A method, device, product, and storage medium for determining the validity of echo data.
By using a sliding window queue for multi-dimensional feature extraction and multi-threshold determination in ultrasonic radar, the problem of insufficient accuracy of echo data filtering methods is solved, enabling accurate filtering and validity judgment of echo data and improving the accuracy of obstacle detection.
Patent Information
- Application Number
- CN202611323000.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-28
- Publication Date
- 2026-11-10
AI Technical Summary
In existing ultrasonic radar technology, the filtering methods for echo data cannot effectively distinguish between valid and interference data, resulting in insufficient data accuracy and affecting the accuracy of obstacle detection.
By maintaining a sliding window queue, multi-dimensional feature extraction is performed, and the validity of echo data is determined by combining multiple feature threshold conditions. This avoids relying on a single threshold filtering method and utilizes the historical environmental information provided by the sliding window queue for continuity judgment.
It improves the accuracy of echo data, reduces false alarms and missed alarms, and ensures the accuracy of downstream data use. It is suitable for functions such as automatic parking, remote parking, and rear cross traffic warning.
Smart Images

Figure CN123066751A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology, and in particular to a method, device, product, and storage medium for determining the validity of echo data. Background Technology
[0002] Ultrasonic sensors (USS) are commonly used short-range environmental perception sensors in Advanced Driver Assistance Systems (ADAS), widely applied in scenarios such as automatic parking, collision warning, and blind spot monitoring. Their working principle is as follows: the ultrasonic sensor emits ultrasonic pulses of a specific frequency (such as 40kHz, 48kHz, or 58kHz), and calculates the distance to the obstacle by measuring the time difference (TOF) of the echo reflected from the obstacle and combining this with the speed of sound. The echo data from the ultrasonic radar, after filtering, can be used for subsequent obstacle detection, distance calculation, and decision-making control.
[0003] However, ultrasonic radar is susceptible to various interference factors in real-world operating environments. On one hand, ultrasonic radars on different vehicles may operate on the same frequency band, generating co-channel interference. On the other hand, reflected echoes from non-target objects such as road surfaces, guardrails, and curbs (e.g., ground echoes) can also be received by the sensor. Furthermore, environmental noise, multipath effects, and sensor-specific aftershocks can also lead to outliers in the echo data. After acquiring the echo data, it is necessary to evaluate it and filter out invalid data. Existing technologies typically employ filtering methods based on fixed threshold filtering, tracking filters, and co-channel interference suppression, but these methods have limited filtering capabilities and cannot guarantee data accuracy. Summary of the Invention
[0004] This invention provides a method, device, product, and storage medium for determining the validity of echo data, enabling accurate determination of the validity of echo data.
[0005] According to one aspect of this application, a method for determining the validity of echo data is provided, comprising: Acquire the currently collected echo data to be determined, wherein the echo data to be determined includes the echo distance; If the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is greater than or equal to the set window length threshold, multi-dimensional feature extraction is performed based on the echo data to be determined and the currently maintained sliding window queue to obtain multi-dimensional features; The validity of the echo data to be judged is determined based on the multidimensional features.
[0006] According to another aspect of this application, an electronic device is provided, the electronic device comprising: At least one processor, and a memory communicatively connected to said at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the method for determining the validity of echo data according to any embodiment of this application.
[0007] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the method for determining the validity of echo data according to any embodiment of this application.
[0008] According to another aspect of this application, a computer program product is provided, the computer program product including a computer program that, when executed by a processor, implements the method for determining the validity of echo data as described in any embodiment of this application.
[0009] The technical solution of this application embodiment acquires currently collected echo data to be judged, which includes echo distance. If the echo distance indicates the detection of an obstacle and the length of the currently maintained sliding window queue is greater than or equal to a set window length threshold, multi-dimensional feature extraction is performed based on the echo data to be judged and the currently maintained sliding window queue to obtain multi-dimensional features. The validity of the echo data to be judged is determined according to the multi-dimensional features. By maintaining the sliding window queue to judge the validity of the echo data, inaccurate results caused by judging the validity based on a single echo data are avoided. Multi-dimensional feature extraction is performed on the echo data to be judged and the sliding window queue to reflect the characteristics of the echo data to be judged from different dimensions, thereby realizing the judgment of the validity of the echo data to be judged and ensuring the accuracy of the results. By judging the validity of the echo data, accurate filtering of interference data can be achieved, providing a basis for downstream data use and avoiding false alarms or missed alarms.
[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description
[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0012] Figure 1 This is a flowchart of a method for determining the validity of echo data according to Embodiment 1 of this application; Figure 2 This is a flowchart of a method for determining the validity of echo data according to Embodiment 2 of this application; Figure 3 This is an example diagram of filtered echo data provided according to Embodiment 2 of this application; Figure 4 This is a schematic diagram of a device for determining the validity of echo data according to Embodiment 3 of this application; Figure 5 This is a schematic diagram of the structure of an electronic device that implements the method for determining the validity of echo data according to the embodiments of this application. Detailed Implementation
[0013] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.
[0014] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0015] Example 1 Figure 1This is a flowchart of a method for determining the validity of echo data provided in Embodiment 1 of this application. This embodiment is applicable to situations where the validity of echo data needs to be determined. The method can be executed by an echo data validity determination device, which can be implemented in hardware and / or software and can be configured in an electronic device.
[0016] Currently, the main technical solutions for filtering vehicle-mounted ultrasonic echoes are as follows: Option 1: Fixed threshold-based filtering method. This method filters echoes by setting a fixed echo intensity threshold or distance threshold, only classifying echoes as valid if their intensity exceeds the threshold. For example, noise can be filtered by setting an intensity threshold slightly higher than the maximum echo intensity. However, different driving scenarios (such as parking and driving scenarios) have different detection requirements for ultrasonic sensors, and a fixed set of thresholds cannot simultaneously meet the requirements of different scenarios.
[0017] Option 2: A filtering method based on a tracking filter. This method checks the occurrence of echoes over a relatively long time interval using a tracking filter. If an echo appears only occasionally (i.e., it is identified only once and does not reappear in subsequent ultrasound pulses), it is determined to be interference and filtered out. Furthermore, cross-verification can be performed using measurement results from multiple ultrasound sensors. However, this method mainly targets randomly occurring interference echoes and lacks effective means to handle continuously occurring abnormal echoes (such as outliers caused by sensor malfunction or continuous environmental interference).
[0018] Option 3: Filtering methods based on co-channel interference suppression. This type of approach avoids co-channel interference by adjusting the time interval of the detection period, using random delays, or employing prime number random arrays to misalign the transmitted and received waveforms. Additionally, some methods achieve interference removal by segmenting the interference in the echo signal. However, these methods often focus on suppressing co-channel interference and have limited filtering capabilities for echoes reflected from non-target objects (such as ground echoes and multipath echoes) and abnormal data generated by the sensor itself.
[0019] This application achieves the validity determination of echo data by maintaining a sliding window queue and performing multi-dimensional feature extraction on the echo data to be determined and the currently maintained sliding window queue.
[0020] The method provided in this application is applicable to vehicles equipped with ultrasonic radar, and is particularly suitable for functions such as Automatic Parking Assist (APA), Remote Parking Assist (RPA), Rear Cross Traffic Alert (RCTA), and Automatic Emergency Braking-Low Speed (AEB-L). In practical applications, the ultrasonic radar sends echo data to the domain controller at a fixed period (e.g., 50ms). The data frame contains information such as a timestamp, echo distance (in mm), echo intensity, and enable flag. The method provided in this application can run as a software module in the domain controller, performing real-time validity determination on each frame of data from each radar probe, thus achieving real-time filtering.
[0021] like Figure 1 As shown, the method includes: S101. Obtain the currently collected echo data to be judged, which includes the echo distance.
[0022] In this embodiment, the echo data to be judged can be understood as echo data with validity determination requirements. The echo distance is the distance between the target and the transmitting source; in this embodiment, the echo distance can be the distance between a detected obstacle and a vehicle. The ultrasonic radar collects echo data in real time. After collecting each frame of echo data, this frame is used as the echo data to be judged, and its validity is determined through the following steps. After judging one frame of echo data, the system waits for the next frame of echo data and continues the judgment process. The echo data to be judged includes the echo distance; for example, the unit of echo distance is mm. The echo data to be judged may also include information such as timestamps, echo intensity, and enable flags.
[0023] S102. If the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is greater than or equal to the set window length threshold, multi-dimensional feature extraction is performed based on the echo data to be determined and the currently maintained sliding window queue to obtain multi-dimensional features.
[0024] In this embodiment, a sliding window queue is used to maintain a set of echo data. The window length threshold can be preset; for example, a window length threshold of 15 means that the length of the sliding window queue is at most 15, that is, the number of echo data in the sliding window queue is at most 15. Multidimensional features can be understood as information that reflects the characteristics of echo data from different dimensions, such as mean, error, and offset.
[0025] The echo distance is used to determine whether an obstacle has been detected. Generally, if an obstacle is detected, its distance is used as the echo distance. If no obstacle is detected, a default value is returned as the echo distance. For example, the default distance when the ultrasonic probe does not detect an object can be set to 7000mm. This embodiment can directly determine whether an obstacle has been detected based on the echo distance. If the echo distance is the default value, it can be considered that no obstacle has been detected; otherwise, it is confirmed that an obstacle has been detected. If the echo distance indicates that an obstacle has been detected, the validity of the echo data to be judged is determined based on the length of the currently maintained sliding window queue.
[0026] The sliding window queue consists of echo data, which is updated accordingly as echo data is collected. The length of the currently maintained sliding window queue is determined and compared with a set window length threshold. If the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is greater than or equal to the set window length threshold, the validity of the echo data to be judged can be determined based on the sliding window queue. Each echo data point in the sliding window queue is analyzed, and features of the echo data are extracted from different dimensions, such as the mean and standard deviation. Based on the mean of the data and the echo data to be judged, it is determined whether there is a significant deviation in the echo data to be judged. Through feature extraction, multidimensional features are obtained.
[0027] S103. Determine the validity of the echo data to be judged based on multidimensional features. Validity is used to indicate whether the echo data to be judged is valid.
[0028] In this embodiment, validity can be represented by natural language, flag bits, etc. For example, validity can be represented by flag bits, with a flag bit of 1 indicating validity and a flag bit of 0 indicating invalidity. Different discrimination conditions for different features are preset, and the extracted multidimensional features are compared with the corresponding discrimination conditions. If the conditions are met, the data is valid; otherwise, the data is invalid. For example, the discrimination conditions could be that the mean does not exceed a set threshold, or that the mean is within a certain range, etc.
[0029] This application provides a method for determining the validity of echo data. It maintains a sliding window queue to judge the validity of echo data. The sliding window queue provides historical environmental information about the current frame's vicinity, enabling filtering decisions to be based on the continuity of the time series rather than judging a single frame in isolation. This avoids inaccurate results caused by judging validity based on a single echo data point. Multi-dimensional feature extraction is performed on the echo data to be judged and the sliding window queue, reflecting the characteristics of the echo data from different dimensions, thereby achieving the validity judgment of the echo data and ensuring the accuracy of the results. By judging the validity of echo data, accurate filtering of interfering data can be achieved, providing a foundation for downstream data use and avoiding false alarms or missed alarms.
[0030] Example 2 Figure 2 This is a flowchart illustrating a method for determining the validity of echo data according to Embodiment 2 of this application. This embodiment is a refinement based on the above embodiment. Figure 2 As shown, the method includes: S201. Obtain the currently collected echo data to be judged, which includes the echo distance.
[0031] S202. If the echo distance indicator does not detect any obstacle, the validity of the echo data to be judged is determined to be invalid.
[0032] Analyze the echo distance value to determine whether an obstacle was detected. For example, a default echo distance indicates that no obstacle was detected. If the echo distance indicates that no obstacle was detected, the validity of the echo data to be judged is directly determined to be invalid, and the complex judgment process is not performed. Proceed to step S207.
[0033] S203. If the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is less than the set window length threshold, the validity of the echo data to be judged is determined to be valid.
[0034] If the echo distance indicator detects an obstacle, the length of the currently maintained sliding window queue is determined. If this length is less than the set window length threshold, the validity of the echo data to be judged is directly determined as valid, and the complex judgment process is no longer executed.
[0035] For example, the window length threshold is N, and the echo distance of the echo data to be judged is x. If the length of the currently maintained sliding window queue is less than N, a flag is set based on whether x is equal to the default value (the default distance value when the ultrasonic probe does not detect an object; this application takes the default value of 7000mm as an example). (If it is equal to 7000, the flag is 0; otherwise, it is 1), and the process continues to wait for the next frame. If the currently maintained sliding window queue has reached N, and the echo distance x is not the default value, multi-dimensional feature extraction is performed based on the echo data to be judged and the currently maintained sliding window queue, and the validity of the echo data to be judged is determined based on the extracted multi-dimensional features. If the echo distance x is the default value (i.e., x=7000), it is directly judged as invalid (flag 0), and no complex discrimination is performed. The process continues to execute step S207.
[0036] S204. If the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is greater than or equal to the set window length threshold, multi-dimensional feature extraction is performed based on the echo data to be determined and the currently maintained sliding window queue to obtain multi-dimensional features.
[0037] Optional multidimensional features include: effective mean, first coefficient of variation, offset distance, and relative error.
[0038] In this embodiment, the effective mean can be understood as the mean of the effective echo data; the first coefficient of variation is a parameter used to measure the degree of fluctuation of historical data; the offset distance is used to represent the offset of the echo data to be judged relative to the effective array. In this embodiment, the offset distance is the distance determined by the joint time-distance distribution; the relative error is used to represent the deviation of the echo distance from the trend of historical data change. The distance of the echo data to be judged is predicted, and the relative error between the echo distance and the predicted distance is calculated. In this embodiment, the relative error is determined based on the predicted distance. The predicted distance can be obtained by curve fitting. The trend of historical data change is determined by curve fitting, and then the echo distance of the echo data to be judged is predicted.
[0039] S205. When the multidimensional features meet the first condition and / or the second condition, the validity of the echo data to be judged is determined to be invalid.
[0040] S206. When the multidimensional features do not meet the first condition and do not meet the second condition, the validity of the echo data to be judged shall be determined as valid.
[0041] First and second conditions are preset. Multidimensional features are compared with the first and second conditions to determine if the multidimensional features satisfy the first condition. If satisfied, the echo data to be judged is deemed invalid. If not satisfied, the multidimensional features are then judged to satisfy the second condition. If satisfied, the echo data to be judged is deemed invalid. If neither condition is satisfied, the echo data to be judged is deemed valid. In this embodiment, the flag for valid echo data can be set to 1, and the flag for invalid echo data can be set to 0.
[0042] The first condition includes: the first coefficient of variation is less than the first coefficient threshold, the relative error is within the error range, and the echo distance of the echo data to be determined is not less than the effective mean. The second condition includes: the offset distance is greater than the offset distance threshold, the relative error is greater than the error threshold, and the echo distance of the echo data to be determined is not less than the effective mean or the first coefficient of variation is less than the second coefficient threshold.
[0043] In this embodiment, the first coefficient threshold is greater than the second coefficient threshold. For example, the first coefficient threshold is 10, the second coefficient threshold is 0.3, the error range can be (0.15, 0.8), the offset distance threshold is 10, and the error threshold is 3%.
[0044] For example, the validity determination process for the echo data to be evaluated is as follows: First condition (curve prediction filtering): When all of the following conditions are met, the echo data to be judged is marked as invalid (filtered). The first coefficient of variation (CV) is less than 10 (historical data is relatively stable, indicating greater confidence in the curve prediction). 0.15 < relative error <0.8 (significantly deviates from the predicted value, but not too much to avoid misjudgment); The echo distance x of the echo data to be judged is greater than or equal to the effective mean μ (only positive deviations are filtered out, because negative deviations usually correspond to closer distances and are mostly real obstacles).
[0045] Second condition (offset distance filtering): If the first condition is not met, then check: Offset distance d > 10 (significant deviation from joint distribution); And it satisfies x≥μ or CV<0.3 (positive deviation or historical data stationary); And relative error >3%.
[0046] If the second condition is met, the echo data to be determined is marked as invalid.
[0047] If none of the above conditions are met, the echo data to be judged is determined to be valid, and the flag is set to 1.
[0048] Quadratic fitting can approximate distance change curves (such as uniformly accelerated / decelerated motion). When the distance to obstacles changes gradually, the predicted value is relatively reliable, and the relative error can sensitively reflect anomalies. This application combines the fitted prediction with the coefficient of variation: this criterion is only used when the historical data is stable (CV < 10), avoiding misuse of predictions in scenarios with drastic fluctuations.
[0049] This application, in determining validity, does not rely on a single threshold but combines multiple features and introduces different threshold combinations (e.g., offset distance threshold of 10, relative error threshold of 0.03-0.8, etc.). These thresholds can be calibrated to adapt to different vehicle models and sensor characteristics. Simultaneously, the "≥mean" constraint is used in the conditions to ensure that only positive anomalies (sudden increases in distance) are filtered, because positive jumps are usually interference (such as ground reflections), while negative abrupt changes (sudden decreases in distance) often represent the presence of real obstacles. This application employs multi-level discrimination, first using trend prediction to filter points with obvious trend anomalies, and then using joint statistical distance to filter multi-dimensional anomalies, ensuring accuracy while avoiding the one-sidedness of a single criterion. Furthermore, the stationarity of historical data (i.e., the first coefficient of variation) is used as a weight to dynamically adjust the stringency of the filtering.
[0050] S207. Determine whether the length of the currently maintained sliding window queue is greater than or equal to the set window length threshold. If yes, execute S208; otherwise, execute S209.
[0051] S208. Remove the first element of the currently maintained sliding window queue, add the echo data to be judged to the tail of the currently maintained sliding window queue, and save the validity of the echo data to be judged accordingly.
[0052] S209. Add the echo data to be judged to the tail of the currently maintained sliding window queue, and save the validity of the echo data to be judged accordingly.
[0053] If the length of the currently maintained sliding window queue is greater than or equal to the set window length threshold, after determining the validity of the echo data to be judged, the echo distance of the echo data to be judged is added to the tail of the sliding window queue, and the head element is removed, keeping the length of the sliding window queue at N, thus updating the sliding window queue; and the validity of the echo data to be judged is saved accordingly. If the length of the currently maintained sliding window queue is less than the set window length threshold, the echo data to be judged is directly added to the tail of the currently maintained sliding window queue, and the validity of the echo data to be judged is saved accordingly.
[0054] This application embodiment can store echo data and validity data in two queues respectively; for example, define two queues of length N: `slide_window`: stores echo data in chronological order (or only stores echo distances from the echo data), initially empty; `valid_flag_window`: stores validity flags (1 for valid, 0 for invalid) at the corresponding position, initially empty. This application embodiment constructs a dynamic sliding window to save the distance values of the most recent N frames and maintains the validity flag corresponding to each value, providing historical context for subsequent statistical judgment. The sliding window provides historical environmental information about the vicinity of the current frame, enabling filtering decisions to be based on the continuity of the time series, rather than judging a single frame in isolation.
[0055] After determining the validity of the echo data to be judged, the valid distance value of the echo data to be judged (output echo distance x if valid, discard or output invalid mark if invalid) can be sent to the downstream fusion module (such as obstacle tracking, parking space detection, etc.).
[0056] Optionally, multi-dimensional feature extraction is performed based on the echo data to be determined and the currently maintained sliding window queue to obtain multi-dimensional features, including steps A1-A4: A1. Generate a valid array based on the currently maintained sliding window queue.
[0057] In this embodiment, the valid array can be understood as an array composed of valid echo data, which includes the echo data. The echo data included in the currently maintained sliding window queue is determined, and the validity of each echo data is determined. For example, validity is determined based on the flag bit of each echo data, or the echo data is processed to determine validity, etc. The valid echo data is extracted to form the valid array.
[0058] A2. Determine the effective mean and the first coefficient of variation based on the effective array.
[0059] The echo data in the valid array is analyzed, the mean is calculated to obtain the valid mean, the fluctuation of the echo data is analyzed, and the first coefficient of variation is calculated. For example, in the embodiments of this application, the calculation formula of the first variation parameter can be preset, the echo data is substituted into the formula for calculation, and the first coefficient of variation is obtained.
[0060] A3. Determine the offset distance and prediction distance of the echo data to be judged based on the echo data to be judged and the valid array.
[0061] Analyze the historical variation patterns of the echo data in the valid array, compare the echo data to be judged with the historical variation patterns, determine the offset distance, and use the offset distance to measure whether the echo data to be judged conforms to the data variation trend. Based on the historical variation patterns determined from the valid array, the echo distance of the echo data to be judged is predicted according to the historical variation patterns, and the obtained echo distance is recorded as the predicted distance.
[0062] A4. Determine the relative error based on the echo distance and predicted distance in the echo data to be determined.
[0063] Determine the echo distance in the echo data to be determined, compare this echo distance with the predicted distance, calculate the error value, and record it as the relative error. For example, the relative error between the echo distance and the predicted distance in the echo data to be determined... Where x is the echo distance in the echo data to be determined. To predict the distance; if =0, then let It is infinitely large.
[0064] Optionally, a valid array is generated based on the currently maintained sliding window queue, including: correcting misfiltering in the currently maintained sliding window queue to obtain the target sliding window queue; and extracting valid echo data from the target sliding window queue to form a valid array.
[0065] In this embodiment, the target sliding window queue can be understood as a queue formed by corrected echo data. The echo data included in the target sliding window queue is the same as the echo data included in the sliding window queue, but their validity may differ.
[0066] Analyze the echo data in the currently maintained sliding window queue. Based on factors such as the validity of the echo data, the validity of the preceding and following frames, and the trend of change, determine whether the validity of the echo data has been misjudged. If misjudged, correct the error, determine the validity of each corrected echo data, and form the target sliding window queue. Determine the validity of each echo data in the target sliding window queue, and extract the valid echo data to form a valid array.
[0067] Optionally, perform error filtering correction on the currently maintained sliding window queue to obtain the target sliding window queue, including steps B1-B3: B1. Obtain the last M frames of echo data from the sliding window queue, and use the M frames of echo data as the data to be corrected, where M is a positive integer.
[0068] In this embodiment, the data to be corrected can be understood as data whose validity needs to be corrected; M can be preset, and M is a positive integer. The M frames of echo data at the tail of the sliding window queue are obtained, and these M frames of echo data are used as the data to be corrected.
[0069] B2. Obtain the validity of each data to be corrected. If all the data to be corrected are invalid, calculate the second coefficient of variation based on the echo distance of each data to be corrected.
[0070] In this embodiment, the second coefficient of variation is a parameter used to measure the degree of fluctuation in historical data. The validity of each data point to be corrected is determined. It is then determined whether all data points to be corrected are invalid. If so, subsequent steps are needed to correct the validity of each data point. The echo distance of each data point to be corrected is determined, and the standard deviation and mean are calculated based on each echo distance. The ratio of the standard deviation to the mean is calculated and used as the second coefficient of variation. Otherwise, it is determined that there are no misclassifications in the data points to be corrected, and the validity of each data point to be corrected remains unchanged.
[0071] B3. If the second coefficient of variation is less than the set threshold and the echo distance of each data to be corrected indicates that an obstacle has been detected, it is determined to be a false filter, and the validity of each data to be corrected is updated to valid.
[0072] In this embodiment, the threshold can be set according to actual conditions, for example, a threshold of 3%. The second coefficient of variation is compared with the set threshold. If the second coefficient of variation is less than the set threshold, it is determined whether the echo distance of each piece of data to be corrected indicates that an obstacle has been detected. For example, if the echo distance of the data to be corrected is not the default value, it indicates that an obstacle has been detected. If the second coefficient of variation is less than the set threshold and the echo distance of each piece of data to be corrected indicates that an obstacle has been detected, it is determined to be a false filter, and the validity of each piece of data to be corrected is updated from invalid to valid.
[0073] This application's embodiments address the problem of consecutive valid frames being misclassified as invalid due to sudden environmental changes or overly strict discrimination criteria through a false filtering feedback correction mechanism. It automatically restores the validity of these frames, improving data utilization. This method monitors the statistical characteristics of consecutive invalid frames at the end of a sliding window; when they are found to be highly consistent and stable, they are automatically restored to their valid state. This mechanism achieves "error correction" using the inherent consistency of the data without requiring additional sensor information, significantly improving data utilization. This mechanism effectively compensates for the shortcomings of a single discrimination criterion, reducing the false detection rate, and is particularly suitable for scenarios where obstacle distances are stable but brief signal fluctuations cause filtering. Experimental data shows that this correction mechanism can improve data utilization efficiency by approximately 15% to 20%, reducing functional degradation caused by data loss.
[0074] For example, taking M=3 as an example, and setting the threshold to 3%, the implementation of the false filtering feedback correction mechanism can be as follows: In each judgment cycle, first check the valid flag bits of the last 3 frames (indexed as N-2, N-1, N) at the end of the sliding window queue. If all the flag bits of these 3 frames are 0 (i.e., they are all judged as invalid), and the following two conditions are met simultaneously: 1. The coefficient of variation (standard deviation / mean × 100%) of the distance values in these 3 frames is less than 3%; 2. There is no default value in these 3 frames; then it is judged as "false filtering", and the valid flag bits of these 3 frames are forcibly set to 1.
[0075] Optionally, determine the effective mean and the first coefficient of variation based on the effective array, including steps C1-C3: C1. Calculate the mean of the effective echo distances in the effective array, denoted as the effective mean.
[0076] In this embodiment, the effective echo distance refers to the echo distance in the effective array. The echo distances in the effective array are determined and recorded as the effective echo distances. The mean of each effective echo distance is calculated and recorded as the effective mean.
[0077] For example, a valid array is obtained through express, The effective echo distance is denoted by m, where m is the length of the effective array (i.e., the number of effective echo distances in the effective array); the effective mean is... .
[0078] C2. Calculate the standard deviation of the effective echo distance in the effective array, denoted as the effective standard deviation.
[0079] Based on the standard deviation calculation formula, the standard deviation of each effective echo distance is calculated, and the obtained standard deviation is recorded as the effective standard deviation.
[0080] For example, the effective standard deviation .
[0081] C3. Calculate the ratio of the effective standard deviation to the effective mean to obtain the first coefficient of variation.
[0082] The ratio of the effective standard deviation to the effective mean is denoted as the first coefficient of variation.
[0083] Statistics such as the effective mean, effective standard deviation, and first coefficient of variation reflect the central tendency and dispersion of historical distances, serving as a benchmark for subsequent anomaly detection.
[0084] Optionally, the offset distance of the echo data to be judged, including D1-D4, is determined based on the echo data to be judged and the valid array: D1. For each valid echo distance in the valid array, construct the corresponding valid echo vector based on the valid echo distance and its original index position. The original index position is the index position of the valid echo distance in the currently maintained sliding window queue.
[0085] In this embodiment, the effective echo vector is a two-dimensional vector of the effective frames. The effective value refers to the echo data in the effective array, which can also be called the effective echo data. This vector includes the original index position and the effective echo distance. The original index position is the index position of the effective echo distance in the currently maintained sliding window queue.
[0086] Each echo distance in the sliding window queue is sorted chronologically, and its corresponding index is determined based on its position. For example, the index of the first echo distance is 1. After generating the valid array V, the index position of each valid echo distance in the sliding window queue is extracted to obtain the original index position of each valid echo distance. These original index positions can be represented by an array. It means that among them The original index position is used (values range from 1 to N, where N is the window length threshold). For each valid echo distance, this valid echo distance and its original index position are used as elements of a vector to construct a vector, resulting in the vector corresponding to this valid echo distance, denoted as the valid echo vector. For example, the valid echo vector is... .
[0087] D2. Determine the covariance matrix and mean vector based on each effective echo vector.
[0088] According to the formula for calculating the covariance matrix, the effective echo vectors are processed to generate a covariance matrix. The average of the original index positions corresponding to each effective echo vector is taken to obtain the index mean. The average of the effective echo distances corresponding to each effective echo vector is taken to obtain the distance mean (i.e., the effective mean in this embodiment). A mean vector is constructed based on the index mean and the distance mean. The mean vector is... ,in, , μ is the effective mean.
[0089] D3. Generate the echo vector to be determined based on the echo distance and expected index position of the echo data to be determined.
[0090] In this embodiment, the expected index position is the location where the echo data to be determined will be added. The echo vector to be determined can be understood as a two-dimensional vector corresponding to the echo data to be determined. The echo data to be determined includes the echo distance. The expected index position of the echo data to be determined is determined according to the window length threshold. For example, the expected index position is N+1, where N is the window length threshold, and the echo vector to be determined is z = (N+1, x), where x is the echo distance of the echo data to be determined.
[0091] D4. Calculate the distance based on the echo vector to be determined, the covariance matrix, and the mean vector to obtain the offset distance of the echo data to be determined.
[0092] The distance between the echo vector to be determined and the covariance matrix and mean vector is calculated. For example, the two-dimensional Mahalanobis distance is calculated as the offset distance of the echo data to be determined. The offset distance calculated in this embodiment of the application not only considers the distance value, but also whether its occurrence time and position are consistent with historical patterns. This distance combines the distance value and time continuity, and can effectively distinguish those abnormal points (such as sudden pulse interference) whose distance values seem normal but whose occurrence time does not conform to the trend.
[0093] For example, this application provides a formula for calculating the offset distance d: , where S is the covariance matrix. S is a 2×2 covariance matrix, using an unbiased estimate.
[0094] This application determines distance offset using two-dimensional Mahalanobis distance, which incorporates time indexing as another dimension to construct a joint distribution of distance evolution over time. For example, as a vehicle approaches an obstacle at a constant speed, the distance value decreases linearly, and its time-distance points should be distributed along a straight line. If an anomalous echo appears at an inappropriate time point (such as a sudden jump), even if its distance value is within the historical range, its two-dimensional Mahalanobis distance will still be large, thus enabling effective identification. This fully leverages time series correlation and utilizes temporal context to determine the validity of echoes.
[0095] Optionally, the predicted distance of the echo data to be determined is determined based on the echo data to be determined and the valid array, including: solving the parameters of the pre-built fitting model based on the valid echo distance in the valid array and the corresponding original index position to obtain the model parameters; obtaining the current fitting model based on the model parameters; and predicting the distance based on the current fitting model and the expected index position corresponding to the echo data to be determined to obtain the predicted distance.
[0096] In this embodiment, the fitting model can be a mathematical model containing unknown parameters; the parameters of the current fitting model are known quantities, which can be substituted into the independent variables to solve for the dependent variable. A fitting model is pre-constructed, extracting multiple valid echo distances from the valid array, and the original index position corresponding to each valid echo distance. Data fitting is performed based on the valid echo distances and the original index positions to solve for the model parameters of the fitting model. The number of model parameters can be one or more. The solved model parameters are substituted into the fitting model to obtain the current fitting model. The expected index position corresponding to the echo data to be determined is substituted into the current fitting model for distance prediction to obtain the predicted distance. In this embodiment, when solving the fitting model, all valid echo distances in the valid array can be used, or only a portion of the valid echo distances can be selected. To ensure the accuracy of the fitting, when the length m of the valid array is greater than the threshold L, the model parameters can be obtained by solving the fitting model using the valid echo distances and their corresponding original index positions. When m is less than or equal to L, a default model is used for distance prediction, or the validity of the echo data to be determined is directly confirmed as invalid, and so on.
[0097] The fitting prediction in this application embodiment reflects the changing trend of obstacle distance (e.g., the vehicle gradually approaches during parking). If the current value deviates significantly from the predicted value, it is likely an abnormal echo.
[0098] For example, the fitted model is as follows: The coefficients are solved using the least squares method. , , Then, predict the distance of the echo data to be determined (corresponding to the expected index position N+1). .
[0099] All calculations in this application involve only basic statistics (mean, standard deviation), 2×2 matrix inversion, and quadratic polynomial least squares fitting. All these operations are analytical, requiring no iterative optimization and exhibiting low computational complexity. On typical ADAS domain controllers (such as the Infineon TC397 or NXP S32K series), the processing time for a single frame of data can be controlled at the microsecond level, fully meeting the real-time processing requirements of ultrasonic radar with a 50ms cycle, without burdening the overall computing power of the domain controller. The discrimination thresholds in this application (such as coefficient of variation thresholds of 3% and 10%, offset distance threshold of 10, and relative error thresholds of 0.03~0.8, etc.) are all configurable parameters and can be flexibly adjusted through vehicle calibration to adapt to the sensor installation positions of different vehicle models, the noise characteristics of different ultrasonic radar modules, and the filtering requirements of different functional scenarios, demonstrating good engineering portability and generalization capabilities.
[0100] For example, Figure 3An example graph of filtered echo data is provided. The horizontal axis represents time, and the vertical axis represents the distance of the ultrasonic echo. Each dot corresponds to one ultrasonic echo; solid red dots represent the filtered echoes, and the remaining hollow blue dots represent the normally output echoes after filtering. Figure 3 It is evident that clutter was effectively identified and filtered.
[0101] This application provides a method for determining the validity of echo data. By constructing discriminative features such as predicted distance and using multi-level joint discriminative logic, it achieves a comprehensive multi-angle judgment of abnormal echoes. Compared with existing filtering methods using a single threshold or single criterion, this application can effectively distinguish various types of abnormal data, such as co-channel interference, ground reflection echoes, multipath effect echoes, and sensor aftershocks, significantly reducing the probability of missed and false filtering. Real-world testing data shows that in typical parking scenarios, the accuracy of this method in identifying abnormal echoes can reach over 95%, significantly outperforming traditional fixed threshold methods. The false filtering feedback correction mechanism provided in this application monitors the coefficient of variation of multiple consecutive frames at the end of the sliding window. When consecutive invalid frames are found to have highly consistent data and no out-of-range values, these frames are automatically restored to a valid state. This mechanism effectively solves the problem of existing technologies erroneously filtering out consecutive valid data during brief signal fluctuations, and is particularly suitable for scenarios where the distance to obstacles remains stable while the vehicle is stationary or moving slowly. This application uses a fixed-length sliding window to store historical frame data and calculates the offset distance and predicted distance based on the historical valid frames, incorporating the continuity and trend of the echo distance over time into the discrimination criteria. The offset distance jointly considers the distance value and its corresponding time index, which can effectively identify abnormal echoes that, although the distance value is within the historical range, appear at times that do not conform to the motion pattern; quadratic curve fitting can predict the trend of obstacle distance changes and accurately filter out outliers that deviate from the trend. This allows this application to maintain good filtering performance even in dynamic driving scenarios (such as vehicles slowly approaching or moving away from obstacles).
[0102] Example 3 Figure 4 This is a schematic diagram of a device for determining the validity of echo data provided in Embodiment 3 of this application. Figure 4 As shown, the device includes: a data acquisition module 31 to be determined, a multi-dimensional feature extraction module 32, and a first validity determination module 33.
[0103] The data acquisition module 31 is used to acquire the currently collected echo data to be determined, which includes the echo distance. The multidimensional feature extraction module 32 is used to perform multidimensional feature extraction based on the echo data to be determined and the currently maintained sliding window queue if the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is greater than or equal to a set window length threshold, so as to obtain multidimensional features. The first validity determination module 33 is used to determine the validity of the echo data to be determined based on the multidimensional features, wherein the validity is used to indicate whether the echo data to be determined is valid.
[0104] This application provides a device for determining the validity of echo data. By maintaining a sliding window queue, the device judges the validity of echo data, avoiding inaccurate results caused by judging validity based on a single echo data point. Multi-dimensional feature extraction is performed on the echo data to be judged and the sliding window queue, reflecting the characteristics of the echo data from different dimensions, thereby achieving the validity judgment of the echo data and ensuring the accuracy of the results. By judging the validity of echo data, accurate filtering of interfering data can be achieved, providing a foundation for downstream data use and avoiding false alarms or missed alarms.
[0105] Optionally, the device may also include: The second validity determination module is used to determine that the validity of the echo data to be determined is invalid if the echo distance indicator does not detect an obstacle. The third validity determination module is used to determine the validity of the echo data to be judged as valid if the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is less than a set window length threshold.
[0106] Optionally, the multidimensional features include: effective mean, first coefficient of variation, offset distance, and relative error; Multidimensional feature extraction module 32 includes: The valid array generation unit is used to generate a valid array based on the currently maintained sliding window queue; An effective mean determination unit is used to determine the effective mean and the first coefficient of variation based on the effective array; A distance determination unit is used to determine the offset distance and predicted distance of the echo data to be determined based on the echo data to be determined and the valid array; The relative error determination unit is used to determine the relative error based on the echo distance in the echo data to be determined and the predicted distance.
[0107] Optionally, the effective array generation unit is specifically used for: correcting misfiltering in the currently maintained sliding window queue to obtain the target sliding window queue; and extracting valid echo data from the target sliding window queue to form an effective array.
[0108] Optionally, the step of correcting the erroneous filtering of the currently maintained sliding window queue to obtain the target sliding window queue includes: obtaining the last M frames of echo data of the sliding window queue, using the M frames of echo data as data to be corrected, where M is a positive integer; obtaining the validity of each of the data to be corrected; if all the data to be corrected are invalid, calculating a second coefficient of variation based on the echo distance of each of the data to be corrected; if the second coefficient of variation is less than a set threshold and the echo distance of each of the data to be corrected indicates that an obstacle has been detected, determining it as erroneous filtering, and updating the validity of each of the data to be corrected to valid.
[0109] Optionally, the effective mean determination unit is specifically used for: calculating the mean of the effective echo distances in the effective array, denoted as the effective mean; calculating the standard deviation of the effective echo distances in the effective array, denoted as the effective standard deviation; and calculating the ratio of the effective standard deviation to the effective mean to obtain the first coefficient of variation.
[0110] Optionally, the distance determination unit is specifically used for: constructing a corresponding effective echo vector for each effective echo distance in the effective array based on the effective echo distance and its original index position, wherein the original index position is the index position of the effective echo distance in the currently maintained sliding window queue; determining a covariance matrix and a mean vector based on each effective echo vector; generating a pending echo vector for the pending echo data based on the echo distance and expected index position of the pending echo data; and calculating the distance based on the pending echo vector, the covariance matrix, and the mean vector to obtain the offset distance of the pending echo data.
[0111] Optionally, the distance determination unit is specifically used for: solving the parameters of the pre-constructed fitting model based on the effective echo distances in the effective array and the corresponding original index positions to obtain model parameters; obtaining the current fitting model based on the model parameters; and performing distance prediction based on the current fitting model and the expected index positions corresponding to the echo data to be determined to obtain the predicted distance.
[0112] The first validity determination module 33 is specifically used for: determining the validity of the echo data to be determined as invalid when the multidimensional feature satisfies the first condition and / or the second condition; and determining the validity of the echo data to be determined as valid when the multidimensional feature does not satisfy the first condition and does not satisfy the second condition. The first condition includes: the first coefficient of variation is less than the first coefficient threshold, the relative error is within the error range, and the echo distance of the echo data to be determined is not less than the effective mean. The second condition includes: the offset distance is greater than the offset distance threshold, the relative error is greater than the error threshold, and the echo distance of the echo data to be determined is not less than the effective mean or the first coefficient of variation is less than the second coefficient threshold.
[0113] Optionally, the device may also include: The sliding window update module is used to determine whether the length of the currently maintained sliding window queue is greater than or equal to a set window length threshold. If so, the first element of the currently maintained sliding window queue is removed, the echo data to be determined is added to the tail of the currently maintained sliding window queue, and the validity of the echo data to be determined is saved accordingly. Otherwise, the echo data to be determined is added to the tail of the currently maintained sliding window queue, and the validity of the echo data to be determined is saved accordingly.
[0114] The echo data validity determination device provided in this application embodiment can execute the echo data validity determination method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects of executing the method.
[0115] Example 4 Figure 5 A schematic diagram of an electronic device 40 that can be used to implement embodiments of this application is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the application described and / or claimed herein.
[0116] like Figure 5 As shown, the electronic device 40 includes at least one processor 41 and a memory, such as a read-only memory (ROM) 42 or a random access memory (RAM) 43, communicatively connected to the at least one processor 41. The memory stores computer programs executable by the at least one processor. The processor 41 can perform various appropriate actions and processes based on the computer program stored in the ROM 42 or loaded from storage unit 48 into the RAM 43. The RAM 43 may also store various programs and data required for the operation of the electronic device 40. The processor 41, ROM 42, and RAM 43 are interconnected via a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.
[0117] Multiple components in electronic device 40 are connected to I / O interface 45, including: input unit 46, such as keyboard, mouse, etc.; output unit 47, such as various types of monitors, speakers, etc.; storage unit 48, such as disk, optical disk, etc.; and communication unit 49, such as network card, modem, wireless transceiver, etc. Communication unit 49 allows electronic device 40 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0118] Processor 41 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 41 performs the various methods and processes described above, such as methods for determining the validity of echo data.
[0119] In some embodiments, the method for determining the validity of echo data may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 40 via ROM 42 and / or communication unit 49. When the computer program is loaded into RAM 43 and executed by processor 41, one or more steps of the method for determining the validity of echo data described above may be performed. Alternatively, in other embodiments, processor 41 may be configured to perform the method for determining the validity of echo data by any other suitable means (e.g., by means of firmware).
[0120] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0121] Computer programs used to implement the methods of this application may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0122] This application provides a computer program product, which includes a computer program that, when executed by a processor, implements the method for determining the validity of echo data as described in any embodiment of this application.
[0123] In the context of this application, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
[0124] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0125] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.
[0126] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.
[0127] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this application can be achieved, and this is not limited herein.
[0128] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.
Claims
1. A method for determining the validity of echo data, characterized in that, include: Acquire the currently collected echo data to be determined, wherein the echo data to be determined includes the echo distance; If the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is greater than or equal to the set window length threshold, multi-dimensional feature extraction is performed based on the echo data to be determined and the currently maintained sliding window queue to obtain multi-dimensional features; The validity of the echo data to be determined is determined based on the multidimensional features, and the validity is used to indicate whether the echo data to be determined is valid.
2. The method according to claim 1, characterized in that, Also includes: If the echo distance indicator does not detect an obstacle, the validity of the echo data to be determined is invalid. If the echo distance indicator detects an obstacle and the length of the currently maintained sliding window queue is less than the set window length threshold, the validity of the echo data to be judged is determined to be valid.
3. The method according to claim 1, characterized in that, The multidimensional features include: effective mean, first coefficient of variation, offset distance, and relative error; the multidimensional feature extraction based on the echo data to be determined and the currently maintained sliding window queue yields multidimensional features, including: Generate a valid array based on the currently maintained sliding window queue; The effective mean and the first coefficient of variation are determined based on the effective array; The offset distance and prediction distance of the echo data to be determined are determined based on the echo data to be determined and the valid array; The relative error is determined based on the echo distance in the echo data to be determined and the predicted distance.
4. The method according to claim 3, characterized in that, The step of generating a valid array based on the currently maintained sliding window queue includes: Perform error filtering correction on the currently maintained sliding window queue to obtain the target sliding window queue; Extract valid echo data from the target sliding window queue to form a valid array.
5. The method according to claim 4, characterized in that, The step of correcting erroneous filtering in the currently maintained sliding window queue to obtain the target sliding window queue includes: Obtain the last M frames of echo data from the sliding window queue, and use the M frames of echo data as the data to be corrected, where M is a positive integer; The validity of each of the data to be corrected is obtained. If all of the data to be corrected are invalid, a second coefficient of variation is calculated based on the echo distance of each of the data to be corrected. If the second coefficient of variation is less than the set threshold and the echo distance of each of the data to be corrected indicates that an obstacle has been detected, it is determined to be a false filter, and the validity of each of the data to be corrected is updated to valid.
6. The method according to claim 3, characterized in that, The step of determining the effective mean and the first coefficient of variation based on the effective array includes: Calculate the mean of the effective echo distances in the effective array, and denot it as the effective mean; Calculate the standard deviation of the effective echo distance in the effective array, and denote it as the effective standard deviation; The first coefficient of variation is obtained by calculating the ratio of the effective standard deviation to the effective mean.
7. The method according to claim 3, characterized in that, Determining the offset distance of the echo data to be determined based on the echo data to be determined and the valid array includes: For each valid echo distance in the valid array, a corresponding valid echo vector is constructed based on the valid echo distance and its original index position, wherein the original index position is the index position of the valid echo distance in the currently maintained sliding window queue; Determine the covariance matrix and mean vector based on each of the effective echo vectors described; Based on the echo distance and expected index position of the echo data to be determined, generate the echo vector to be determined of the echo data to be determined; The offset distance of the echo data to be determined is obtained by calculating the distance based on the echo vector to be determined, the covariance matrix, and the mean vector.
8. The method according to claim 3, characterized in that, The step of determining the predicted distance of the echo data to be determined based on the echo data to be determined and the valid array includes: The pre-built fitting model is solved for parameters based on the effective echo distances in the effective array and the corresponding original index positions to obtain model parameters, and the current fitting model is obtained based on the model parameters. Distance prediction is performed based on the current fitted model and the expected index position corresponding to the echo data to be determined, and the predicted distance is obtained.
9. The method according to claim 1, characterized in that, The multidimensional features include: effective mean, first coefficient of variation, offset distance, and relative error; determining the validity of the echo data to be judged based on the multidimensional features includes: When the multidimensional features satisfy the first condition and / or the second condition, the validity of the echo data to be determined is invalid. When the multidimensional feature does not meet the first condition and does not meet the second condition, the validity of the echo data to be determined is determined to be valid; The first condition includes: the first coefficient of variation is less than the first coefficient threshold, the relative error is within the error range, and the echo distance of the echo data to be determined is not less than the effective mean. The second condition includes: the offset distance is greater than the offset distance threshold, the relative error is greater than the error threshold, and the echo distance of the echo data to be determined is not less than the effective mean or the first coefficient of variation is less than the second coefficient threshold.
10. The method according to any one of claims 1-9, characterized in that, Also includes: Determine whether the length of the currently maintained sliding window queue is greater than or equal to the set window length threshold. If so, remove the first element of the currently maintained sliding window queue, add the echo data to be determined to the tail of the currently maintained sliding window queue, and save the validity of the echo data to be determined accordingly. Otherwise, the echo data to be determined is added to the tail of the currently maintained sliding window queue, and the validity of the echo data to be determined is saved accordingly.
11. An electronic device, characterized in that, The electronic device includes: At least one processor, and a memory communicatively connected to said at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the method for determining the validity of echo data according to any one of claims 1-10.
12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the method for determining the validity of echo data according to any one of claims 1-10.
13. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method for determining the validity of echo data according to any one of claims 1-10.
