Prediction of ground subsidence risk using ground reflection characteristics and vehicle control system and method

Abstract

The present invention relates to a ground subsidence prediction and vehicle control system and method using ground reflection characteristics, which analyzes ground reflection characteristics based on radar signals to predict a risk of ground subsidence and automatically controls the vehicle to provide a driver warning to ensure safety. The method implements a ground subsidence prediction and vehicle control method using ground reflection characteristics by collecting radar frame data output from a forward radar, extracting ground reflection components that are distinguishable from moving objects from the collected radar frame data, analyzing ground reflection characteristics to calculate ground subsidence from the extracted ground reflection components, calculating the subsidence risk by comparing the analyzed ground reflection characteristics with reference road surface characteristics, suppressing false detections by removing specific reflection patterns to determine false detections regarding the calculated subsidence risk, warning the driver according to the calculated subsidence risk, and automatically controlling the vehicle to avoid the risk.

Description

Technology Field

[0001] The present invention relates to predicting ground subsidence using ground reflection characteristics, and in particular, to a ground subsidence prediction and vehicle control system and method using ground reflection characteristics that analyzes ground reflection characteristics based on radar signals to predict a risk of ground subsidence and, when such risk is predicted, provides a driver warning and automatically controls the vehicle to ensure safety. Background Technology

[0002] Recently, sinkhole accidents caused by ground weakening or cavities have been occurring on urban roads, and drivers often find it difficult to detect them in advance.

[0003] Generally, a sinkhole refers to a hole or depression formed when the surface of the earth suddenly subsides, usually occurring when underground rocks are dissolved by water or when caves or underground spaces collapse. They vary in size and depth, ranging from as small as 1 meter to hundreds of meters, and can be caused by a combination of natural and artificial factors.

[0004] The causes of sinkholes include natural causes such as the dissolution of soluble rocks like limestone by groundwater or rainwater, cave collapses, soil erosion, changes in groundwater levels, and earthquakes, as well as artificial causes such as excessive groundwater extraction, construction and development activities, leakage in sewers and water supply systems, underground excavation work, and changes in surface water flow.

[0005] It is extremely difficult for a driver to recognize ground subsidence occurring in this way while driving.

[0006] Meanwhile, vehicles primarily use forward radar for detecting objects ahead; however, since this is intended for detecting moving objects such as vehicles and pedestrians, most of the road surface reflection components are removed as clutter.

[0007] In addition, vehicles also utilize camera technology for road surface recognition; however, camera-based road surface recognition technology has limitations in ensuring reliability due to vulnerability to lighting, shadows, rain, and nighttime environments.

[0008] Alternatively, technologies such as GPR for ground cavity detection are vehicle- or infrastructure-centric, and there are limitations in terms of cost and configuration when applying them to ADAS in mass-produced vehicles.

[0009] Therefore, there is a need for a technology that utilizes the forward radar generally equipped in existing vehicles as is, but adds a function to analyze only ground reflection signals from the signals acquired by the forward radar to predict the risk of ground subsidence in advance and ensure safety. Prior art literature

[0010] Republic of Korea Published Patent 10-2025-0072833 (Published May 26, 2025) Republic of Korea Published Patent 10-2025-0132770 (Published September 5, 2025) Republic of Korea Published Patent 10-2025-0178680 (Published December 29, 2025) Republic of Korea Published Patent 10-2026-0012636 (Published January 27, 2026) Republic of Korea Registered Patent 10-2370908 (Published March 7, 2022) Republic of Korea Registered Patent 10-2712602 (Published October 2, 2024) Republic of Korea Registered Patent 10-2738057 (Published December 4, 2024) The problem to be solved

[0011] Accordingly, the present invention is proposed to resolve the limitations of determining or predicting ground subsidence using technology that recognizes objects or the ground using front radar and cameras in conventional vehicles as described above. The purpose of the invention is to provide a ground subsidence prediction and vehicle control system and method utilizing ground reflection characteristics, which analyzes ground reflection characteristics based on radar signals to predict a risk of ground subsidence and automatically controls the vehicle and warns the driver to ensure safety.

[0012] Another objective of the present invention is to provide a ground subsidence prediction and vehicle control system and method using ground reflection characteristics that predicts ground subsidence by analyzing ground reflection characteristics and improves driving safety by performing warnings and automatic vehicle control according to the risk level. means of solving the problem

[0013] To achieve the above-mentioned objectives, the "ground subsidence prediction and vehicle control system using ground reflection characteristics" according to the present invention comprises:

[0014] A forward radar signal collection unit that collects radar frame data output from a forward radar;

[0015] A ground reflection component separation unit that extracts ground reflection components distinguished from moving objects from the above-mentioned collected radar frame data;

[0016] A ground reflection characteristic analysis unit that analyzes ground reflection characteristics to predict ground subsidence risk based on abnormalities in ground reflection characteristics from the ground reflection component extracted above;

[0017] A sinkhole risk calculation unit that calculates the sinkhole risk by comparing the ground reflection characteristics analyzed above with the standard road surface characteristics;

[0018] A false positive suppression judgment unit that suppresses false positives by removing a specific reflection pattern to determine false positives regarding the above-calculated sinkhole risk;

[0019] A warning output unit that warns the driver of the risk of sinking according to the sinking risk calculated by the sinking risk calculation unit above;

[0020] It is characterized by including a vehicle control unit that automatically controls the vehicle to avoid danger according to the sinkhole risk calculated by the sinkhole risk calculation unit.

[0021] Preferably, the forward radar signal acquisition unit is,

[0022] It is characterized by acquiring range, amplitude (RCS), Doppler, and angle-based data from radar frame data output from the aforementioned forward radar, and outputting the acquired data after performing preprocessing to align the time axis so that comparison between frames is possible.

[0023] Preferably, the ground reflection component separation unit is,

[0024] Considering the radar installation height / pitch (vehicle attitude), a candidate ROI (Region of Interest) for the road surface is set for a certain distance ahead, and within the set ROI, moving objects are removed based on relative speed, and only stationary and low-speed components are separated.

[0025] Preferably, the ground reflection characteristic analysis unit is,

[0026] It is characterized by extracting the reflection intensity abrupt change index, distance distribution discontinuity index, scattering (dispersion) index, and multiple reflection possibility index as ground reflection characteristics.

[0027] Preferably, the sinkhole risk calculation unit,

[0028] It is characterized by calculating an anomaly score by comparing the extracted ground reflection characteristic value with a normal road surface standard value, and calculating the sinkhole risk level by calculating the calculated anomaly score and the persistence index.

[0029] Preferably, the above false positive suppression judgment unit is,

[0030] It is characterized by removing repetitive reflection patterns at specific widths and lengths using steel plate / speed bump patterns, excluding wet road surfaces and puddles by distinguishing between uniform reflection changes over wide areas and depression types, excluding small, single-occurrence events from the risk of depression or limiting them to a low level, and suppressing false positives by lowering the risk level through vehicle dynamics cross-verification if there is a discrepancy with changes in suspension and acceleration.

[0031] Preferably, the above false positive suppression judgment unit is,

[0032] It is characterized by reducing false positives by increasing the confidence weight when repeated detections occur in the same section.

[0033] Preferably, the above sinkhole risk level is,

[0034] It is classified into levels 0 to 3, and at level 1, it provides a visual warning to alert the driver, at level 2, it performs audiovisual / haptic warnings and lowers the ACC speed target, and at level 3, it is characterized by recommending automatic deceleration and lane change inhibition or avoidance routes.

[0035] In addition, the "method for predicting ground subsidence and controlling vehicles using ground reflection characteristics" according to the present invention is,

[0036] (a) A step of collecting radar frame data output from a forward radar;

[0037] (b) a step of extracting ground reflection components that are distinguishable from moving objects from the collected radar frame data;

[0038] (c) A step of analyzing ground reflection characteristics to calculate ground depression from the ground reflection component extracted above;

[0039] (d) A step of calculating the sinkhole risk by comparing the ground reflection characteristics analyzed above with the reference road surface characteristics;

[0040] (e) a step of suppressing false positives by removing a specific reflection pattern to determine false positives regarding the sinkhole risk calculated above; and

[0041] (f) It is characterized by including a step of warning the driver according to the calculated sinkhole risk and automatically controlling the vehicle to avoid the risk. Effects of the invention

[0042] According to the present invention, there is an advantage in that the risk of ground subsidence can be predicted without additional sensors by utilizing existing forward radar hardware as is.

[0043] In addition, according to the present invention, there is also the advantage of being able to expand the range of application of existing ADAS radars through ground reflection component separation and characteristic analysis technology.

[0044] In addition, according to the present invention, reliability in actual driving environments can be improved through false positive suppression logic, and there is also the advantage of effectively improving driving safety by providing warnings and vehicle control before an accident occurs due to ground subsidence. Brief explanation of the drawing

[0045] FIG. 1 is a block diagram of a ground subsidence risk prediction and vehicle control system using ground reflection characteristics according to the present invention, and FIG. 2 is a flowchart showing a method for predicting ground subsidence risk and controlling a vehicle using ground reflection characteristics according to the present invention. Specific details for implementing the invention

[0046] A system and method for predicting ground subsidence risk and controlling vehicles using ground reflection characteristics according to a preferred embodiment of the present invention will be described in detail below with reference to the attached drawings.

[0047] The terms or words used in the present invention described below should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the present invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention.

[0048] Therefore, the embodiments described in this specification and the configurations illustrated in the drawings are merely preferred embodiments of the present invention and do not represent all technical concepts of the present invention; thus, it should be understood that various equivalents and modifications that can replace them may exist at the time of filing this application.

[0049] FIG. 1 is a configuration diagram of a ground subsidence risk prediction and vehicle control system using ground reflection characteristics according to a preferred embodiment of the present invention, and may include a forward radar signal collection unit (10), a ground reflection component separation unit (20), a ground reflection characteristic analysis unit (30), a subsidence risk calculation unit (40), a false positive suppression judgment unit (50), a warning output unit (60), and a vehicle control unit (70).

[0050] The forward radar signal collection unit (10) serves to collect radar frame data output from the ADAS forward radar.

[0051] This forward radar signal acquisition unit (10) acquires range, amplitude (RCS), doppler, and angle-based data from radar frame data output from the ADAS forward radar, and outputs the acquired data after performing preprocessing to align the time axis so that comparison between frames is possible.

[0052] The ground reflection component separation unit (20) serves to extract ground reflection components that are distinguished from moving objects from the collected radar frame data.

[0053] This ground reflection component separation unit (20) can set a road surface candidate ROI (Region of Interest) for a certain distance section in front, taking into account the radar installation height / pitch (vehicle attitude), and remove moving objects based on relative speed within the set ROI and separate only stationary and low-speed components.

[0054] The ground reflection characteristic analysis unit (30) plays the role of analyzing ground reflection characteristics to predict ground subsidence risk based on abnormalities in ground reflection characteristics from the extracted ground reflection component.

[0055] This ground reflection characteristic analysis unit (30) can extract the reflection intensity rapid change index, distance distribution discontinuity index, scattering (dispersion) index, and multiple reflection possibility index as ground reflection characteristics.

[0056] The sinkhole risk calculation unit (40) can calculate the sinkhole risk by comparing the ground reflection characteristics analyzed above with the standard road surface characteristics.

[0057] The sinkhole risk calculation unit (40) can calculate an anomaly score by comparing the extracted ground reflection characteristic value with a normal road surface standard value, and calculate the sinkhole risk level by calculating the anomaly score and a preset persistence index.

[0058] The false positive suppression judgment unit (50) plays the role of suppressing false positives by removing a specific reflection pattern to determine false positives regarding the above-calculated sinkhole risk.

[0059] This false detection suppression judgment unit (50) can suppress false detections by removing reflection patterns that repeat at a specific width and length as steel plate / speed bump patterns, distinguishing between uniform reflection changes and depressions over a wide area to exclude wet road surfaces and puddles, excluding small and single events from the risk of depression or limiting them to a low level, and lowering the risk level if there is a discrepancy with suspension and acceleration changes through vehicle dynamics cross-verification.

[0060] The warning output unit (60) serves to warn the driver of the risk of sinking according to the sinking risk calculated by the sinking risk calculation unit (40).

[0061] The vehicle control unit (70) plays the role of automatically controlling the vehicle according to the sinkhole risk calculated by the sinkhole risk calculation unit (40) to avoid the risk.

[0062] FIG. 2 is also a flowchart showing a “method for predicting ground subsidence and controlling a vehicle using ground reflection characteristics” according to the present invention, which may include: (a) a step of collecting radar frame data output from a forward radar (S101); (b) a step of extracting ground reflection components that are distinguished from moving objects from the collected radar frame data (S102); (c) a step of analyzing ground reflection characteristics to calculate ground subsidence from the extracted ground reflection components (S103); (d) a step of calculating a subsidence risk by comparing the analyzed ground reflection characteristics with reference road surface characteristics (S104); (e) a step of suppressing false detection by removing a specific reflection pattern to determine false detection regarding the calculated subsidence risk (S105-S107); and (f) a step of warning the driver according to the calculated subsidence risk and automatically controlling the vehicle to avoid danger (S108-S109).

[0063] The ground subsidence risk prediction and vehicle control system and method using ground reflection characteristics according to a preferred embodiment of the present invention configured as described above will be specifically explained as follows.

[0064] First, when the vehicle starts operating, the front radar signal collection unit (10) collects radar frame data output from the ADAS front radar. Here, for vehicles equipped with an ADAS system, the ADAS front radar can be used, and for vehicles not equipped with an ADAS system, a separate front radar must be installed to collect radar signals.

[0065] The above-described forward radar signal acquisition unit (10) acquires data based on range, amplitude (RCS), Doppler, and azimuth / angle from frame-unit radar data output from the ADAS forward radar, performs preprocessing to align the time axis so that frame-to-frame comparison is possible, and outputs the acquired data to the ground reflection component separation unit (20) (S101).

[0066] The above ground reflection component separation unit (20) extracts ground reflection components that are distinguished from moving objects from the collected radar frame data.

[0067] For example, the ground reflection component separation unit (20) sets a road surface candidate ROI (Region of Interest) for extracting road surface information for a certain distance section in front, taking into account the radar installation height / pitch (vehicle attitude), and within the set ROI, removes moving objects based on relative speed (Doppler) and separates only stationary and low-speed components that can be considered as road surface reflections and transmits them to the ground reflection characteristic analysis unit (30) (S102).

[0068] The above ground reflection characteristic analysis unit (30) analyzes ground reflection characteristics to calculate ground depression from the extracted ground reflection component (S103).

[0069] For example, the ground reflection characteristic analysis unit (30) can extract a reflection intensity abrupt change index, a distance distribution discontinuity index, a scattering (dispersion) index, and a multiple reflection possibility index as ground reflection characteristics. Here, the reflection intensity abrupt change index is calculated using the average / median reflection signal intensity (amplitude) change amount between frames, and the distance distribution discontinuity index is extracted using abnormal peaks / gaps of the distance histogram within the ROI. In addition, the scattering (dispersion) index is calculated using the variance / entropy of the reflection signal intensity or point density, and the multiple reflection possibility index can be extracted by checking whether it is maintained for more than N frames (temporal consistency) at the same location.

[0070] The ground reflection characteristic value extracted in this way is transmitted to the sinkhole risk calculation unit (40).

[0071] The sinkhole risk calculation unit (40) calculates the sinkhole risk by comparing the analyzed ground reflection characteristics with the preset standard road surface characteristics (S104).

[0072] For example, the sinkhole risk calculation unit (40) calculates an anomaly score by comparing the extracted ground reflection characteristic value with a normal road surface reference value (feature baseline). Here, the normal road surface reference value can be selected and set from a factory default value, a learned vehicle-specific standard, and a cumulative standard for the same driving section. Then, the sinkhole risk level is calculated by calculating the calculated anomaly score and the persistence index (multiple reflection persistence index).

[0073] Here, the sinkhole risk level is classified from Level 0 to Level 3. Level 1 provides a visual warning to alert the driver, Level 2 provides an audiovisual / haptic warning and lowers the ACC speed target, and Level 3 recommends automatic deceleration, lane change inhibition, or an avoidance route.

[0074] When the risk level is calculated, it is not used immediately, but rather the presence or absence of a false positive is determined to suppress false positives as much as possible (S105). That is, the false positive suppression judgment unit (50) suppresses false positives by removing a specific reflection pattern to determine a false positive with respect to the above-calculated sinkhole risk level.

[0075] For example, the false detection suppression judgment unit (50) recognizes a reflection pattern that repeats at a specific width and length as a steel plate / speed bump pattern and removes it from the reflection pattern, distinguishes between a uniform reflection change (low frequency) over a wide area and a sinkhole type (local sudden change) to determine it as a wet road surface and puddle and excludes it from the reflection pattern, and excludes potholes by excluding small and single events (low persistence) from the sinkhole risk or limiting them to a low level. Additionally, false detection may be suppressed by lowering the risk level through vehicle dynamics cross-verification if there is a discrepancy with changes in suspension and acceleration.

[0076] In addition, the false positive suppression judgment unit (50) can further reduce false positives by increasing the reliability weight when repeated detections occur in the same section. It can also store event logs (location, features, level) to be used for post-analysis / map updates.

[0077] In this way, a specific reflection pattern for suppressing false positives is removed, and if the part for which the sinkhole risk was calculated is the removed specific reflection pattern, the risk is lowered or invalidated and the risk is recalculated (S106). The sinkhole risk calculation unit (40) recalculates the anomaly score by comparing the remaining ground reflection characteristics with the normal road surface reference value (feature baseline), excluding the removed reflection characteristic value. Subsequently, the sinkhole risk level is finally calculated by calculating the calculated anomaly score and the persistence index (multiple reflection persistence index) (S107).

[0078] Afterwards, if the frame in which the above sinkhole latitude level is calculated does not continue for more than a preset number of consecutive frames, it is judged as a one-time event and the risk level is lowered or invalidated. On the other hand, if the calculated sinkhole risk level continues for more than a preset number of consecutive frames, the driver is warned of the predicted sinkhole risk level through the warning output unit (60) and the vehicle control unit (70) according to the sinkhole risk level, and safety is ensured through vehicle control (S108 - S109).

[0079] For example, the warning output unit (60) warns the driver of the risk of sinking according to the sinking risk calculated by the sinking risk calculation unit (40), and if the sinking risk level is 1, it instructs the driver to pay attention through a visual warning, and if the level is 2, it performs a warning using audiovisual / haptic, etc., and instructs the driver to lower the ACC target speed or request deceleration.

[0080] In addition, the vehicle control unit (70) automatically controls the vehicle according to the sinkhole risk calculated by the sinkhole risk calculation unit (40) to avoid the risk.

[0081] For example, if the sinkhole risk level is 3, it automatically slows down, prevents lane changes (or recommends an avoidance route), and records emergency lights / remote events if necessary.

[0082] According to the present invention described above, the risk of ground subsidence can be predicted without additional sensors by utilizing existing forward radar hardware as is, and the range of application of existing ADAS radar can be expanded through ground reflection component separation and characteristic analysis technology.

[0083] In addition, according to the present invention, reliability in actual driving environments can be improved through false positive suppression logic, and driving safety can be effectively improved by providing warnings and vehicle control before an accident occurs due to ground subsidence.

[0084] Although the invention made by the inventors has been specifically described according to the above embodiments, it is obvious to those skilled in the art that the invention is not limited to the above embodiments and can be modified in various ways without departing from the gist thereof. Explanation of the symbols

[0085] 10: Forward radar signal acquisition unit 20: Ground reflection component separation unit 30: Ground Reflection Characteristics Analysis Unit 40: Subsidence Risk Calculation Unit 50: False positive suppression judgment unit 60: Warning output unit 70: Vehicle control unit

Claims

Claim 1 A forward radar signal collection unit for collecting radar frame data output from a forward radar; a ground reflection component separation unit for extracting ground reflection components distinguished from moving objects from the collected radar frame data; a ground reflection characteristic analysis unit for analyzing ground reflection characteristics to predict ground sinkage risk based on abnormalities in ground reflection characteristics from the extracted ground reflection components; a sinkage risk calculation unit for calculating sinkage risk by comparing the analyzed ground reflection characteristics with reference road surface characteristics; a false positive suppression judgment unit for suppressing false positives by removing specific reflection patterns to determine false positives regarding the calculated sinkage risk; and a warning output unit for warning the driver of sinkage risk according to the sinkage risk calculated by the sinkage risk calculation unit.and includes a vehicle control unit that automatically controls the vehicle to avoid danger according to the sinkhole risk calculated by the sinkhole risk calculation unit; the ground reflection characteristic analysis unit calculates a reflection intensity abrupt change index using the average / median reflection signal intensity (amplitude) change amount between frames, extracts a distance distribution discontinuity index using abnormal peaks / gaps of the distance histogram within the ROI, calculates a scattering (variance) index using the variance / entropy of the reflection signal intensity or point density, and extracts a multiple reflection possibility index by verifying temporal consistency maintained at the same location for more than N frames; the false positive suppression judgment unit reduces false positives by increasing the reliability weight when repeated detection occurs in the same section; the sinkhole risk calculation unit calculates an anomaly score by comparing the extracted ground reflection characteristic value with a normal road surface reference value, and calculates a sinkhole risk level by performing calculations on the calculated anomaly score and the persistence index, wherein the sinkhole risk level is differentially classified into Level 1 to Level 3, and visual A ground subsidence prediction and vehicle control system utilizing ground reflection characteristics, characterized by being classified into Level 1, which issues a warning to alert the driver; Level 2, which performs audiovisual / haptic warnings and lowers the ACC speed target; and Level 3, which performs automatic deceleration and inhibits lane changes or recommends an avoidance path. Claim 2 A ground subsidence prediction and vehicle control system using ground reflection characteristics, wherein, in claim 1, the forward radar signal acquisition unit acquires range, amplitude (RCS), Doppler, and angle-based data from radar frame data output from the forward radar, and outputs the acquired data after performing preprocessing to align the time axis so that comparison between frames is possible. Claim 3 Claim 1, wherein the ground reflection component separation unit sets a road surface candidate ROI (Region of Interest) for a certain distance section in front, considering the radar installation height / pitch (vehicle body attitude), and removes moving objects based on relative speed within the set ROI and separates only stationary and low-speed components, thereby forming a ground subsidence prediction and vehicle control system using ground reflection characteristics. Claim 4 delete Claim 5 delete Claim 6 A ground subsidence prediction and vehicle control system using ground reflection characteristics, wherein, in claim 1, the false detection suppression judgment unit removes reflection patterns repeating at a specific width and length as steel plate / speed bump patterns, excludes wet road surfaces and puddles by distinguishing uniform reflection changes over a wide area and subsidence types, and excludes potholes by excluding small, single-occurrence events from the risk of subsidence or limiting them to a low level. Claim 7 A method for predicting ground subsidence and controlling a vehicle using ground reflection characteristics, comprising: (a) collecting radar frame data output from a forward radar; (b) extracting ground reflection components that are distinguishable from moving objects from the collected radar frame data; (c) analyzing ground reflection characteristics to calculate ground subsidence from the extracted ground reflection components; (d) calculating a subsidence risk by comparing the analyzed ground reflection characteristics with reference road surface characteristics; (e) suppressing false detections by removing specific reflection patterns to determine false detections regarding the calculated subsidence risk; and (f) warning a driver according to the calculated subsidence risk and automatically controlling the vehicle to avoid danger.

Images