous vehicle 110, i.e., the historical acceleration, is equal to a. 01 a 02 a 03 and a 04 The actual time intervals are t1, t2, t3, and t4. Based on the time corresponding to the expected acceleration and the time corresponding to the actual historical acceleration, the response delay ΔT can be calculated.

[0056] In some embodiments, the response delay ΔT is estimated by minimizing a cost function, which is as follows:

[0057]

[0058] Where n* represents the optimal index, i.e., the estimated value of the response delay ΔT, that minimizes the discrete cost function within the set sliding window; acc control Indicates the acceleration corresponding to the control signal; acc imu This indicates the actual acceleration detected.

[0059] In some embodiments, the second acceleration a is combined with the above-described time delay estimation results. B With the third acceleration a C Subtracting the signals yields the passive acceleration. That is, the acceleration information of the autonomous vehicle 110 is equal to the second acceleration a. B Subtract the third acceleration a before time ΔT C The calculation formula is as follows:

[0060] acc passive [k] = acc imu [k]-acc control [kn * ]

[0061] Among them, acc passive [k] represents the passive acceleration of the autonomous vehicle 110 at time k, i.e., the target acceleration indicated by the acceleration information of the autonomous vehicle 110, independent of the control signal of the autonomous vehicle 110; acc imu [k] represents the actual acceleration of autonomous vehicle 110 at time k, i.e., the second acceleration a of autonomous vehicle 110. B ;acc control [kn*] represents the target acceleration corresponding to the control signal of the autonomous vehicle 110 at time kn*, i.e., the third acceleration a of the autonomous vehicle 110. C .

[0062] The target acceleration acc of the autonomous vehicle 110, independent of its control signal, is obtained through the above calculations. In some embodiments, a first weight is positively correlated with the target acceleration acc indicated by the acceleration information, and a second weight is negatively correlated with the target acceleration acc indicated by the acceleration information. For example, the first weight is acc%, the second weight is (1-acc%), and the sum of the two is 100%. In other embodiments, the specific functional relationship between the first and second weights and the acceleration information can be adjusted according to the different characteristics and driving conditions of the autonomous vehicle 110, and is not limited here.

[0063] At box 240, the weighted information of the first overlap information and the second overlap information is determined based on the first weight and the second weight.

[0064] In some embodiments, based on the first overlap information S1 and the second overlap information S2 between the obstacle 120 and the autonomous vehicle 110, and the first weight acc% and the second weight (1-acc%), the weighted information S1*acc% of the first overlap information and the weighted information S2*(1-acc%) of the second overlap information can be obtained. In this embodiment, the weighted information of the first overlap information / second overlap information is only expressed as a product relationship, indicating that it is calculated based on the first overlap information / second overlap information and the first weight / second weight, without limiting the functional relationship between the two.

[0065] At box 250, based on weighted information, the target confidence level of collision between autonomous vehicle 110 and obstacle 120 is determined.

[0066] In some embodiments, multimodal sensing information is integrated using the following likelihood function:

[0067] likelihood=S1*acc%+S2*(1-acc%)

[0068] When the acceleration changes significantly, it indicates a higher probability of a collision, and the polygonal outline region 325 may be incomplete. In this case, it is preferable to use the relatively large circumscribed rectangular region 320 to improve recall. When the acceleration changes slightly, it may be a grazing or close encounter, and it is preferable to use the fine polygonal outline region 325 to make a judgment to reduce false detections.

[0069] In some embodiments, determining the target confidence level of an autonomous vehicle colliding with an obstacle based on weighted information includes: determining a set of historical confidence levels of the autonomous vehicle colliding with an obstacle at a set of historical times; determining the predicted confidence level corresponding to the current time based on the set of historical confidence levels; determining the observation information of the decision model based on weighted information; and adjusting the predicted confidence level based on the observation information to determine the target confidence level.

[0070] On the one hand, similar to the aforementioned process for correcting delays, historical data feedback correction improves the reliability of the target confidence likelihood detection model, thereby further enhancing the accuracy of collision detection. On the other hand, the target confidence likelihood detection model, corrected based on historical confidence, can predict collision scenarios over a future period, indicating a potential collision between the autonomous vehicle 110 and obstacle 120 at some point in the future, and thus mitigating or avoiding such collisions.

[0071] In some embodiments, based on weighted information, in response to a target confidence level greater than a threshold, it is determined whether collision information associated with an obstacle satisfies a set of preset constraints; and in response to the collision information satisfying a set of preset constraints, a collision signal associated with the obstacle is generated.

[0072] In some embodiments, the autonomous vehicle 110 presets or calculates a threshold P, which can be adjusted according to the characteristics and driving conditions of the autonomous vehicle 110 and the obstacle 120, without limitation. The target confidence level (likelihood) calculated in the above steps is compared with the threshold P. If the target confidence level (likelihood) is less than the threshold P, it is determined that no collision has occurred between the autonomous vehicle 110 and the obstacle 120. If the target confidence level (likelihood) is greater than or equal to the threshold P, it is determined that a collision may have occurred between the autonomous vehicle 110 and the obstacle 120. Further collision determination is needed based on whether the collision information associated with the obstacle meets a set of preset constraints to improve detection accuracy and reduce false detections.

[0073] In some embodiments, a set of preset constraints includes at least one of the following: a first constraint indicates that the overlap area between the polygonal contour region 325 and the target region 310 is greater than a first threshold; a second constraint indicates that the shape information of the obstacle 120 meets preset conditions; a third constraint indicates that when the speed of the autonomous vehicle 110 is less than a threshold, the target acceleration of the autonomous vehicle 110 is greater than a second threshold; and a fourth constraint indicates that the relative driving direction and relative position of the obstacle 120 relative to the autonomous vehicle 110 meet preset conditions.

[0074] In some embodiments, according to the first constraint, on the one hand, when the overlap between the polygonal contour region 325 and the target region 310 is greater than a first threshold, it can be determined that there is a large positional overlap between the obstacle 120 and the autonomous vehicle 110, that is, it can be determined that a collision has occurred between the autonomous vehicle 110 and the obstacle 120, and the autonomous vehicle 110 generates a collision signal associated with the obstacle 120. On the other hand, the polygonal contour region 325 of the autonomous vehicle 110 and the obstacle 120 is checked, and if there is no intersection, it is determined to be a false detection.

[0075] In some embodiments, according to the second constraint, collision instances of obstacles 120 with abnormal shapes (such as being too large or too small) can be filtered out to avoid false detections caused by abnormal shape output of the perception result.

[0076] In some embodiments, according to the third constraint, on the one hand, if the target acceleration of the autonomous vehicle 110 is greater than the second threshold when the speed of the autonomous vehicle 110 is less than the threshold, it can be determined that the autonomous vehicle 110 is subjected to an external force causing a large change in acceleration, that is, it can be determined that a collision has occurred between the autonomous vehicle 110 and the obstacle 120, and the autonomous vehicle 110 generates a collision signal associated with the obstacle 120. On the other hand, when the autonomous vehicle 110 is stationary, collision instances with small changes in acceleration of the autonomous vehicle 110 are filtered out to avoid false detections caused by slight contact with pedestrians or bicycles.

[0077] In some embodiments, according to the fourth constraint, on the one hand, when the relative driving direction and relative position of obstacle 120 relative to autonomous vehicle 110 meet preset conditions, the positions of autonomous vehicle 110 and obstacle 120 after a period of time can be determined, that is, it can be determined that a collision may occur between autonomous vehicle 110 and obstacle 120, thereby autonomous vehicle 110 generating a collision signal associated with obstacle 120. On the other hand, collision instances that move parallel to autonomous vehicle 110 and are located beside autonomous vehicle 110 are filtered to reduce false detections in close parallel scenarios of autonomous vehicle 110 and obstacle 120.

[0078] In some embodiments, the airbag sensors integrated into the chassis of the autonomous vehicle 110 will emit a collision response signal in the event of a severe collision. The collision algorithm of the autonomous vehicle 110 will access this signal and forward it through a dedicated field, thereby providing reliable collision detection data, especially in high-intensity collisions, providing crucial support for the system's emergency response.

[0079] Example devices and equipment

[0080] Embodiments of this disclosure also provide corresponding apparatus for implementing the above methods or processes. Figure 4 A schematic structural block diagram of a collision detection apparatus 400 according to certain embodiments of the present disclosure is shown. The apparatus 400 may be implemented as or included in the computing device 130 discussed above. The various modules / components in the apparatus 400 may be implemented by hardware, software, firmware, or any combination thereof.

[0081] like Figure 4 As shown, the device 400 includes an information acquisition module 410 configured to acquire the position information of an obstacle, the position information indicating the circumscribed rectangular region and the polygonal outline region of the obstacle; a first determination module 420 configured to determine first overlap information between the circumscribed rectangular region and the target region of the autonomous vehicle, and second overlap information between the polygonal outline region and the target region; a second determination module 430 configured to determine a first weight corresponding to the first overlap information and a second weight corresponding to the second overlap information based on the acceleration information of the autonomous vehicle, the acceleration information indicating the target acceleration independent of the control signal of the autonomous vehicle; a third determination module 440 configured to determine weighted information of the first overlap information and the second overlap information based on the first weight and the second weight; and a fourth determination module 450 configured to determine the target confidence level of a collision between the autonomous vehicle and the obstacle based on the weighted information.

[0082] In some embodiments, the information acquisition module 410 is configured to determine the circumscribed rectangular region of the obstacle based on the obstacle's tracking information; and / or to determine the polygonal outline region of the obstacle based on point cloud data associated with the obstacle.

[0083] In some embodiments, the first overlap information or the second overlap information indicates at least one of the following:

[0084] The distance from the circumscribed rectangular or polygonal outline region to the target region;

[0085] The overlap ratio between the circumscribed rectangular or polygonal outline region and the target region.

[0086] In some embodiments, the second determining module 430 is configured to acquire a first acceleration collected by the inertial navigation device of the autonomous vehicle;

[0087] Based on the installation location of the inertial navigation device, the first acceleration is converted into a second acceleration corresponding to the center of mass of the autonomous vehicle; and

[0088] Based on the second acceleration, the acceleration information of the autonomous vehicle is determined.

[0089] In some embodiments, determining the acceleration information of the autonomous vehicle based on the second acceleration includes:

[0090] The third acceleration is determined based on the target control signals of the autonomous vehicle; and

[0091] The acceleration information of the autonomous vehicle is determined based on the second and third accelerations.

[0092] In some embodiments, obtaining the first acceleration collected by the inertial navigation device of the autonomous vehicle includes:

[0093] Low-pass filtering is performed on the acceleration data collected by the inertial navigation device to determine the first acceleration.

[0094] In some embodiments, determining the third acceleration based on the target control signal of the autonomous vehicle includes:

[0095] Determine the response latency corresponding to autonomous vehicles; and

[0096] Based on response delay, the target control signal is determined from the historical control signals of the autonomous vehicle.

[0097] In some embodiments, the response delay is determined based on a comparison between a set of expected accelerations corresponding to a set of historical control signals and a set of historical accelerations.

[0098] In some embodiments, determining the target confidence level of a collision between an autonomous vehicle and an obstacle based on weighted information includes:

[0099] Determine a set of historical confidence levels for an autonomous vehicle colliding with an obstacle at a set of historical moments;

[0100] Based on a set of historical confidence levels, determine the prediction confidence level corresponding to the current moment;

[0101] Based on weighted information, determine the observation information for the decision-making model; and

[0102] Based on observational information, the prediction confidence level is adjusted to determine the target confidence level.

[0103] In some embodiments, a first weight is positively correlated with the target acceleration indicated by the acceleration information, and a second weight is negatively correlated with the target acceleration indicated by the acceleration information.

[0104] In some embodiments, the fourth determining module 450 is configured to determine whether collision information associated with an obstacle satisfies a set of preset constraints in response to a target confidence level greater than a threshold; and

[0105] In response to collision information satisfying a set of preset constraints, a collision signal associated with the obstacle is generated.

[0106] In some embodiments, a set of preset constraints includes at least one of the following:

[0107] The first constraint indicates that the overlap between the polygonal contour region and the target region is greater than a first threshold.

[0108] The second constraint indicates that the shape information of the obstacle meets the preset conditions.

[0109] The third constraint indicates that when the speed of the autonomous vehicle is less than the threshold, the target acceleration of the autonomous vehicle is greater than the second threshold.

[0110] The fourth constraint indicates that the relative driving direction and relative position of the obstacle with respect to the autonomous vehicle meet preset conditions.

[0111] In some embodiments, the circumscribed rectangular region and the polygonal outline region correspond to the overhead view.

[0112] The units included in device 400 can be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some embodiments, one or more units may be implemented using software and / or firmware, such as machine-executable instructions stored on a storage medium. In addition to or as an alternative to machine-executable instructions, some or all of the units in device 400 may be implemented at least partially by one or more hardware logic components. By way of example and not limitation, exemplary types of hardware logic components that may be used include field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-chips (SoCs), complex programmable logic devices (CPLDs), and so on.

[0113] Figure 5 A block diagram of an electronic device 500 in which one or more embodiments of the present disclosure may be implemented is shown. It should be understood that... Figure 5 The electronic device 500 shown is merely exemplary and should not be construed as limiting the functionality and scope of the embodiments described herein. Figure 5 The electronic device 500 shown can be used to achieve Figure 1 The computing device 130 shown.

[0114] like Figure 5 As shown, electronic device 500 is in the form of a general-purpose electronic device. Components of electronic device 500 may include, but are not limited to, one or more processors or processing units 510, memory 520, storage device 530, one or more communication units 540, one or more input devices 550, and one or more output devices 560. Processing unit 510 may be a physical or virtual processor and is capable of performing various processes according to programs stored in memory 520. In a multiprocessor system, multiple processing units execute computer-executable instructions in parallel to improve the parallel processing capability of electronic device 500.

[0115] Electronic device 500 typically includes multiple computer storage media. Such media can be any accessible media that is accessible to electronic device 500, including but not limited to volatile and non-volatile media, removable and non-removable media. Memory 520 can be volatile memory (e.g., registers, cache, random access memory (RAM)), non-volatile memory (e.g., read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. Storage device 530 can be a removable or non-removable medium and can include machine-readable media, such as flash drives, disks, or any other media that can be used to store information and / or data (e.g., training data for training) and can be accessed within electronic device 500.

[0116] Electronic device 500 may further include additional removable / non-removable, volatile / non-volatile storage media. Although not explicitly stated... Figure 5 As shown, disk drives for reading from or writing to removable, non-volatile disks (e.g., "floppy disks") and optical disk drives for reading from or writing to removable, non-volatile optical disks can be provided. In these cases, each drive can be connected to a bus (not shown) via one or more data media interfaces. Memory 520 may include computer program product 525 having one or more program modules configured to perform various methods or actions of various embodiments of this disclosure.

[0117] Communication unit 540 enables communication with other electronic devices via a communication medium. Additionally, the functionality of components of electronic device 500 can be implemented using a single computing cluster or multiple computing machines capable of communicating via communication connections. Therefore, electronic device 500 can operate in a networked environment using logical connections to one or more other servers, network personal computers (PCs), or another network node.

[0118] Input device 550 can be one or more input devices, such as a mouse, keyboard, trackball, etc. Output device 560 can be one or more output devices, such as a monitor, speaker, printer, etc. Electronic device 500 can also communicate with one or more external devices (not shown) via communication unit 540 as needed. These external devices include storage devices, display devices, etc., and can communicate with one or more devices that enable user interaction with electronic device 500, or with any device that enables electronic device 500 to communicate with one or more other electronic devices (e.g., network card, modem, etc.). Such communication can be performed via input / output (I / O) interface (not shown).

[0119] According to an exemplary implementation of this disclosure, a computer-readable storage medium is provided that stores computer-executable instructions thereon, wherein the computer-executable instructions are executed by a processor to implement the methods described above. According to an exemplary implementation of this disclosure, a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, which are executed by a processor to implement the methods described above.

[0120] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, devices, and computer program products implemented according to this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0121] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0122] Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions that execute on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0123] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0124] Various implementations of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed implementations. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described implementations. The terminology used herein is chosen to best explain the principles, practical applications, or improvements to technology in the market, or to enable others skilled in the art to understand the various implementations disclosed herein.

Claims

1. A collision detection method, comprising: Obtain the location information of the obstacle, wherein the location information indicates the bounding rectangular region and polygonal outline region of the obstacle; Determine the first overlap information between the circumscribed rectangular region and the target region of the autonomous vehicle, and the second overlap information between the polygonal contour region and the target region; Based on the acceleration information of the autonomous vehicle, a first weight corresponding to the first overlapping information and a second weight corresponding to the second overlapping information are determined, wherein the acceleration information indicates a target acceleration independent of the control signal of the autonomous vehicle; as well as Based on the first weight and the second weight, the weighted information of the first overlap information and the second overlap information is determined; as well as Based on the weighted information, the target confidence level of the collision between the autonomous vehicle and the obstacle is determined.

2. The method according to claim 1, further comprising: Based on the tracking information of the obstacle, the circumscribed rectangular region of the obstacle is determined; and / or Based on point cloud data associated with the obstacle, the polygonal outline region of the obstacle is determined.

3. The method of claim 1, wherein the first overlap information or the second overlap information indicates at least one of the following: The distance from the circumscribed rectangular region or the polygonal outline region to the target region; The overlap ratio between the circumscribed rectangular region or the polygonal outline region and the target region.

4. The method according to claim 1, further comprising: Acquire the first acceleration collected by the inertial navigation device of the autonomous vehicle; Based on the installation location of the inertial navigation device, the first acceleration is converted into a second acceleration corresponding to the center of mass of the autonomous vehicle; as well as Based on the second acceleration, the acceleration information of the autonomous vehicle is determined.

5. The method according to claim 4, wherein determining the acceleration information of the autonomous vehicle based on the second acceleration includes: Based on the target control signal of the autonomous vehicle, the third acceleration is determined; as well as The acceleration information of the autonomous vehicle is determined based on the second acceleration and the third acceleration.

6. The method of claim 4, wherein acquiring the first acceleration collected by the inertial navigation device of the autonomous vehicle comprises: The acceleration data collected by the inertial navigation device is subjected to low-pass filtering to determine the first acceleration.

7. The method of claim 4, wherein determining the third acceleration based on the target control signal of the autonomous vehicle comprises: Determine the response delay corresponding to the autonomous vehicle; as well as Based on the response delay, the target control signal is determined from the historical control signals of the autonomous vehicle.

8. The method of claim 7, wherein the response delay is determined based on a comparison between a set of expected accelerations corresponding to a set of historical control signals and a set of historical accelerations.

9. The method according to claim 1, wherein determining the target confidence level of the collision between the autonomous vehicle and the obstacle based on the weighted information includes: Determine a set of historical confidence levels for the collision between the autonomous vehicle and the obstacle at a set of historical moments; Based on the set of historical confidence levels, determine the prediction confidence level corresponding to the current moment; Based on the weighted information, the observation information for the decision-making model is determined; as well as Based on the observation information, the prediction confidence is adjusted to determine the target confidence.

10. The method of claim 1, wherein the first weight is positively correlated with the target acceleration indicated by the acceleration information, and the second weight is negatively correlated with the target acceleration indicated by the acceleration information.

11. The method according to claim 1, further comprising: In response to the target confidence being greater than a threshold, it is determined whether the collision information associated with the obstacle satisfies a set of preset constraints; as well as In response to the collision information satisfying the set of preset constraints, a collision signal associated with the obstacle is generated.

12. The method of claim 11, wherein the set of preset constraints includes at least one of the following: A first constraint indicates that the overlap between the polygonal contour region and the target region is greater than a first threshold. The second constraint indicates that the shape information of the obstacle meets a preset condition; A third constraint indicates that, when the speed of the autonomous vehicle is less than a threshold, the target acceleration of the autonomous vehicle is greater than a second threshold. The fourth constraint indicates that the relative driving direction and relative position of the obstacle with respect to the autonomous vehicle meet preset conditions.

13. The method of claim 1, wherein the circumscribed rectangular region and the polygonal outline region correspond to a top-down view.

14. An apparatus for collision detection, comprising: The information acquisition module is configured to acquire the location information of an obstacle, wherein the location information indicates the bounding rectangular region and the polygonal outline region of the obstacle. The first determining module is configured to determine first overlap information between the circumscribed rectangular region and the target region of the autonomous vehicle, and second overlap information between the polygonal outline region and the target region. The second determining module is configured to determine a first weight corresponding to the first overlapping information and a second weight corresponding to the second overlapping information based on the acceleration information of the autonomous vehicle, wherein the acceleration information indicates a target acceleration independent of the control signal of the autonomous vehicle; as well as The third determining module is configured to determine the weighted information of the first overlap information and the second overlap information based on the first weight and the second weight; as well as The fourth determining module is configured to determine the target confidence level of the collision between the autonomous vehicle and the obstacle based on the weighted information.

15. An electronic device comprising: At least one processing unit; as well as At least one memory, coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, which, when executed by the at least one processing unit, cause the electronic device to perform the method according to any one of claims 1 to 13.

16. A computer-readable storage medium having a computer program stored thereon, the computer program being executable by a processor to implement the method according to any one of claims 1 to 13.

Images