Vehicle blind spot detection method and system based on ultrasonic sensors

Abstract

The application relates to a kind of vehicle blind spot detection method and system based on ultrasonic sensor, it relates to vehicle driving environment perception technical field, it includes the detection range of ultrasonic sensor in the periphery of vehicle is divided into several windows according to distance, obtains the echo data of the ultrasonic sensor;The echo data is preprocessed to obtain feature information, and the type of obstacle corresponding to the echo data is distinguished based on the feature information;The feature point of the obstacle in the window is calculated;According to the current motion characteristics of vehicle, the window and position of the feature point after vehicle movement are predicted;Based on the window and position of the feature point after vehicle movement, the object position around the vehicle body is identified.The application has the effect that vehicle blind spot detection function is realized only through ultrasonic sensor.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of vehicle driving environment perception technology, and in particular to a vehicle blind spot detection method and system based on an ultrasonic sensor. BACKGROUND

[0002] The blind spot detection function of a passenger vehicle relies on a vehicle-mounted perception system, which needs to locate obstacles around the vehicle body. Distance perception sensors commonly used in the field of autonomous driving include ultrasonic radar, millimeter wave radar and laser radar.

[0003] For the blind spot detection function, laser radar brings high cost, so the mainstream solution uses two rear millimeter wave radars installed at the corners of the rear bumper to monitor vehicles in adjacent lanes and rear areas. When a moving object is detected in the warning area, a warning will be issued. For ultrasonic radar, due to the limited detection distance of ultrasonic waves, and when the vehicle is running at high speed, the ultrasonic sensor will be severely disturbed. If the echo data of the ultrasonic sensor is directly used to triangulate the obstacle, the position of the identified obstacle will be distorted. At the same time, since the ultrasonic sensor positioning has a certain delay, it is impossible to achieve the blind spot detection function by accurately positioning the obstacles around the vehicle through the ultrasonic sensor. SUMMARY

[0004] To solve the above technical problems, the present application provides a vehicle blind spot detection method and system based on an ultrasonic sensor, which can accurately locate the obstacles around the vehicle through the ultrasonic sensor to achieve the blind spot detection function.

[0005] In a first aspect, the present application provides a vehicle blind spot detection method based on an ultrasonic sensor, which adopts the following technical solution:

[0006] A vehicle blind spot detection method based on an ultrasonic sensor is applied to a vehicle, the vehicle is provided with a plurality of ultrasonic sensors on the side, the plurality of ultrasonic sensors form a continuous detection range, comprising:

[0007] The detection range on the side of the vehicle is divided into a plurality of windows according to the distance, and the echo data of the ultrasonic sensor is obtained;

[0008] The echo data is preprocessed to obtain feature information, and the type of the obstacle corresponding to the echo data is distinguished based on the feature information;

[0009] The feature points of the obstacle in the window are calculated;

[0010] According to the current motion characteristics of the vehicle, the window and position of the feature points after the vehicle moves are predicted;

[0011] The object position around the vehicle body is identified based on the window and the position of the feature point after the vehicle moves.

[0012] Further, in the vehicle blind spot detection method based on the ultrasonic sensor, the pre-processing of the echo data to obtain feature information and the distinguishing of the obstacle type corresponding to the echo data based on the feature information include:

[0013] The effective echo number and echo data variance in the detection range are counted based on the echo data.

[0014] The echo data is determined to belong to noise or an actual object based on the effective echo number.

[0015] The actual object corresponding to the echo data is determined to belong to a stationary object or a moving object based on the echo data variance.

[0016] Further, in the vehicle blind spot detection method based on the ultrasonic sensor, the calculation of the feature point of the obstacle in the window includes:

[0017] The feature point is stored in each window, and the attributes of the feature point include the average measurement value and the covariance of the echo used to generate the feature point.

[0018] The feature point is updated and determined in the window based on the measurement value of the ultrasonic sensor and the covariance corresponding to the measurement value.

[0019] Further, in the vehicle blind spot detection method based on the ultrasonic sensor, the number of feature points stored in each window does not exceed four.

[0020] Further, in the vehicle blind spot detection method based on the ultrasonic sensor, the updating and determination of the feature point in the window based on the measurement value of the ultrasonic sensor and the covariance corresponding to the measurement value include:

[0021] When the feature point is located in a certain window, the feature point is adjusted based on the measurement value of the ultrasonic sensor in the window according to a preset adjustment method until the feature point leaves the window; wherein the adjustment method includes:

[0022] If the feature point is still in the window and the newly transmitted measurement value is greater than the feature point, the attributes of the feature point remain unchanged.

[0023] If the feature point is still in the window and the newly transmitted measurement value is less than the feature point, the new measurement value and its covariance are assigned to the window and the attributes of the feature point are updated.

[0024] Furthermore, in the above-mentioned vehicle blind spot detection method based on ultrasonic sensors, the step of identifying the position of objects around the vehicle body based on the window and position of the feature point after the vehicle moves includes:

[0025] While the vehicle is in motion, the positions of feature points detected in the front window of the vehicle are acquired;

[0026] Based on the location of the feature points, the predicted location of the feature points after the vehicle moves is obtained according to a preset inference algorithm.

[0027] The predicted position of the feature point is compared with the position of the feature point detected in the rear window, and a judgment is made on whether the objects around the vehicle body have moved based on the comparison result.

[0028] Furthermore, in the above-mentioned vehicle blind spot detection method based on ultrasonic sensors, the inference algorithm includes:

[0029] Fuzzy reasoning is used to quantify statements representing various judgment results into quantized values;

[0030] Define the initial state of the measured values;

[0031] The initial state and the preset state transition matrix are input into the Markov chain to obtain the inference result;

[0032] The predicted location of the feature point is obtained by comparing the inference result with the quantization value.

[0033] Furthermore, in the above-mentioned vehicle blind spot detection method based on ultrasonic sensors, the method further includes:

[0034] Based on the object's location, a probability threshold, and the driver's driving intentions, an alarm is triggered.

[0035] Secondly, the present invention provides a vehicle blind spot detection system based on an ultrasonic sensor, employing the following technical solution:

[0036] A vehicle blind spot detection system based on ultrasonic sensors includes;

[0037] The data acquisition module includes multiple ultrasonic sensors mounted on both sides of the vehicle, used to generate ultrasonic waves and receive echo data reflected by obstacles;

[0038] The data preprocessing module is used to receive the echo data and preprocess the echo data;

[0039] The feature point prediction module is used to divide the window and calculate the feature points within the serial port;

[0040] The position reasoning module is used to reason about the position of objects around the vehicle.

[0041] Further, the vehicle blind spot detection system based on the ultrasonic sensor further comprises:

[0042] The warning module is signal connected to the car system to obtain the driving intention of the driver, and simultaneously obtains the position of the object around the vehicle body, and is used for warning and prompting based on the object position and the driving intention.

[0043] In a third aspect, the present application provides a vehicle, which adopts the following technical scheme:

[0044] The vehicle comprises the vehicle blind spot detection system based on the ultrasonic sensor.

[0045] In a fourth aspect, the present application provides a computer readable storage medium, which adopts the following technical scheme:

[0046] The computer readable storage medium stores the logic language corresponding to the vehicle blind spot detection method based on the ultrasonic sensor.

[0047] In summary, the present application has at least one beneficial technical effect:

[0048] 1. The present application can realize the blind spot detection function of the vehicle at high speed by using only the ultrasonic sensor, and the effective information in the ultrasonic sensor at high speed is extracted by an algorithm, the detection range is divided into a plurality of windows according to the distance, and the echo data in the window is used to realize the blind spot detection function; in addition, the algorithm can be optimized from the angle, which is convenient for application and arrangement in the embedded system.

[0049] 2. The present application is a method for realizing the blind spot detection function of the vehicle at high speed by using the ultrasonic sensor, which does not need to depend on the millimeter wave sensor, so that the low-configured vehicle (the vehicle only installed with the ultrasonic sensor) has the same blind spot detection function. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 is a flow chart of one embodiment of the vehicle blind spot detection method based on the ultrasonic sensor.

[0051] Figure 2 is a flow chart of another embodiment of the vehicle blind spot detection method based on the ultrasonic sensor.

[0052] Figure 3 is a judgment standard diagram of the signal-to-noise ratio (SNR) of the echo data in the vehicle blind spot detection method based on the ultrasonic sensor.

[0053] Figure 4is a flow chart of another embodiment of the vehicle blind spot detection method based on the ultrasonic sensor.

[0054] Figure 5 is an example diagram of the moving window division in the vehicle blind spot detection method based on the ultrasonic sensor.

[0055] Figure 6 is an example diagram of the feature point storage in the vehicle blind spot detection method based on the ultrasonic sensor.

[0056] Figure 7 is a flow chart of another embodiment of the vehicle blind spot detection method based on the ultrasonic sensor.

[0057] Figure 8 is an example diagram of the Markov chain in the vehicle blind spot detection method based on the ultrasonic sensor.

[0058] Figure 9 is a topological diagram of an embodiment of the vehicle blind spot detection system based on the ultrasonic sensor.

[0059] Legend: 1, data acquisition module; 2, data preprocessing module; 3, feature point prediction module; 4, position inference module; 5, warning module. DETAILED DESCRIPTION

[0060] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0061] The execution order of the method steps described in the embodiments of the present application can be executed in the order described in the specific implementation manner, or can be adjusted according to actual needs on the premise of solving the technical problems. The execution order of each step is not listed here.

[0062] The following will be described in detail with reference to the drawings in the embodiments of the present application. Figures 1-3 The present application will be further described in detail.

[0063] The embodiments of the present application disclose a vehicle blind spot detection method based on an ultrasonic sensor, referring to Figure 1 The vehicle blind spot detection method based on the ultrasonic sensor comprises,

[0064] S1, dividing the detection range of the vehicle into several windows according to the distance, and obtaining echo data of the ultrasonic sensor;

[0065] S2, pre-processing the echo data to obtain feature information, and distinguishing the type of the obstacle corresponding to the echo data based on the feature information;

[0066] S3, calculating the feature point of the obstacle in the window;

[0067] S4, predicting the window and position of the feature point after the movement of the vehicle according to the current movement characteristics of the vehicle;

[0068] S5, identifying the position of the object around the vehicle body based on the window and position of the feature point after the movement of the vehicle.

[0069] In the vehicle of the present application, a plurality of ultrasonic sensors are arranged on both sides of the vehicle. Specifically, in the present embodiment, a total of 8 ultrasonic sensors are arranged on both sides of the vehicle, which are respectively located at the front left corner, the front right corner, the front left side, the front right side, the rear left side, the rear right side, the rear left corner and the rear right corner of the vehicle body. At the same time, the vehicle needs to provide the vehicle speed, the wheel speed and the turning angle information for algorithm calculation.

[0070] Firstly, the echo data received by the ultrasonic sensors on both sides of the vehicle body is obtained. Then, the feature extraction is performed on the echo data, and the type of the obstacle corresponding to the echo data is distinguished based on the extracted features. Then, the area on the side of the vehicle body is divided into several windows, and the feature points are stored in each window. Then, the feature points are updated based on the current movement characteristics of the vehicle. Finally, the positions of the objects around the vehicle body are identified based on the feature points after the movement of the vehicle.

[0071] In order to detect the objects in the blind area on both sides of the vehicle body, especially the moving objects, the present application does not need to install expensive millimeter wave radar or laser radar on both sides of the vehicle body, but only uses ultrasonic sensors to realize the blind area detection function when the vehicle is running at high speed.

[0072] Further, as one embodiment of the present application, referring to Figure 2 , step S2, pre-processing the echo data to obtain feature information, and distinguishing the type of the obstacle corresponding to the echo data based on the feature information, comprising:

[0073] S21, counting the number of effective echoes and the variance of the echo data in the detection range of the ultrasonic sensor based on the echo data;

[0074] S22, determining whether the echo data belongs to noise or actual object based on the number of effective echoes;

[0075] S23, determining whether the echo data belongs to a stationary object or a moving object based on the variance of the echo data.

[0076] The effective echo refers to the ultrasonic echo with a distance greater than 0 and less than the maximum detection distance of the ultrasonic sensor. Meanwhile, in order to facilitate subsequent calculation, the invalid echo beyond the maximum detection distance of the ultrasonic sensor is assigned as 0.

[0077] The step S21 specifically comprises: distinguishing whether the object detected by the ultrasonic sensor is noise or a stationary object or a moving object on the highway by counting the number of effective echoes detected by the ultrasonic sensor and the variance of the echo data.

[0078] Specifically, the step S22, determining whether the echo data belongs to noise or an actual object based on the number of effective echoes comprises: determining the signal-to-noise ratio based on the number of effective echoes, and distinguishing the noise and the actual object according to the determined signal-to-noise ratio. Refer to Figure 3 The signal-to-noise ratio SNR generated by the noise is obviously lower than the signal-to-noise ratio SNR calculated by the actual object. In order to facilitate actual calculation, the threshold between the signal-to-noise ratio SNR generated by the noise and the signal-to-noise ratio SNR generated by the actual object is set to 0.67 in the embodiment, that is, the signal-to-noise ratio SNR is considered to be non-noise when it is greater than 0.67, and the confidence of the measurement data is 0 / 1.

[0079] Further, the step S23, determining whether the echo data belongs to a stationary object or a moving object based on the variance of the echo data. The occurrence rate of low noise (low covariance) measurement generated by the moving object is obviously higher than that generated by the stationary object or random noise, so the variance of the echo data can be used to determine whether the echo data belongs to a stationary object or a moving object.

[0080] Further, as an embodiment of the present application, refer to Figure 4 and Figure 5 The step S3, calculating the feature points of the obstacle in the window, comprises:

[0081] S31, storing the feature points in each window, and the attributes contained in the feature points include the average measurement value and the covariance of the echo used to generate the point. Specifically, refer to Figure 5 The area of the side of the vehicle is divided into a plurality of windows, and the feature points in the window are calculated. In order to reduce the calculation amount, at most four feature points are stored in each window, and the attributes contained in the feature points are the average measurement value and the covariance of the echo used to generate the point.

[0082] S32, updating and determining the feature points based on the measurement value of the ultrasonic sensor and the covariance corresponding to the measurement value.

[0083] Further, as an embodiment of the present application, the step S32 is specifically:

[0084] When the feature point is located in a certain window, the feature point is adjusted based on the measurement value of the ultrasonic sensor in the window according to a preset adjustment method until the feature point leaves the window, and the adjustment method includes:

[0085] If the feature point is still in the window and the newly incoming measurement value is greater than the feature point, the attribute of the feature point remains unchanged;

[0086] If the feature point is still in the window and the newly incoming measurement value is less than the feature point, the new measurement value and its covariance are assigned to the storage window and the attribute of the feature point is updated.

[0087] Referring to Figure 6 , as an example, as long as the feature point (the circular point) is located in the first window, it will be adjusted based on the new measurement value (the square point). If the feature point (the circular point) is still in the first window and the newly incoming measurement value (the square point) is smaller, the new measurement value (the square point) and its covariance are assigned to the storage window and the feature is updated. Only when the feature point (the circular point) leaves the first window, the parameters of the feature point (the circular point) are fixed. Finally, as the vehicle moves, the feature point with the smallest distance is extracted from each window.

[0088] Further, as an embodiment of the present application, the step S4 is specifically: predicting the window and position of the feature point after the movement of the vehicle according to the current motion characteristics of the vehicle, including the current speed and the steering angle.

[0089] Further, as an embodiment of the present application, referring to Figure 7 , the step S5 is specifically: identifying the position of the object around the vehicle body based on the window and position of the feature point after the movement of the vehicle, including:

[0090] S51, obtaining the position of the feature point detected in the front window of the vehicle when the vehicle moves;

[0091] S52, obtaining the predicted position of the feature point after the movement of the vehicle according to a preset inference algorithm based on the position of the feature point;

[0092] S53, comparing the predicted position of the feature point with the position of the feature point detected in the rear window, and making a judgment on whether the object around the vehicle body moves based on the comparison result.

[0093] Step S51 acquires the position of the feature point in real time when the vehicle is moving, and after the vehicle has run for a period of time, step S52 estimates the predicted position of the feature point after the vehicle moves according to a preset inference algorithm. Finally, in step S53, by comparing the feature point calculated by the rear sensor of the ego vehicle in real time and the predicted position of the feature point after the front feature point moves with the vehicle, it can be inferred whether the moving object is around the vehicle body. If the deviation between the feature point detected by the rear sensor and the predicted position of the front feature point is not large, it can be considered that it is only a stationary object, and no blind spot warning is needed. If the deviation between the feature point detected by the rear sensor and the predicted position of the front feature point is large, it means that there is a moving object around the vehicle, and the driver needs to be prompted.

[0094] Further, as an embodiment of the present application, the inference algorithm comprises: quantizing sentences representing various judgment results into quantized values through fuzzy inference; defining an initial state of a measurement value; inputting the initial state and a preset state transition matrix into a Markov chain to obtain an inference result; and comparing the inference result with the quantized values to obtain a predicted position of the feature point.

[0095] In the embodiment, quantizing sentences representing various judgment results into quantized values through fuzzy inference is specifically:

[0096] 1. A(x): This value indicates that the signal-to-noise ratio of the rear sensor x determines that the measurement data contains non-noise;

[0097] 2. B_n: This value indicates that the vertical distance between the measured object and the nth expected object feature point is very "small";

[0098] 3. C_n: This value indicates that the Mahalanobis distance between the measured object and the nth expected object feature point is very "small";

[0099] and other quantized values calculated according to A(x), B_n, C_n,

[0100] 4. I1_n: I1_n = B_n||C_n, this value indicates that the nth measurement value is reliable;

[0101] 5. I2_n: I2_n = I1_n&&A(x), this value indicates that by comparing the measurement value and the predicted value, it is shown that the nth measurement value comes from a stationary object;

[0102] 6. I3(x): I3(x) = I2_1||I2_2...||I2_n, this value indicates that the measurement of the rear sensor x is effective;

[0103] 7. I4(x): I4(x) = (1-I3(x))&&A(x), this value indicates that the measurement value of the rear sensor x comes from a moving object;

[0104] 8, I5(x): I5(x) = 1-A(x), which indicates that the measurement of the rear sensor x is random noise.

[0105] wherein the definition of "small" is that the measurement-prediction is negative, and equals 1 if B_n / C_n is true; otherwise equals 0.

[0106] Further, referring to Figure 8 , the initial state of the measurement is defined, and the initial state and the preset state transition matrix are input into the Markov chain to obtain the inference result, specifically:

[0107] Three initial states S1, S2 and S3 in the Markov chain are defined. S1 = max(I5(a), I5(b)) indicates that the measurement is derived from random noise. S2 = max(I3(a), I3(b)) indicates that the measurement is derived from a stationary object. S3 = max(I4(a), I4(b)) indicates that the measurement is derived from a moving object. Wherein a and b represent different ultrasonic sensors at the rear, the rear / rear corner ultrasonic sensors apply the above quantified fuzzy statements to the Markov chain.

[0108] The state transition matrix P in the Markov chain is defined as T :

[0109] ,

[0110] Wherein, each weight coefficient a11 to a33 is calibrated by a calibration personnel in advance.

[0111] Finally, by setting the initial state in the Markov chain and the state transition matrix P T , the position and probability of whether there is a moving object around the vehicle can be finally obtained.

[0112] Further, as an embodiment of the present application, the vehicle blind spot detection method based on ultrasonic sensors further comprises:

[0113] Based on the object position, the probability threshold and the driving intention of the driver are matched to give an alarm prompt. Wherein, the driving intention is obtained through the car system, and the car system judges based on the signals of the steering wheel and brake signal sensors. Increasing the alarm prompt is beneficial to timely reminding the driver when there is an object, especially a moving object, at the blind spot of the vehicle body, thereby improving the driving safety.

[0114] The implementation principle of an embodiment of the vehicle blind spot detection method based on ultrasonic sensors of the present application is:

[0115] Firstly, data preprocessing is performed, and the effective echo detected by the ultrasonic sensor is counted;

[0116] Secondly, feature information is calculated based on the effective echo, and the detected obstacle is distinguished from noise or actual object by counting the variance of the echo data and the number of effective echoes. The actual object is a stationary object or a moving object.

[0117] Then, the vehicle side area is divided into moving windows, and the feature points of each window are calculated.

[0118] Next, according to the current motion characteristics of the vehicle, including the current speed and the steering angle, the window and the position of the feature point after the vehicle moves are predicted.

[0119] Finally, a fuzzy Markov chain is used to calculate the probability of the existence of a moving obstacle around the vehicle. By comparing the feature points calculated by the rear sensor of the vehicle in real time with the predicted values of the feature points in front of the vehicle after the vehicle moves, it can be inferred whether the moving object is around the vehicle.

[0120] Based on the above ultrasonic sensor-based vehicle blind spot detection method, the embodiment of the present application also discloses an ultrasonic sensor-based blind spot detection system.

[0121] Referring to Figure 9 An ultrasonic sensor-based blind spot detection system includes a data acquisition module 1, a data preprocessing module 2, a feature point prediction module 3, and a position reasoning module 4.

[0122] As an embodiment of the present application, the data acquisition module 1 includes a plurality of ultrasonic sensors arranged on both sides of the vehicle for emitting ultrasonic waves and receiving echo data reflected by obstacles. As an embodiment, the number of ultrasonic sensors in the present embodiment can be set to 8, which are respectively located at the front left corner, front right corner, front left side, front right side, rear left side, rear right side, rear left corner and rear right corner of the vehicle body.

[0123] At the same time, in order to further facilitate the data processing of the subsequent modules, the data acquisition module 1 is also connected to the vehicle infotainment system for acquiring the vehicle speed, tire speed and steering angle information of the vehicle.

[0124] Further, as an embodiment of the present application, the data preprocessing module 2 is connected to the data acquisition module 1 for receiving the echo data acquired by the data acquisition module 1 and preprocessing the echo data to obtain the variance of the echo data and the signal-to-noise ratio.

[0125] Further, as an embodiment of the present application, the feature point prediction module 3 is data-connected to the data preprocessing module 2 and the data acquisition module 1, for dividing windows and calculating feature points in the windows. Specifically, the feature point prediction module 3 divides the blind spot area on both sides of the vehicle body into multiple windows, and calculates the feature points of each window and the attributes of the feature points based on the data preprocessing module 2. Meanwhile, the feature point prediction module 3 obtains the current motion characteristics of the vehicle through the data acquisition module 1, including the current speed and the steering angle, and predicts the window and position of the feature points after the vehicle moves.

[0126] Further, as an embodiment of the present application, the position inference module 4 is data-connected to the feature point prediction module 3, for inferring the position of the object around the vehicle body. The position inference module 4 is built-in with a fuzzy inference algorithm to calculate the probability of the existence of a moving obstacle around the vehicle. By comparing the feature points calculated by the rear sensor of the vehicle in real time and the predicted values of the front feature points after the vehicle moves, the position inference module 4 can infer whether a moving object is around the vehicle body. If the deviation between the feature points detected by the rear sensor and the predicted position of the front feature points is not large, it can be considered that it is only a stationary object, and no blind spot warning is needed. If the deviation between the feature points detected by the rear sensor and the predicted position of the front feature points is large, it means that there is a moving object around the vehicle, and the driver needs to be prompted.

[0127] Further, as an embodiment of the present application, the blind spot detection system based on the ultrasonic sensor further comprises a warning module 5.

[0128] Further, as an embodiment of the present application, the warning module 5 is signal-connected to the vehicle system, for obtaining the driving intention of the driver, and obtaining the position of the object around the vehicle body, for warning and prompting based on the position of the object and the driving intention. The warning module 5 is beneficial to timely remind the driver when there is an object, especially a moving object, in the blind spot of the vehicle body, thereby improving the driving safety.

[0129] Based on the above-mentioned blind spot detection system based on the ultrasonic sensor, the present application further discloses a vehicle.

[0130] The vehicle comprises the blind spot detection system according to any one of the above-mentioned embodiments. The vehicle can realize the blind spot detection function at high speed by using only the ultrasonic sensor, and the millimeter wave radar is omitted. The effective information in the ultrasonic sensor during high-speed driving is extracted by the algorithm, and whether there is a moving object around the vehicle is judged by using the fuzzy Markov chain, for realizing the blind spot detection function.

[0131] Further, the present application further provides a computer readable storage medium,

[0132] A computer readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the ultrasonic sensor-based vehicle blind spot detection method of any one of the preceding embodiments.

[0133] The computer readable storage medium can include any entity or device capable of storing the computer program, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution medium, etc. The computer program includes computer program code. The computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer readable storage medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution medium, etc.

[0134] Any process or method descriptions or any other descriptions in flow diagrams or otherwise described herein can be understood as representing one or more steps of a method or process, including a set of executable instructions executed by one or more computers and / or any other suitable means for performing the steps of the method or process, and that the various embodiments of the present application can include additional or fewer steps performing the same or equivalent functions.

[0135] The logic and / or steps represented in flow diagrams or otherwise described herein, for example, can be considered as a set of steps to be executed by one or more computers and / or any other suitable means for performing the steps of the method or process, and that the various embodiments of the present application can include additional or fewer steps performing the same or equivalent functions.

[0136] The above embodiments are only used to illustrate the technical solutions of the present application, not limit them; although the above embodiments of the present application are described in detail, those skilled in the art should understand that the technical solutions recorded in the above embodiments can be modified, or some technical features can be replaced by equivalent ones; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An ultrasonic sensor-based blind spot detection method for a vehicle, the vehicle having a plurality of ultrasonic sensors arranged on the side of the vehicle, the plurality of ultrasonic sensors forming a continuous detection range, the method comprising: The method comprises: ​ dividing the detection range around the vehicle into a plurality of windows according to distance, and obtaining echo data of the ultrasonic sensor; preprocessing the echo data to obtain feature information, and distinguishing the type of the obstacle corresponding to the echo data based on the feature information; calculating feature points of the obstacle in the window, including: storing feature points in each window, the attributes of the feature points including average measurement value and covariance of the echo used to generate the feature point; updating and determining the feature points in the window based on the measurement value of the ultrasonic sensor and the covariance corresponding to the measurement value; predicting the window and position of the feature point after the movement of the vehicle according to the current movement characteristics of the vehicle; identifying the position of the object around the vehicle body based on the window and position of the feature point after the movement of the vehicle, including: obtaining the position of the feature point detected in the front window of the vehicle when the vehicle is moving; obtaining the predicted position of the feature point after the movement of the vehicle based on the position of the feature point according to a preset inference algorithm; comparing the predicted position of the feature point with the position of the feature point detected in the rear window, and making a judgment on whether the object around the vehicle body moves based on the comparison result.

2. The ultrasonic sensor-based vehicle blind spot detection method of claim 1, wherein, The preprocessing of the echo data to obtain feature information, and the distinguishing of the type of the obstacle corresponding to the echo data based on the feature information, includes: counting the number of effective echoes and echo data variance in the detection range based on the echo data; determining whether the echo data belongs to noise or actual object based on the number of effective echoes; determining whether the actual object corresponding to the echo data belongs to stationary object or moving object based on the echo data variance.

3. The ultrasonic sensor-based vehicle blind spot detection method of claim 1, wherein, The number of feature points stored in each window is not more than four.

4. The ultrasonic sensor-based vehicle blind spot detection method of claim 1, wherein, The updating and determination of the feature points in the window based on the measurement value of the ultrasonic sensor and the covariance corresponding to the measurement value, includes: when the feature point is located in a window, adjusting the feature point based on the measurement value of the ultrasonic sensor in the window according to a preset adjustment method until the feature point leaves the window; wherein the adjustment method includes: if the feature point is still in the window and the newly transmitted measurement value is greater than the feature point, the attributes of the feature point remain unchanged; if the feature point is still in the window and the newly transmitted measurement value is less than the feature point, the new measurement value and its covariance are assigned to the window and the attributes of the feature point are updated.

5. The ultrasonic sensor-based vehicle blind spot detection method of claim 1, wherein, The inference algorithm includes: quantifying the statements representing various judgment results into quantized values through fuzzy inference algorithm; defining the initial state of the measurement value; inputting the initial state and the preset state transition matrix into Markov chain to obtain inference result; comparing the inference result with the quantized value to obtain the predicted position of the feature point.

6. The ultrasonic sensor-based vehicle blind spot detection method of claim 1, wherein, The method further comprises: based on the object position, combining a probability threshold and the driving intention of the driver to give an alarm prompt.

7. An ultrasonic sensor based vehicle blind spot detection system, applied to the method according to any one of claims 1 to 6, characterized in that, The method comprises: a data acquisition module (1) comprising a plurality of ultrasonic sensors arranged on both sides of the vehicle for emitting ultrasonic waves and receiving echo data reflected by obstacles; a data preprocessing module (2) for receiving and preprocessing the echo data; a feature point prediction module (3) for dividing windows and calculating feature points within the serial ports; a position inference module (4) for inferring the positions of objects around the vehicle body.

8. The ultrasonic sensor-based vehicle blind spot detection system of claim 7, wherein, The system further comprises: a warning module (5) connected to the vehicle system to obtain the driving intention of the driver and the positions of objects around the vehicle body, and to give a warning based on the positions of the objects and the driving intention.

9. A vehicle characterized by comprising: The vehicle is provided with the system according to any one of claims 7-8.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a logic language corresponding to the method according to any one of claims 1-6.

Images