A vehicle-in-the-loop test method and test system

Abstract

The application discloses a kind of whole vehicle in loop test method, test system.For the problem that target is lost due to sensing distortion of physical simulation equipment under extremely close distance, the method comprises: controlling physical simulation equipment to output physical excitation signal in far zone, so that domain controller identifies target and generates unique target tracking ID;When target distance or collision time is reduced to below the preset physical performance boundary threshold, smooth relay transition is triggered;In transition, physical excitation signal is gradually attenuated, and sensing result message is injected into vehicle bus synchronously, wherein the target ID of injected message is set as the aforementioned target tracking ID, and data confidence is inversely improved with signal attenuation;After physical signal is zeroed, only message is continuously injected into bus with full confidence.This application effectively maintains target without loss under extremely close distance through seamless handover of virtual and real signals, and realizes high-fidelity verification of vehicle control response under limit safety scenario.

Description

Technical Field

[0001] This invention relates to the field of vehicle testing technology, specifically to a whole vehicle in-loop testing method and system. Background Technology

[0002] With the rapid development of autonomous driving technology, Vehicle-in-the-Loop (VIL) testing has become an indispensable core component for verifying the control strategies of Advanced Driver Assistance Systems (ADAS) and autonomous driving domain controllers. To maximize the replication of the real physical world, existing VIL testing generally employs physical excitation methods to provide a testing environment for real vehicle sensors. For example, a high-resolution display screen is placed in front of the test vehicle for onboard cameras to capture environmental images, while a radar simulator transmits echo signals with specific distance delays and Doppler frequency shifts to the onboard millimeter-wave radar. This testing method based on real physical perception preserves the underlying characteristics of the sensor hardware and offers high test fidelity.

[0003] As autonomous driving active safety testing standards become increasingly stringent, the demand for automatic emergency braking (AEB) and emergency avoidance testing in extreme scenarios such as extremely close-range emergency lane-cutting and pedestrians suddenly appearing (e.g., target distance less than 3 meters) is constantly increasing. When facing such extremely close-range scenarios, due to limitations in optical imaging principles, when onboard binocular or multi-view cameras capture images of a two-dimensional display at extremely close range, they cannot obtain true depth parallax information, resulting in severe parallax distortion. This can easily lead to misjudgments or failures in visual perception algorithms. Furthermore, due to limitations in microwave near-field effects and the physical isolation of antenna transmission and reception, radar simulators are prone to multipath interference and have physical simulation blind spots when simulating dynamic targets at extremely close range, making it impossible to output high signal-to-noise ratio radar echoes at extremely short ranges.

[0004] The aforementioned limitations of physical simulation equipment under near-field conditions can lead to severe signal distortion in the vehicle's real sensors as the test target approaches the test vehicle from a distance. Once the target crosses the physical performance boundary threshold of the equipment, this sudden degradation in perception can cause anomalies in the target tracking algorithm (such as Kalman filtering) within the domain controller, resulting in target information jumps, target tracking ID resets, or even instantaneous target loss. Because the domain controller cannot maintain continuous and stable target perception at extremely close ranges, the test vehicle often fails to trigger the expected control response under extreme conditions such as automatic emergency braking. This forces the interruption of vehicle-in-the-loop testing in such high-risk close-range scenarios, preventing the formation of a complete test loop.

[0005] Therefore, overcoming the limitations of physical simulation equipment in perceptual distortion at extremely close distances and ensuring that the target ID is not lost during the entire motion cycle from far to near, thereby achieving high-fidelity verification of vehicle control response in extreme scenarios at extremely close distances without compromising the integrity of the vehicle system, is a technical challenge that urgently needs to be solved in the field of vehicle-in-the-loop testing. Summary of the Invention

[0006] In view of the above-mentioned shortcomings mentioned in the background art, the purpose of this invention is to provide a vehicle-in-the-loop testing method and system.

[0007] This invention provides a method for in-loop testing of a whole vehicle, the method comprising the following steps:

[0008] In a virtual simulation scenario, test targets are generated, and the physical simulation equipment is controlled to continuously output physical excitation signals corresponding to the test targets to the real sensors of the test vehicle. This enables the domain controller of the test vehicle to identify the test targets based on real physical perception and generate a unique target tracking ID for the test targets.

[0009] The relative distance or collision time between the test target and the test vehicle in the virtual simulation scene is acquired in real time. When the relative distance or collision time is reduced to below the preset physical performance boundary threshold, a smooth relay transition is triggered.

[0010] During the smooth relay transition, the strength of the physical excitation signal and the visibility of the image output by the physical simulation device gradually decrease, and the perception result message of the test target is injected into the domain controller through the vehicle bus at the same time; wherein, the target ID in the perception result message is set as the unique target tracking ID, and the data confidence of the perception result message increases inversely as the physical excitation signal decreases.

[0011] When the physical excitation signal strength decays to zero, the signal output of the physical simulation device is completely stopped. Only the perception result message of the test target is continuously injected into the domain controller with full confidence through the vehicle bus to keep the test target from being lost, thereby verifying the control response of the test vehicle in a very close-range scenario.

[0012] The present invention also provides a vehicle-in-the-loop testing system, the system comprising:

[0013] The physical stimulus and recognition module is used to generate test targets in a virtual simulation scenario, control the physical simulation equipment to continuously output physical stimulus signals corresponding to the test targets to the real sensors of the test vehicle, so that the domain controller of the test vehicle can recognize the test targets based on real physical perception and generate a unique target tracking ID for the test targets;

[0014] The status monitoring and triggering module is used to obtain the relative distance or collision time between the test target and the test vehicle in the virtual simulation scene in real time. When the relative distance or collision time is reduced to below the preset physical performance boundary threshold, the smooth relay transition is triggered.

[0015] The smooth relay transition module is used to control the physical excitation signal strength and screen visibility output by the physical simulation device to gradually decrease during the smooth relay transition process, and simultaneously inject the perception result message of the test target into the domain controller through the vehicle bus; wherein, the target ID in the perception result message is set as the unique target tracking ID, and the data confidence of the perception result message increases inversely as the physical excitation signal decreases.

[0016] The near-field full injection verification module is used to completely stop the signal output of the physical simulation device when the physical excitation signal strength decays to zero. It continuously injects the perception result message of the test target into the domain controller with full confidence only through the vehicle bus to keep the test target from being lost, thereby verifying the control response of the test vehicle in extremely close-range scenarios.

[0017] This invention overcomes the parallax distortion and near-field physical blind spots of existing physical simulation equipment at extremely close distances by employing realistic physical excitation in the far-field region and triggering a smooth relay transition mechanism when the test target approaches the physical performance boundary threshold. During the transition, this invention achieves seamless and smooth handover of the target perception data source by smoothly attenuating the physical excitation signal and synchronously injecting bus data, and by forcing the injected perception message to inherit the target tracking ID generated by the original physical perception, thus inversely improving the data confidence. The solution of this invention can effectively avoid target jumps and loss caused by sensor perception degradation at extremely close distances, ensuring continuous and stable tracking of the same test target by the domain controller. This solves the near-field testing bottleneck and achieves high-fidelity verification of vehicle control response in extremely close-range extreme safety scenarios without compromising the integrity of the entire vehicle system. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of a vehicle-in-the-loop testing method disclosed in an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of the overall process of injecting the perception result message of the test target into the domain controller through the vehicle bus, as disclosed in an embodiment of the present invention.

[0020] Figure 3 This is a partial flowchart illustrating the process of injecting a perception result message of a test target into a domain controller via a vehicle bus, as disclosed in an embodiment of the present invention.

[0021] Figure 4This is a schematic diagram of the structure of a vehicle-in-the-loop testing system disclosed in an embodiment of the present invention;

[0022] Figure 5 This is a schematic diagram of the unfolded structure of a vehicle-in-the-loop testing system disclosed in an embodiment of the present invention. Detailed Implementation

[0023] The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this disclosure.

[0024] Reference Figure 1 As shown, this embodiment of the invention provides a vehicle-in-the-loop testing method, which includes the following steps S1 to S4:

[0025] S1: Generate test targets in a virtual simulation scenario, control the physical simulation equipment to continuously output physical excitation signals corresponding to the test targets to the real sensors of the test vehicle, so that the domain controller of the test vehicle can identify the test targets based on real physical perception and generate a unique target tracking ID for the test targets;

[0026] In this embodiment, the method can be deployed in the main control computer, industrial control computer, embedded processor, or cloud server of a vehicle-in-the-loop test bench. The test bench can be equipped with a vehicle-in-the-loop test management system, physical simulation equipment, and a gateway injection device connected to the test vehicle's bus; the physical simulation equipment may include a front-facing high-resolution display and a radar simulator; the test vehicle may be equipped with a real binocular camera, a forward-facing millimeter-wave radar, and an intelligent driving domain controller.

[0027] In practice, the vehicle-in-the-loop testing system first generates a test target in a virtual simulation scenario. This test target can be a vehicle ahead, a pedestrian, a two-wheeled vehicle, or other road users. Taking an extremely close-range emergency braking test scenario as an example, the virtual simulation scenario can generate a test target located 50 meters in front of the test vehicle with a relative speed of -20 km / h, representing the vehicle decelerating and approaching. This test target has a corresponding three-dimensional contour, motion state, and relative pose, and can be updated in real time by the virtual simulation scenario.

[0028] Then, the test system controls the physical simulation equipment to continuously output physical excitation signals corresponding to the test target to the real sensors of the test vehicle. Specifically, the high-resolution display screen in front of the physical simulation equipment renders the image of the test target for the real camera of the test vehicle to capture; the radar simulator emits radar waves with corresponding range delay and Doppler frequency shift for the real millimeter-wave radar of the test vehicle to receive. Thus, the real sensors of the test vehicle do not receive virtual perception results directly injected by the vehicle bus, but rather physical excitation signals provided by the physical simulation equipment. Continuous output means that during this far-field phase, the physical simulation equipment continuously provides visual and radar excitations to the real sensors according to the motion state of the test target in the virtual simulation scenario, until the triggering conditions for a smooth transition are met.

[0029] The test vehicle's domain controller identifies the test target based on real physical perception. Specifically, a real binocular camera captures the target image on the display screen, a forward-facing millimeter-wave radar receives the echo signal output from the radar simulator, and the domain controller's internal sensor fusion algorithm fuses the visual and radar information to identify the test target. After identifying the test target, the domain controller assigns a unique tracking ID to it, such as ID_001, and begins stable tracking. It should be understood that this unique target tracking ID is generated by the domain controller based on real physical perception, not pre-specified by the test system, nor created by vehicle bus injection messages.

[0030] It should be noted that when the test target is far from the test vehicle and the physical simulation equipment is still within its effective performance range, the physical simulation equipment should provide physical stimuli first, enabling the domain controller to complete target identification and generate a unique target tracking ID based on the real-world perception link. This unique target tracking ID is the key foundation for achieving target identity inheritance and preventing target loss during the subsequent smooth handover. If bus injection is performed directly in the far-field stage, the domain controller cannot establish a target tracking ID based on real-world physical perception, and the subsequent handover of virtual and real signals will lack an inheritance target.

[0031] It is understood that the specific values ​​and identifiers such as 50 meters, -20 km / h, and ID_001 mentioned above are only used to illustrate the technical approach of this embodiment and are not intended to limit the scope of protection of this invention. Without departing from the core logic of virtual-real data transfer in this invention, adaptive adjustments can be made according to specific test scenarios, sensor hardware characteristics, and test bench environments.

[0032] S2: Real-time acquisition of the relative distance or collision time between the test target and the test vehicle in the virtual simulation scene. When the relative distance or collision time is reduced to below the preset physical performance boundary threshold, a smooth relay transition is triggered.

[0033] In this embodiment, the testing system acquires the relative distance or collision time between the test target and the test vehicle in the virtual simulation scene in real time. Specifically, the pose and motion state of the test target in the virtual simulation scene are updated in real time, and the testing system can calculate the relative distance between the test target and the test vehicle based on their real-time poses.

[0034] The relative distance can be represented as the distance between the geometric center of the test target and the reference point of the test vehicle, or as the closest distance between the bounding box of the test target's 3D contour and the actual sensor of the test vehicle. The collision time can be calculated from the relative distance and relative velocity; for example, dividing the relative distance by the absolute value of the relative velocity yields the collision time under the current motion trend. The testing system can periodically acquire the aforementioned relative distance or collision time, for example, performing real-time monitoring with periods of 10ms, 20ms, or 50ms.

[0035] When the relative distance or collision time decreases below a preset physical performance boundary threshold, the test system triggers a smooth relay transition. It's important to note that a smooth relay transition refers to the process of transitioning from a phase primarily driven by physical excitation signals provided by the physical simulation equipment to a phase primarily driven by sensing result messages injected into the vehicle bus. It should be understood that if the smooth relay transition is triggered too early, the physical simulation equipment can still provide high-fidelity physical excitation at a considerable distance, and premature injection of bus messages will reduce the realistic physical perception component of the test. If the smooth relay transition is triggered too late, the real sensors of the test vehicle have already entered the near-field distortion zone of the physical simulation equipment, and the domain controller may have already experienced target jumps, target tracking ID resets, or even target loss, causing subsequent transitions to lose their inheriting object. Therefore, this step precisely triggers a smooth relay transition by real-time monitoring of the relative distance or collision time, before the physical perception is about to deteriorate but has not yet deteriorated.

[0036] As an example, the preset physical performance boundary threshold is taken from the maximum of the minimum effective microwave simulation distance of the radar simulator in the physical simulation device and the critical distance at which the camera of the test vehicle produces parallax distortion when shooting the display screen of the physical simulation device.

[0037] In practice, physical simulation equipment exhibits physical performance limitations at extremely close ranges. On one hand, radar simulators are physically constrained by near-field microwave effects, antenna transmission / reception isolation, and multipath interference, resulting in a minimum effective microwave simulation distance. When the distance between the test target and the test vehicle is less than this minimum effective microwave simulation distance, the radar simulator cannot output high signal-to-noise ratio radar echoes at extremely short ranges, and may even generate multipath interference, causing the test vehicle's forward-facing millimeter-wave radar to be unable to stably identify the target. This minimum effective microwave simulation distance can be obtained through calibration experiments, for example, calibrated to 2.5 meters. On the other hand, when the test vehicle's binocular or multi-view cameras capture images of the physical simulation equipment's flat display screen at extremely close range, the lack of realistic depth parallax information leads to severe parallax distortion, causing misjudgments or failures in the visual perception algorithm. The critical distance for this parallax distortion can also be obtained through calibration experiments, for example, calibrated to 3.0 meters.

[0038] Therefore, the preset physical performance boundary threshold is taken from the maximum value of the minimum effective microwave simulation distance and the parallax distortion critical distance. Taking the example above, the minimum effective microwave simulation distance is 2.5 meters, and the parallax distortion critical distance is 3.0 meters. The maximum value is 3.0 meters, so the test system uses 3.0 meters as the preset physical performance boundary threshold. By taking the maximum value, a smooth transition can be triggered in time before either the radar simulator or the camera reaches the physical performance boundary. In other words, as long as the camera reaches the parallax distortion critical distance before the radar, or the radar reaches the minimum effective microwave simulation distance before the camera, the test system will initiate the transition at the first reached physical performance boundary, thereby preventing any real sensor from experiencing perception degradation before the transition begins.

[0039] It should be noted that the specific values ​​of 2.5 meters, 3.0 meters, etc., mentioned above are only used to illustrate the technical approach of this embodiment and are not intended to limit the scope of protection of this invention. The minimum effective microwave simulation distance and parallax distortion critical distance can be adaptively calibrated and adjusted according to different radar simulator models, different camera parameters, different display screen arrangements, and different test bench layouts. For example, the parallax distortion critical distance may be different for binocular cameras with different focal lengths; the minimum effective microwave simulation distance may also be different for radar simulators with different frequency bands and antenna layouts. The test system can pre-store the above calibration results and select the corresponding physical performance boundary thresholds according to the actual test configuration during the test.

[0040] In practice, the testing system can simultaneously monitor relative distance and collision time, and set distance thresholds and collision time thresholds. When the relative distance is less than 3.0 meters, or the collision time is less than the collision time threshold corresponding to 3.0 meters, the testing system determines that the triggering conditions are met and triggers the smooth relay transition. After triggering, the testing system enters step S3, begins to control the physical excitation signal strength and image visibility output by the physical simulation equipment to gradually decay, and simultaneously injects the perception result message of the test target into the domain controller through the vehicle bus.

[0041] Understandably, by setting the physical performance boundary threshold to the maximum of the minimum effective microwave simulation range of the radar simulator and the critical distance for camera parallax distortion, the test system can initiate the handover between virtual and real signals before the physical simulation equipment can provide effective excitation but is about to enter the near-field distortion zone. This prevents the domain controller from losing targets when physical perception deteriorates. This triggering mechanism, in conjunction with the aforementioned generated unique target tracking ID, enables the perception result messages injected in subsequent steps to inherit the same target identity, achieving a smooth and seamless handover transition.

[0042] S3: During the smooth relay transition, the strength of the physical excitation signal and the visibility of the screen output by the physical simulation device are gradually reduced, and the perception result message of the test target is injected into the domain controller through the vehicle bus at the same time; wherein, the target ID in the perception result message is set as the unique target tracking ID, and the data confidence of the perception result message increases inversely as the physical excitation signal decreases.

[0043] In this embodiment, the test system simultaneously performs the following operations within a very short time window, such as 0.5 seconds: 1) gradually attenuating the strength of the physical excitation signal output by the physical simulation device and the visibility of the image; 2) injecting the perception result message of the test target into the domain controller through the vehicle bus. The above two operations are carried out simultaneously, enabling seamless and smooth handover of the target information received by the domain controller at the data source.

[0044] In practice, the intensity of the physical excitation signal can gradually decrease from its initial strength to zero. For example, the echo intensity output by the radar simulator can decrease smoothly according to a preset curve; simultaneously, the transparency of the test target image on the high-resolution display screen gradually increases, causing the target image to gradually fade until it becomes completely transparent. It should be understood that the decrease in image visibility corresponds to the increase in transparency, in order to avoid a sudden disappearance of the target image that would cause abrupt changes in visual perception.

[0045] Simultaneously with the attenuation of the physical stimulus signal strength, the gateway injection device begins injecting perception result messages of the test target into the vehicle bus. The data confidence level in the perception result messages increases inversely as the physical stimulus signal attenuates. For example, when the physical stimulus signal strength is 100% of its initial strength, the data confidence level of the injected message can be 0%; when the physical stimulus signal strength attenuates to 80%, the data confidence level increases to 20%; and when the physical stimulus signal strength attenuates to zero, the data confidence level increases to 100%. Through this inverse increase, the domain controller gradually obtains bus-injected perception results with increasingly higher confidence levels as the physical perception gradually weakens, thus preventing target jumps due to sudden changes in the data source.

[0046] It should be noted that the target ID in the perception result message is forcibly set to the aforementioned unique target tracking ID, such as ID_001. In other words, the injection message does not create a new virtual target, but rather inherits the target identity that the domain controller has already generated and stably tracked in step S1. If the injection message uses a new target ID, the target tracking algorithm inside the domain controller may identify the injected target as a new target, leading to the cancellation of the original target ID, target switching, or target loss, thereby interrupting the very close-range AEB test. Therefore, target ID inheritance is the key to achieving seamless handover between virtual and real signals.

[0047] As an example, the gradual attenuation of the physical excitation signal strength and the reverse increase of the data confidence of the perception result message follow a linear mapping relationship or an S-shaped curve mapping relationship; and during the gradual attenuation of the screen visibility, the transparency of the test target screen on the display screen in the physical simulation device gradually increases with time or the shortening of the relative distance until it is completely transparent.

[0048] In practice, the gradual attenuation of the physical excitation signal strength and the reverse increase in the data confidence level of the sensing result message can follow a linear mapping relationship. For example, suppose the physical excitation signal strength is... The data confidence level is Then the linear mapping relationship can be expressed as:

[0049]

[0050] in, This is the initial value of the physical excitation signal strength.

[0051] As another example, the two can also follow an S-shaped curve mapping relationship, for example:

[0052]

[0053] in, This is the curve steepness coefficient. This represents the midpoint of the transition. The S-shaped curve changes relatively gently at the beginning and end of the transition, but changes more rapidly in the middle stage. This allows the data source handover to be smoother at the beginning and end, preventing the domain controller from sensing sudden changes at the beginning or end of the transition.

[0054] As the visibility of the image gradually decreases, the transparency of the test target image on the display screen in the physical simulation device gradually increases with time or a decrease in relative distance, until it becomes completely transparent. In practice, the target vehicle image on the display screen does not disappear instantly; instead, its transparency gradually increases following an S-shaped or linear curve, gradually fading until it becomes completely transparent. This increase in transparency corresponds to the decrease in image visibility, preventing the target image captured by the camera from suddenly disappearing, thus avoiding target loss or misjudgment by the visual perception algorithm due to sudden changes in the image.

[0055] As an example, see Figure 2 The perception result message of the test target is injected into the domain controller via the vehicle bus, specifically including two sub-steps, S31 and S32:

[0056] S31: Obtain the size of the three-dimensional contour bounding box and the coordinates of the geometric center of the test target in the virtual simulation scene. Based on the real-time relative pose of the test vehicle and the test target, determine the physical contour point closest to the test vehicle on the edge surface of the three-dimensional contour bounding box and define it as the equivalent physical detection anchor point.

[0057] The testing system acquires the dimensions and geometric center coordinates of the 3D bounding box of the test target in the virtual simulation scene. The 3D bounding box describes the outer contour of the test target in space, and its geometric center coordinates are a commonly used target position representation in simulation software. In specific implementation, the 3D bounding box can be represented as:

[0058]

[0059] in, The coordinates of the geometric center of the 3D bounding box are: , , These represent the length, width, and height of the 3D bounding box. , , These are the three unit direction vectors in the target's body coordinate system. The testing system can directly read these dimensions and geometric center coordinates from the virtual simulation scene, or it can calculate them in real time based on the target's 3D model.

[0060] Furthermore, based on the real-time relative pose of the test vehicle and the test target, the testing system determines the physical contour point closest to the test vehicle on the edge surface of the 3D contour bounding box. Specifically, the real-time relative pose of the test vehicle and the test target can be represented as a rotation matrix of the test target's body coordinate system relative to the test vehicle's coordinate system. Translation vector For any point on the 3D bounding box... Its coordinates in the test vehicle coordinate system can be expressed as:

[0061]

[0062] The testing system searches the edge surface of the 3D bounding box for the physical contour point closest to the test vehicle reference point. The test vehicle reference point can be the actual sensor mounting location of the test vehicle, such as the camera or radar mounting location, or the origin of the test vehicle's coordinate system. Through this search, the testing system obtains the physical contour point located on the edge surface of the 3D bounding box and defines it as the equivalent physical probe anchor point. It should be noted that the equivalent physical probe anchor point is not the geometric center of the test target, but rather a surface point simulating the nearest detection point of the actual sensor, thus better aligning with the domain controller's tracking logic for physical targets.

[0063] It should be noted that if the geometric center coordinates of the 3D contour bounding box are directly injected into the bus as the target position coordinates, at extremely close distances, the domain controller may perceive the target as instantly jumping back by half the vehicle length, leading to control algorithm divergence or abnormal target tracking. Therefore, this step determines equivalent physical detection anchor points to ensure that the target position in the injected message spatially aligns with the target surface position detected by the actual physical sensors before the transition, avoiding spatial jumps caused by center coordinate injection.

[0064] As an example, refer to Figure 3 As shown, the physical contour point closest to the test vehicle is determined on the edge surface of the three-dimensional contour bounding box and defined as the equivalent physical detection anchor point, including:

[0065] S311: Obtain the effective physical field of view model of the real sensor on the test vehicle, and based on the real-time relative pose of the effective physical field of view model and the test target, calculate the spatial intersection part of the three-dimensional contour bounding box intercepted by the effective physical field of view model, and define the exposed surface of the spatial intersection part as the effective visible perception area.

[0066] In this step, the testing system acquires the effective physical field-of-view model of the real sensors on the test vehicle. This effective physical field-of-view model describes the effective detection range of the real sensors in space. Since test vehicles typically carry both cameras and millimeter-wave radar, the effective physical field-of-view models can be established separately or merged into a unified spatial constraint model.

[0067] As an example, the effective physical detection field of view model for a camera on a test vehicle can be represented as a frustum model. This frustum model can be determined by the following parameters: the camera's mounting position in the test vehicle coordinate system. Unit vector of camera optical axis direction Horizontal field of view Vertical field of view and maximum effective detection range In the camera coordinate system, the view frustum can be represented as a set of spatial points that satisfy the following conditions:

[0068]

[0069] in, The coordinates of the spatial point in the camera coordinate system.

[0070] For a millimeter-wave radar on a test vehicle, its effective physical detection field of view model can be represented as a detection beam model. This detection beam model can be determined by the following parameters: the radar's installation position in the test vehicle coordinate system. Unit vector in the direction of radar beam center Horizontal beam angle Vertical beam angle Minimum effective detection range and maximum effective detection range In the radar coordinate system, the detection beam can be represented as a set of spatial points satisfying the following conditions:

[0071]

[0072] All of the above parameters can be obtained in advance through calibration experiments and stored in the main control computer of the test system. It should be noted that the camera view frustum model and the radar detection beam model can be established separately and used separately for subsequent calculations; or, depending on the test requirements, the two can be merged into a unified effective physical detection field of view model, that is, only the spatial region that falls within the camera view frustum model or the radar detection beam model at the same time is considered to be the area that the real sensor can theoretically perceive.

[0073] Furthermore, based on the real-time relative pose of the effective physical field-of-view model and the test target, the testing system calculates the spatial intersection of the 3D contour bounding box intercepted by the effective physical field-of-view model. In specific implementation, the 3D contour bounding box, as defined above, has the following dimensions: , , The geometric center coordinates are The testing system transforms the 3D bounding box from the world coordinate system to the sensor coordinate system through coordinate transformation. Let the rotation matrix from the world coordinate system to the sensor coordinate system be... The translation vector is Then any point on the bounding box The coordinates in the sensor coordinate system can be expressed as:

[0074]

[0075] Wherein, rotation matrix Translation vector The relative pose between the test vehicle and the test target in real time, as well as the installation position of the sensors on the test vehicle, are determined.

[0076] After transforming the 3D bounding box to the sensor coordinate system, the test system calculates the spatial intersection of the bounding box and the effective physical detection field of view model. Specifically, for the camera frustum model, the test system can use a frustum clipping algorithm, such as the Sutherland-Hodgman polygon clipping algorithm, to clip each face of the 3D bounding box to the boundary plane of the frustum, retaining the area inside the frustum. For the radar detection beam model, the test system can similarly clip the 3D bounding box to the boundary plane of the detection beam, and further calculate based on the minimum effective detection range. and maximum effective detection range Distance constraints are applied to the clipping results. After the above clipping, the portion of the 3D contour bounding box that lies within the effective physical detection field of view model is the spatial intersection.

[0077] After obtaining the spatial intersection portion, the testing system further determines the exposed surface of this spatial intersection portion and defines it as the effective visible sensing area. In specific implementation, the spatial intersection portion consists of several surface patches of a three-dimensional contour bounding box. For each surface patch, the testing system determines the angle between its normal vector and the direction vector from the sensor to that surface patch.

[0078] Let the normal vector of a certain surface patch be... The direction vector from the sensor to any point on the surface plate is If satisfied This indicates that the surface patch faces the sensor and is an exposed surface; if it satisfies This indicates that the surface patch faces away from the sensor and is not an exposed surface.

[0079] The testing system combines all surface patches facing the sensor to obtain the exposed surface of the spatially intersecting portion, which is defined as the effective visible sensing area. This effective visible sensing area is the area of ​​the test target surface that a real sensor can theoretically see, and it will not fall within the blind spot or detection blind zone of the real sensor.

[0080] It should be noted that the effective visible sensing area obtained through this step is not the entire surface of the 3D contour bounding box of the test target, but rather a surface area that simultaneously satisfies the following two conditions: the surface area is located within the spatial range defined by the effective physical detection field of view model; and the surface area faces the actual sensors of the test vehicle. Therefore, the effective visible sensing area can accurately reflect the actual perceptible surface range of the test target by the camera or millimeter-wave radar under the current positional relationship.

[0081] S312: Within the effective visible sensing area, extract the geometric point closest to the test vehicle as the physical contour point, and define it as the equivalent physical detection anchor point.

[0082] In this step, the testing system extracts the geometric point closest to the test vehicle within the obtained effective visible sensing area as a physical contour point and defines it as an equivalent physical detection anchor point. In specific implementation, the effective visible sensing area consists of several surface patches. The testing system can calculate the geometric point closest to the test vehicle for each surface patch, and then select the closest point to the test vehicle from all the closest points of all surface patches as the final equivalent physical detection anchor point.

[0083] For any surface patch, the geometric point closest to the test vehicle is calculated as follows:

[0084] Let the plane containing the surface patch be... The test vehicle reference point or sensor installation location is The reference point for the test vehicle can be the camera mounting location, the radar mounting location, or the origin of the test vehicle's coordinate system. Projected onto a plane , obtain the projection point :

[0085]

[0086] in, For any point on the surface sheet, Let be the unit normal vector of the surface patch. If the projection point... If it falls within the boundary range of the surface sheet, then This is the geometric point on the surface piece closest to the test vehicle. If the projection point... If the point falls outside the boundary of the surface patch, then the geometric point closest to the test vehicle is determined on the boundary line segment of the surface patch. For example, for the boundary line segment... The nearest point on it can be represented as:

[0087]

[0088] in,

[0089]

[0090] This indicates that the value is limited to the range [0,1]. The test system can calculate... Find the nearest point to each boundary line segment and select the one with the smallest distance. The calculation of the nearest point can be achieved using point-plane projection and point-line projection methods in spatial geometry.

[0091] After obtaining the nearest point corresponding to each surface patch, the test system calculates the distance between each nearest point and the reference point of the test vehicle. The distance between them is calculated, and the point with the smallest distance is selected as the final physical contour point. Let the first point be... The nearest point corresponding to each surface patch is Then the equivalent physical detection anchor point can be expressed as:

[0092]

[0093] in, This is the equivalent physical detection anchor point. The testing system will... The coordinates are used as the target location coordinates and injected into the perception result message.

[0094] Using the above method, the testing system can extract the geometric point closest to the test vehicle within the effective visible perception area, serving as an equivalent physical detection anchor point. This equivalent physical detection anchor point is strictly limited to the area theoretically perceptible by a real sensor. If the geometric nearest point is only calculated on the entire 3D bounding box, when the test target is obstructed at a large angle and at extremely close range, the nearest point may fall outside the physical field of view of the test vehicle's camera. Upon receiving the target coordinates located within the blind spot, the domain controller may trigger a rationality safety check, determine the data as abnormal, and report a fault code. Therefore, by capturing the effective visible perception area through the field of view model and then extracting the nearest point within that area, the injected data can be made consistent with the rationality of the field of view of a real sensor, thereby avoiding anti-spoofing errors from the domain controller.

[0095] It should be noted that the equivalent physical detection anchor point is not the geometric center of the test target, but rather a surface point simulating the nearest detection point of the real sensor. Through the above processing, the target position in the injected message can be spatially aligned with the target surface position detected by the real physical sensor before the transition, avoiding spatial jumps caused by center coordinate injection, thereby enabling the domain controller to maintain continuous and stable tracking of the same test target.

[0096] As an example, when the relative movement of the test target causes a change in the effective visible sensing area, the geometric point closest to the test vehicle within the effective visible sensing area is extracted in real time, so that the equivalent physical detection anchor point is always updated by sliding within the dynamic boundary of the effective visible sensing area.

[0097] In practice, as the test target continuously approaches or makes a large-angle entry, the relative pose between the test vehicle and the test target changes constantly. Consequently, the spatial intersection portion captured by the effective physical detection field-of-view model also changes, meaning the effective visible perception area dynamically changes. At this point, the test system re-executes steps S311 and S312 in real time, recalculating the spatial intersection portion, re-determining the exposed surface, and re-extracting the geometric point closest to the test vehicle within the effective visible perception area. This ensures that the equivalent physical detection anchor point is constantly sliding and updating within the dynamic boundary of the effective visible perception area.

[0098] For example, in an emergency lane-changing scenario, the test target cuts into front of the test vehicle from an adjacent lane, and the area of ​​the test target's surface visible to the test vehicle's cameras and radar will change accordingly. The equivalent physical detection anchor point can continuously slide from one corner of the rear bumper to another corner or a visible point on the side. This slide update can reproduce the phenomenon of microwave reflection points sliding on the target vehicle's surface in the real physical world, ensuring that the target position received by the domain controller is continuous and reasonable, and avoiding target tracking anomalies triggered by fixed or abrupt anchor points.

[0099] S32: Use the coordinates of the equivalent physical detection anchor point as the target position coordinates and inject them into the sensing result message.

[0100] The perception result message includes fields such as target ID, target location coordinates, and data confidence level. The target ID is set to the aforementioned unique target tracking ID, such as ID_001; the target location coordinates use the coordinates of the equivalent physical detection anchor point, rather than the geometric center coordinates of the test target; the data confidence level increases inversely according to the attenuation of the physical excitation signal. This perception result message is injected into the domain controller via the vehicle bus, enabling the domain controller to obtain target information that is continuously located, consistent in identity, and gradually increases in confidence level as physical perception gradually weakens.

[0101] For example, the testing system can perform the following operations in each control cycle: acquire the dimensions and geometric center coordinates of the 3D bounding box of the test target; acquire the real-time relative pose of the test vehicle and the test target; acquire the effective physical detection field of view model of the real sensors of the test vehicle; calculate the spatial intersection of the 3D bounding box and the effective physical detection field of view model to obtain the effective visible sensing area; extract the geometric point closest to the test vehicle within the effective visible sensing area as the equivalent physical detection anchor point; use the coordinates of the equivalent physical detection anchor point as the target position coordinates and fill them into the sensing result message; set the target ID in the sensing result message as a unique target tracking ID; determine the data confidence level of the sensing result message based on the attenuation degree of the physical excitation signal strength; and inject the sensing result message into the domain controller via the vehicle bus. Thus, anchor point determination and coordinate injection can be completed dynamically and continuously in each control cycle, ensuring that the target position does not jump or get lost.

[0102] It should be noted that steps S31 and S32 can be continuously executed during the smooth relay transition phase, or they can continue during the near-field full injection phase after the physical excitation signal strength has attenuated to zero. During the near-field full injection phase, since the physical simulation equipment is completely silent, the test system continuously injects the perception result messages of the test target into the domain controller via the vehicle bus with full confidence. The target position coordinates in the messages still use the equivalent physical detection anchor point coordinates after effective visible sensing area truncation and dynamic sliding updates. Therefore, even at extremely close distances, the target position received by the domain controller remains continuous and reasonable, without jumps due to differences between the target's geometric center and the actual detection surface.

[0103] S4: When the physical excitation signal strength decays to zero, the signal output of the physical simulation device is completely stopped. The perception result message of the test target is continuously injected into the domain controller with full confidence through the vehicle bus to ensure that the test target is not lost, thereby verifying the control response of the test vehicle in a very close-range scenario.

[0104] In this embodiment, when the physical excitation signal strength attenuates to zero, the test system completely stops the signal output of the physical simulation device. Specifically, the high-resolution display screen in front of the physical simulation device no longer displays the test target image, the radar simulator becomes completely silent, and no longer emits radar waves to the real sensors of the test vehicle. At this time, the physical simulation device can enter a standby state or a shutdown state to avoid interference from residual physical excitation signals to the real sensors of the test vehicle.

[0105] It's important to note that in this step, the test system does not simultaneously stop all data sources. Instead, it completely halts the signal output of the physical simulation equipment, continuously injecting perception result messages of the test target into the domain controller solely through the vehicle bus. That is, the target information received by the domain controller in the near-field phase no longer comes from the actual physical sensors, but from the perception result messages injected via the vehicle bus, which are injected with full confidence. Full confidence can be set to 100%, meaning the data confidence field reaches its defined maximum value. Through full-confidence injection, the domain controller can treat the bus-injected perception result messages as a completely reliable perception data source, thus maintaining stable tracking of the test target even in extremely close-range areas where physical sensors are ineffective due to near-field distortion.

[0106] In practice, the test system can continuously inject perception result messages from the test target at high frequency via the vehicle bus. The injection frequency can be matched to the domain controller's control cycle, for example, with a cycle of 10ms, 20ms, or 50ms. The perception result message includes fields such as target ID, target location coordinates, and data confidence level. The target ID is set to the aforementioned unique target tracking ID, such as ID_001; the target location coordinates are the equivalent physical detection anchor point coordinates after effective visible perception area interception and dynamic sliding updates; and the data confidence level is set to full confidence, such as 100%. Thus, the target information received by the domain controller at extremely close range maintains consistency in identity with the previous physical perception stage, remains continuous with the nearest detection point of the actual sensor in location, and meets the domain controller's requirements for the reliability of the perception results.

[0107] For example, in an extremely close-range emergency lane-changing automatic emergency braking test scenario, when the smooth transition ends, the test target is only 1.5 meters, or even 0.5 meters, away from the test vehicle. At this point, the high-resolution display screen no longer shows the test target, and the radar simulator is completely silent. The test system continuously injects perception result messages processed by anchor point slippage into the domain controller at high frequency with 100% full confidence, solely through the vehicle bus. Due to the extremely smooth transition in step S3, the target tracking ID inside the domain controller, such as ID_001, is completely preserved. The domain controller continues to stably track the test target based on the continuous, non-abrupt injected data and verifies the control response of the test vehicle according to the preset control strategy. Finally, at an extremely dangerous moment when the distance is extremely close, such as 0.5 meters, the domain controller accurately issues the maximum braking deceleration command based on the continuous, non-abrupt injected data, causing the test vehicle's calipers to lock up, successfully completing the high-fidelity verification of the automatic emergency braking control strategy in the extremely close-range emergency lane-changing scenario.

[0108] It should be noted that during this step, the test system can still continue to execute S31 and S32. Simultaneously, when the relative movement of the test target causes a change in the effective visible sensing area, the test system re-extracts the geometric point closest to the test vehicle within the effective visible sensing area in real time, ensuring that the equivalent physical detection anchor point is always updated by sliding within the dynamic boundary of the effective visible sensing area. Therefore, even after the physical excitation signal has completely stopped, the target position in the injected message remains continuous and reasonable, without jumping due to differences between the target's geometric center and the actual detection surface, and without triggering a reasonableness check error in the domain controller due to anchor point fixation or jumps.

[0109] It should be noted that the specific values ​​such as 1.5 meters, 0.5 meters, and 100% mentioned above are only used to illustrate the technical approach of this embodiment and are not intended to limit the scope of protection of this invention. Without departing from the core logic of virtual-real data transfer and anchor point sliding in this invention, adaptive adjustments can be made according to specific test scenarios, sensor hardware characteristics, and test bench environments.

[0110] Reference Figure 4 As shown, this embodiment of the invention also provides a vehicle-in-the-loop testing system, the system comprising:

[0111] The physical excitation and recognition module 100 is used to generate test targets in a virtual simulation scene, control the physical simulation equipment to continuously output physical excitation signals corresponding to the test targets to the real sensors of the test vehicle, so that the domain controller of the test vehicle can recognize the test targets based on real physical perception and generate a unique target tracking ID for the test targets;

[0112] The status monitoring and triggering module 200 is used to acquire the relative distance or collision time between the test target and the test vehicle in the virtual simulation scene in real time. When the relative distance or collision time is reduced to below the preset physical performance boundary threshold, a smooth relay transition is triggered.

[0113] The smooth relay transition module 300 is used to control the physical excitation signal strength and screen visibility output by the physical simulation device to gradually decrease during the smooth relay transition process, and simultaneously inject the perception result message of the test target into the domain controller through the vehicle bus; wherein, the target ID in the perception result message is set as the unique target tracking ID, and the data confidence of the perception result message increases inversely as the physical excitation signal decreases.

[0114] The near-field full injection verification module 400 is used to completely stop the signal output of the physical simulation device when the physical excitation signal strength decays to zero, and continuously inject the perception result message of the test target into the domain controller with full confidence only through the vehicle bus to keep the test target from being lost, thereby verifying the control response of the test vehicle in extremely close-range scenarios.

[0115] As an example, the preset physical performance boundary threshold is taken from the maximum of the minimum effective microwave simulation distance of the radar simulator in the physical simulation device and the critical distance at which the camera of the test vehicle produces parallax distortion when shooting the display screen of the physical simulation device.

[0116] As an example, the gradual attenuation of the physical excitation signal strength and the reverse increase of the data confidence of the perception result message follow a linear mapping relationship or an S-shaped curve mapping relationship; and during the gradual attenuation of the screen visibility, the transparency of the test target screen on the display screen in the physical simulation device gradually increases with time or the shortening of the relative distance until it is completely transparent.

[0117] As an example, refer to Figure 5 As shown, the smooth relay transition module 300 includes:

[0118] The first extraction and determination unit 301 is used to obtain the size of the three-dimensional contour bounding box and the geometric center coordinates of the test target in the virtual simulation scene. Based on the real-time relative pose of the test vehicle and the test target, it determines the physical contour point closest to the test vehicle on the edge surface of the three-dimensional contour bounding box and defines it as the equivalent physical detection anchor point.

[0119] The first coordinate injection unit 302 is used to inject the coordinates of the equivalent physical detection anchor point as the target position coordinates into the sensing result message.

[0120] As an example, the smooth relay transition module 300 further includes:

[0121] The spatial interception calculation subunit 303 is used to obtain the effective physical detection field of view model of the real sensor on the test vehicle, and calculate the spatial intersection part intercepted by the effective physical detection field of view model and the test target based on the real-time relative pose of the effective physical detection field of view model and the test target, and define the exposed surface of the spatial intersection part as the effective visible perception area.

[0122] The nearest point extraction subunit 304 is used to extract the geometric point closest to the test vehicle within the effective visible perception area as the physical contour point and define it as the equivalent physical detection anchor point.

[0123] As an example, the nearest point extraction subunit 304 is also used to: when the relative movement of the test target causes the effective visible perception area to change, re-extract the geometric point closest to the test vehicle within the effective visible perception area in real time, so that the equivalent physical detection anchor point is always updated by sliding within the dynamic boundary of the effective visible perception area.

[0124] It is understood that the various test methods and control logics described in the embodiments of this application can be implemented by computer program instructions running on a general-purpose computing device (such as a test bench main controller), or can be solidified and executed by application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other hardware logic circuits; the corresponding program code can be tangibly stored in any non-volatile computer-readable storage medium.

[0125] Based on the same overall inventive concept, the method embodiments and system embodiments provided in this application are mirror images of each other, supporting and referencing each other in terms of underlying logic and beneficial effects. To ensure the conciseness and compactness of the entire specification, technical details not elaborated in detail in the system embodiments can be directly referred to in the corresponding descriptions in the foregoing method section.

[0126] Finally, it should be noted that the specific implementation details described in the specification (such as specific physical boundary distances, specific confidence decay curves, etc.) are only for the purpose of helping to thoroughly understand the technical approach of the present invention and are by no means an objective limitation on the scope of protection of the present invention. It should be understood that, without departing from the core logic of virtual-real data transfer and anchor point sliding of the present invention, any equivalent substitution, parameter fine-tuning, or adaptive reorganization of features made according to specific sensor hardware characteristics or test bench environment will necessarily fall within the legal protection scope defined by the appended claims.

Claims

1. A vehicle-in-the-loop testing method, characterized in that, Includes the following steps: In a virtual simulation scenario, test targets are generated, and the physical simulation equipment is controlled to continuously output physical excitation signals corresponding to the test targets to the real sensors of the test vehicle. This enables the domain controller of the test vehicle to identify the test targets based on real physical perception and generate a unique target tracking ID for the test targets. The relative distance or collision time between the test target and the test vehicle in the virtual simulation scene is acquired in real time. When the relative distance or collision time is reduced to below the preset physical performance boundary threshold, a smooth relay transition is triggered. During the smooth relay transition, the strength of the physical excitation signal and the visibility of the image output by the physical simulation device gradually decrease, and the perception result message of the test target is injected into the domain controller through the vehicle bus at the same time; wherein, the target ID in the perception result message is set as the unique target tracking ID, and the data confidence of the perception result message increases inversely as the physical excitation signal decreases. When the physical excitation signal strength decays to zero, the signal output of the physical simulation device is completely stopped. Only the perception result message of the test target is continuously injected into the domain controller with full confidence through the vehicle bus to keep the test target from being lost, thereby verifying the control response of the test vehicle in a very close-range scenario.

2. The vehicle-in-the-loop testing method according to claim 1, characterized in that, The preset physical performance boundary threshold is taken from the maximum value of the minimum effective microwave simulation distance of the radar simulator in the physical simulation equipment and the critical distance at which the camera of the test vehicle produces parallax distortion when shooting the display screen of the physical simulation equipment.

3. The vehicle-in-the-loop testing method according to claim 1, characterized in that, The gradual attenuation of the physical excitation signal strength and the reverse increase of the data confidence of the perception result message follow a linear mapping relationship or an S-shaped curve mapping relationship; and during the gradual attenuation of the screen visibility, the transparency of the test target screen on the display screen in the physical simulation device gradually increases with time or the shortening of the relative distance until it is completely transparent.

4. The vehicle-in-the-loop testing method according to claim 1, characterized in that: The perception result message of the test target is injected into the domain controller via the vehicle bus, including: The size of the three-dimensional contour bounding box and the coordinates of the geometric center of the test target in the virtual simulation scene are obtained. Based on the real-time relative pose of the test vehicle and the test target, the physical contour point closest to the test vehicle is determined on the edge surface of the three-dimensional contour bounding box and defined as the equivalent physical detection anchor point. The coordinates of the equivalent physical detection anchor point are used as the target location coordinates and injected into the sensing result message.

5. The vehicle-in-the-loop testing method according to claim 4, characterized in that: The physical contour point closest to the test vehicle is determined on the edge surface of the three-dimensional contour bounding box and defined as the equivalent physical detection anchor point, including: Obtain the effective physical field of view model of the real sensor on the test vehicle. Based on the real-time relative pose of the effective physical field of view model and the test target, calculate the spatial intersection part of the three-dimensional contour bounding box intercepted by the effective physical field of view model, and define the exposed surface of the spatial intersection part as the effective visible perception area. Within the effective visible sensing area, the geometric point closest to the test vehicle is extracted as the physical contour point and defined as the equivalent physical detection anchor point.

6. The vehicle-in-the-loop testing method according to claim 5, characterized in that: When the relative movement of the test target causes a change in the effective visible sensing area, the geometric point closest to the test vehicle within the effective visible sensing area is extracted in real time, so that the equivalent physical detection anchor point is always updated by sliding within the dynamic boundary of the effective visible sensing area.

7. A vehicle-in-the-loop testing system, characterized in that, The system includes: The physical stimulus and recognition module is used to generate test targets in a virtual simulation scenario, control the physical simulation equipment to continuously output physical stimulus signals corresponding to the test targets to the real sensors of the test vehicle, so that the domain controller of the test vehicle can recognize the test targets based on real physical perception and generate a unique target tracking ID for the test targets; The status monitoring and triggering module is used to obtain the relative distance or collision time between the test target and the test vehicle in the virtual simulation scene in real time. When the relative distance or collision time is reduced to below the preset physical performance boundary threshold, the smooth relay transition is triggered. The smooth relay transition module is used to control the physical excitation signal strength and screen visibility output by the physical simulation device to gradually decrease during the smooth relay transition process, and simultaneously inject the perception result message of the test target into the domain controller through the vehicle bus; wherein, the target ID in the perception result message is set as the unique target tracking ID, and the data confidence of the perception result message increases inversely as the physical excitation signal decreases. The near-field full injection verification module is used to completely stop the signal output of the physical simulation device when the physical excitation signal strength decays to zero. It continuously injects the perception result message of the test target into the domain controller with full confidence only through the vehicle bus to keep the test target from being lost, thereby verifying the control response of the test vehicle in extremely close-range scenarios.

8. A vehicle-in-the-loop testing system according to claim 7, characterized in that, The preset physical performance boundary threshold is taken from the maximum value of the minimum effective microwave simulation distance of the radar simulator in the physical simulation equipment and the critical distance at which the camera of the test vehicle produces parallax distortion when shooting the display screen of the physical simulation equipment.

9. A vehicle-in-the-loop testing system according to claim 7, characterized in that, The gradual attenuation of the physical excitation signal strength and the reverse increase of the data confidence of the perception result message follow a linear mapping relationship or an S-shaped curve mapping relationship; and during the gradual attenuation of the screen visibility, the transparency of the test target screen on the display screen in the physical simulation device gradually increases with time or the shortening of the relative distance until it is completely transparent.

10. A vehicle-in-the-loop testing system according to claim 7, characterized in that: The smooth relay transition module includes: The first extraction and determination unit is used to obtain the size of the three-dimensional contour bounding box and the coordinates of the geometric center of the test target in the virtual simulation scene. Based on the real-time relative pose of the test vehicle and the test target, the unit determines the physical contour point closest to the test vehicle on the edge surface of the three-dimensional contour bounding box and defines it as the equivalent physical detection anchor point. The first coordinate injection unit is used to inject the coordinates of the equivalent physical detection anchor point as the target position coordinates into the sensing result message.

Images