TWI541152B - Traffic safety system and its obstacle screening method - Google Patents

Description

Driving safety system and obstacle screening method thereof

The invention relates to a driving system, in particular to a driving safety system and an obstacle screening method thereof.

With the development of technology, the development of smart driving system has gradually become the trend of the smart car market. The current smart driving system mainly collects data such as speed and front distance, and uses the neural network to analyze the driver's driving habits to find out Driving safety distance and speed, through this technology can set different safety distances for different speeds, so the stability of the driving system and the occurrence of failure are closely related, the driving system is more susceptible to too complicated environment or noise, resulting in Errors are detected when detecting a target vehicle in front (especially in a turn), thus causing system failure and reducing system performance.

With the current development technology of the driving system, such as Volvo's City Safety system, it can cruise in the city between 15-30km/h. When the obstacle is found in front, the City Safety system will brake the car and reduce the speed. It can be completely driven under 15 km. Another example is the SCBS system of the Mazda car, which can use a laser sensor to capture the obstacle data in front and can detect the distance from the obstacle at low speed (4-30km/h) to determine whether to brake. Control, slow down in time and reduce collisions. Another example is the Ford car's front anti-collision system, which is one of the indispensable safety systems for the Ford car. The speed is between 5-30km/h. When the system detects that the front distance is below the warning value, it starts. Independent braking, applicable to road sections with a radius of curvature greater than 20m.

For example, China's invention patent No. I318604 "Applying recursive least squares method for estimating the collision time" (hereinafter referred to as the previous case), the main purpose is to improve the accuracy of calculating the collision time, which is used in the vehicle. a distance sensor for measuring a relative distance between the host vehicle and an alien vehicle or an obstacle, and using an estimation unit provided on the vehicle, the estimation unit for reading the distance perception Measuring the relative distance of the device, and the estimating unit estimates the quadratic curve of the relative distance according to the relative distance of the measurement, and brings the most known coefficients into the quadratic curve equation according to a Recursive Least Square method. And the calculation formula of the recursive least-square method is used to obtain the time point at which the calculated relative distance is zero, and the collision time and time difference are estimated, so that the noise influence can be reduced to prevent the collision from occurring in the case of external noise interference. .

It can be seen from the above that the existing R&D technical system has many restrictions, and although it can slow down and reduce the collision opportunity, it still has the following disadvantages: 1. The traveling speed of the vehicle needs to be less than 30km/h, and only the City Safety part is applicable. Scenes. 2. It is impossible to accurately judge the timing of the brakes or provide warnings in advance so that the driving speed must be lowered to avoid the brakes being too late. 3. The prior art lacks a prediction mechanism, and the lack of accuracy leads to system instability, which is likely to cause an accident when the system fails frequently. The previous case is to improve the accuracy of the driving system to determine the obstacles. It is to perform complex calculations to estimate the collision time and time difference to reduce the influence of noise and prevent collisions. However, it takes not only a lot of time and In addition to computing resources, the filtered objects are all floating in the relative distance between the external vehicle or the obstacle, and the previous case does not judge the presence of the front target. It is directly assumed that the target in front is true. The existence of the entity, so if the application of the previous case is still insufficient in the actual road conditions, therefore, the existing smart driving system technology still has problems such as system instability, insufficient accuracy and complicated methods, which raises the cost, so it is indeed proposed. The necessity of a good program.

In view of the above problems of the prior art, the main object of the present invention is to provide a driving safety system and an obstacle screening method thereof, which are driven by a vehicle equipped with a driving safety system so that the vehicle is not subjected to traveling speed and scene during driving. Restrictions can instantly and accurately determine non-obstructions, environmental reflections, ground objects and other noises, and eliminate them, increasing the accuracy and stability of the vehicle safety system to determine obstacles, in order to improve vehicle safety.

The main technical means adopted to achieve the above objectives is to enable the obstacle screening method of the above-mentioned driving safety system to be mainly connected to an image capturing module and a distance measuring module by a vehicle computer, and the vehicle computer is used. Performing the following steps: receiving more than one obstacle information and corresponding one or more image information; performing a filtering analysis mechanism on the obstacle information to filter the noise, and generating a position information of the obstacle after filtering; The position information of the obstacle is compared with the image information to exclude a ground noise and obtain a front target information.

In the above method, the vehicle computer, the image capturing module, and the distance measuring module are mounted on a vehicle, and the vehicle computer receives the obstacle information through the distance measuring module, and the image capturing The module receives the image information corresponding to the obstacle information, and performs the filtering analysis mechanism on the obstacle information to filter out the reflection noise caused by the environment to generate the position information of the suspected obstacle, and the vehicle computer uses the position information to The image information is compared to eliminate noise generated by the ground, so as to accurately eliminate all non-obstacle information and instantly obtain a target information in front, so that the driving system can predict the driving condition and accurately predict the non-obstacle, Noise such as environmental reflections, ground objects, etc., and eliminate them, increase the accuracy and stability of the vehicle safety system to determine the obstacles, and achieve the purpose of improving driving safety.

Another major technical means adopted to achieve the above objectives is that the driving safety system includes: a ranging module that captures more than one obstacle information; and an image capturing module that captures more than one image. Information relating to the obstacle information; a vehicle computer connected to the distance measuring module and the image capturing module, and receiving the obstacle information and the image information; Performing a filtering analysis mechanism to filter the reflection noise from the non-obstacle, and generating a position information of the obstacle after filtering, the vehicle computer compares the position information of the obstacle with the image information to exclude one The noise generated by the ground objects is used to obtain a target information in front.

According to the above configuration, the driving safety system of the present invention can be installed on a vehicle, and the obstacle measuring module receives the obstacle information, and the image capturing module receives the image information corresponding to the obstacle information. Perform a filtering analysis mechanism on the computer to filter out information or reflection noise from non-obstacle caused by the environment to generate position information of the suspected obstacle. The vehicle computer compares the position information with the image information. It can eliminate the noise generated by the ground objects and instantly obtain the information of the front target, so that the vehicle equipped with the driving system of the invention can predict the situation ahead of the road, and early warning and stabilize the system, and because the pre-determination is accurate Obstacle, environmental reflection, ground objects and other noise, and eliminate it, increase the accuracy and stability of the vehicle safety system to determine the obstacles, in order to achieve the purpose of improving driving safety.

For a system architecture of a preferred embodiment of the driving safety system of the present invention, please refer to FIG. 1 and FIG. 2, which includes a ranging module 10, an image capturing module 20, and a vehicle computer 30, and In the embodiment, the vehicle module 30, a vehicle state module 50, and an alarm module 60 are connected to the distance measuring module 10 and the image capturing module 20, respectively. The brake module 40, the vehicle state module 50, and the alarm module 60.

The ranging module 10 is configured to capture more than one obstacle information and transmit the obstacle information to the vehicle computer 30. The image capturing module 20 is configured to capture more than one image information. The image information system is associated with the obstacle information. The image capturing module 20 also transmits the captured image information to the vehicle computer 30, and provides the vehicle computer 30 for comparative analysis. In this embodiment, The distance measuring module 10 can be a millimeter wave radar. The image capturing module 20 can be a camera. The alarm module 60 includes a display unit and/or a sound unit, and provides instant information display through the display unit. Or the warning screen can also sound an alarm through the sound unit in case of an emergency.

The vehicle computer 30 performs a filtering analysis mechanism on the obstacle information and the image information according to the received obstacle information and the corresponding image information to filter the reflection noise from the non-obstacle and the surrounding environment, and After filtering, position information of an obstacle is generated, and the vehicle computer 30 compares the position information of the obstacle with the image information to exclude a noise generated by the ground object and thereby obtain a forward target information. When the driving system of the present invention is installed in a vehicle, the vehicle computer 30 can instantly determine the driving condition in front, and early warning and stabilize the system, and can accurately determine non-obstacle, environmental reflection, ground objects, and the like. Noise, and eliminate it, increase the accuracy and stability of the vehicle safety system to determine the obstacles, in order to achieve the purpose of improving driving safety.

Further, the brake module 40 receives a brake/de-oil control signal sent by the vehicle computer 30 according to the judgment analysis result of the vehicle computer 30 to drive the vehicle to decelerate; the vehicle state module 50 is used. In order to sense the current driving state of the vehicle and provide a vehicle status information to the vehicle computer 30, the vehicle computer 30 uses the driving status information for other analysis applications. In this embodiment, the driving status information includes a vehicle speed signal. And a corner speed signal, when the vehicle computer 30 respectively captures the image information captured by the image capture module 20, the obstacle information captured by the distance measuring module 10, and returns the vehicle according to the vehicle state module 50 The condition determines a vehicle route, and if the front target information has fallen into a warning range, the brake/de-oil control signal is immediately sent to the brake control module 40 to cause the vehicle to automatically decelerate.

When the vehicle computer 30 obtains driving state information through the vehicle state module 50, the distance measuring module 10 obtains obstacle information, and calculates the driving state information and the obstacle information to obtain an obstacle speed. When the obstacle speed exceeds a speed threshold, it is determined to be a dynamic obstacle, otherwise it is a static obstacle.

In this embodiment, the vehicle computer 30 is mainly composed of a microcontroller 31 connected to a sensing fusion processor 32, as shown in FIG. 2, wherein the microcontroller 31 is separately connected to the ranging module 10 and the brakes. The module 40, the vehicle state module 50, and the alarm module 60 are electrically connected to each other. The sensory fusion processor 32 is electrically connected to the image capturing module 20 and the alarm module 60. In this embodiment, the sensing fusion is performed. The processor 32 can be configured by a digital signal processing (DSP). The sensing fusion processor 32 can accelerate the real-time operation to obtain the front target information. The perceptual fusion processor 32 mainly transmits the obstacle information. The image information indicates the closest obstacle information in the route of the vehicle and determines whether the obstacle information is a static obstacle. If yes, a body identification can be performed according to the position of the obstacle information to determine whether the front target information is For the vehicle, a result of marking the warning pattern is displayed through the alarm module 60.

Regarding the above-mentioned application mode for obstacle screening of the driving safety system of the present invention, when the vehicle is in the course of the path, the vehicle computer 30 receives the obstacle information and phase received by the distance measuring module 10 and the image capturing module 20. Corresponding image information, but due to the influence of the environment, the mirror effect is caused, so that the image capturing module 20 misjudges the obstacles in the distance due to the signal reflection, and the noise is easy to drift, and in this embodiment, The filtering analysis mechanism further includes a first filtering noise method and a second filtering noise method to strengthen the filtering of the reflected noise from the non-obstacle and the surrounding environment in a dual filtering manner.

The first filtering noise method mainly collects all the received signals, as shown in FIG. 3, which includes a horizontal axis and a vertical axis, and the horizontal axis represents a time unit (second/s), and the vertical axis It represents the variation of all signals (Variance, V). Through the vehicle computer 30, the distance information of an obstacle and the estimated distance information are calculated, and the error value of the difference is calculated by a variation amount every other cycle. After the time (such as 6.4 seconds), the re-zeroing update is calculated again. Since the noise generated by the mirror effect is easy to float, it will lead to the characteristics of high variation, so it is judged whether the signal variation has a significant fluctuation oscillation state, and if so It is determined as a noise N1, if otherwise determined as a non-noise N2; in this embodiment, the variation calculation refers to a standard deviation (SD) calculation.

When the first filtering noise method filters out the noise generated by the specular effect, if there is a target information in front, and the signal is reflected back and forth between the target information and the vehicle, the shooting range is doubled to capture. To the signal, the second filtering noise method is used to eliminate the double harmonic noise caused by the back and forth reflection of the signal between the vehicles, as shown in FIG. 4, which includes a horizontal axis, a vertical axis, and a first obstacle. The object position D1 and a second obstacle position D2 representing a unit distance (meter/m) in the X-axis direction centering on the host vehicle S, and representing a unit in the Y-axis direction centering on the host vehicle S The distance (meter/m), when the vehicle computer 30 obtains the information of the first obstacle position D1, the second obstacle position D2, and the distance between the second obstacle position D2 and the first obstacle position D1 When a proportional condition is satisfied (for example, if the Y-axis distance of the second obstacle position D2 is twice the first obstacle position D1), it is determined whether the position information (X coordinate, Y coordinate) of the obstacle positions D1, D2 is Within a distance, if yes The second harmonic is twice the obstacle position D2 noise, Furthermore, the scope of the noise measurement position D1 by the first obstacle is calculated from the standard deviation distribution.

After the vehicle computer 30 executes the first and second filtering noise methods respectively, the position information of the obstacle is generated according to the first obstacle position D1, but the position information may be due to encountering a ground object. The position information generated after reflection (such as metal objects) is further compared by the vehicle computer 30 with the image information to perform a filtering ground noise method, as shown in FIG. 5, which includes a range R of the vehicle routing route, the host vehicle S and the first obstacle position D1, and the Y-axis distance captured by the ranging module 10 to the first obstacle position D1 is a first distance information (eg 9m), the Y-axis distance of the first obstacle position D1 captured by the image capturing module 20 is a second distance information (such as 15m), and the vehicle computer 30 determines the first and second distances. If the information is different, or the second distance information is greater than the first distance information by a set value or more, the D1 is determined to be ground noise, and the filtering ground noise method is performed, mainly by the ranging module 10 Performing the first obstacle position D1 The target is locked, and the obstacle position tracking prediction is performed to generate an estimated position information (X coordinate, Y coordinate), and the estimated position information is compared with the first obstacle position D1 to obtain an absolute value. If the absolute value is less than an error threshold, it is determined that the obstacle is a noise to filter the position information of the obstacle, thereby achieving the effect of continuously eliminating the noise generated by the ground object and accurately obtaining the front target information.

The description of the above embodiments of the present invention and the application manner thereof can further summarize the obstacle screening method of the line safety system, which is mainly connected to the image capturing module 20 and the distance measuring module by the vehicle computer 30. Group 10, as shown in FIG. 6, and the following steps are performed by the vehicle computer 30: receiving one or more obstacle information and corresponding one or more image information (S61); performing a filtering analysis mechanism on the obstacle information to Filtering the noise (S62), and generating a position information of the obstacle after filtering (S63); comparing the position information of the obstacle with the image information to exclude a ground noise (S64), and obtaining a In front of the target information (S65); in this embodiment, the vehicle computer 30 can obtain the driving status information and the obstacle information through the vehicle state module 50 and the distance measuring module 10, respectively, and the driving status information, The obstacle information is calculated to obtain an obstacle speed. When the obstacle speed exceeds a speed threshold, it is determined to be a dynamic obstacle, otherwise it is a static obstacle.

The vehicle computer 30, the image capturing module 20, and the distance measuring module 10 are mounted on a vehicle, and the vehicle computer 30 receives the obstacle information through the distance measuring module 10, and the image capturing module The group 20 receives the image information corresponding to the obstacle information, and performs the filtering analysis mechanism on the obstacle information to filter out the reflection noise caused by the environment to generate the position information of the suspected obstacle, when the above steps are performed to "the obstacle When the information information performs a filtering analysis mechanism to filter the noise (S62), as shown in FIG. 7, the first computer filtering method is further executed by the vehicle computer 30, and the method further comprises the following steps: calculating a barrier The distance information of the object and the estimated distance information (S71); the error value of the distance information and the estimated distance information is subjected to a variation calculation (S72); and whether the signal variation calculation has a significant fluctuation oscillation state ( S73); if yes, it is determined as a noise N1 (S74); if not, it is determined to be a non-noise (S75).

Further, when the above step is performed to the step of performing a filtering analysis mechanism to filter the noise (S62), as shown in FIG. 8, the second computer filtering method is further executed by the vehicle computer 30. The method further includes the following steps: obtaining information of a first obstacle position and a second obstacle position (S81); and when the distance between the second obstacle position and the first obstacle position meets a proportional condition (S82) Determining whether the position information of the position of the obstacle is within a distance range (S83); if so, the second obstacle position is twice the harmonic noise (S84); if not, the noise is not ( S85).

After performing the first and second filtering methods, the vehicle computer 30 compares the location information with the image information to further exclude noise generated from the ground to accurately exclude all non-obstacle information and Instantly obtaining the forward target information, when the above step is performed to the step of "resolving a ground noise (S64) by comparing the position information of the obstacle with the image information", the vehicle computer 30 is based on the first When the second distance information is different, or the second distance information is greater than the first distance information by a set value or more, the vehicle computer 30 performs a filtering ground noise method, as shown in FIG. The method includes the following steps:: capturing obstacle information through the ranging module 10 to target lock the first obstacle position (S91); in the embodiment, the sensing fusion processor 32 of the vehicle computer 30 utilizes the obstacle Information information, the image information, marking the closest obstacle information in the route of the vehicle, and determining whether the obstacle information is a static obstacle, and if so, according to the Obstructing the body information to perform a vehicle body identification to determine whether the front target information is a vehicle; performing an obstacle position tracking estimation (S92) to generate an estimated position information; and the estimated position information and the first obstacle The object position is compared to obtain an absolute value (S93); determining whether the absolute value is less than an error threshold value (S94); if yes, determining that the obstacle is a noise (S95), and filtering the position information of the obstacle If not, it ends (S96).

In summary, the driving system of the present invention is capable of predicting driving conditions, and more accurately pre-determining noises such as non-obstructions, environmental reflections, and ground objects, and avoiding the influence of external noise or roadside obstacles, leading to the front. The target vehicle is caught incorrectly, causing the system to malfunction and fail. Eliminate noise by importing signal processing filters, simplify driving environment information, reduce failure rate, and integrate technology into an automatic emergency braking system (AEB) to increase the accuracy of vehicle safety systems in determining obstacles. And stability, to achieve the purpose of improving driving safety.

10 Ranging Module 20 Image Capture Module 30 Car Computer 31 Microcontroller 32 Perceptual Fusion Processor 40 Brake Module 50 Vehicle Status Module 60 Alarm Module

1 is a system architecture diagram of a preferred embodiment of the present invention. 2 is a block diagram of another system in accordance with a preferred embodiment of the present invention. 3 is a waveform diagram of filtering noise according to a preferred embodiment of the present invention. FIG. 4 is another coordinate diagram of filtering noise according to a preferred embodiment of the present invention. Figure 5 is a schematic diagram showing the state of comparative analysis of a preferred embodiment of the present invention. Figure 6 is a flow chart showing the obstacle screening according to a preferred embodiment of the present invention. FIG. 7 is a flow chart of a first filtering noise method according to a preferred embodiment of the present invention. FIG. 8 is a flow chart of a second method for filtering noise according to a preferred embodiment of the present invention. 9 is a flow chart of a method for filtering ground noise according to a preferred embodiment of the present invention.

10 Ranging Module 20 Image Capture Module 30 Car Computer 40 Brake Module 50 Vehicle Status Module 60 Alarm Module

Claims (14)

An obstacle screening method for a driving safety system is mainly connected to an image capturing module and a distance measuring module by a vehicle computer, and the vehicle computer performs the following steps: receiving more than one obstacle information and Corresponding to more than one image information; performing a filtering analysis mechanism on the obstacle information to filter the noise, and filtering to generate position information of an obstacle; and comparing the position information of the obstacle with the image information Exclude a ground noise and get a forward target information. The obstacle screening method of the driving safety system according to claim 1, wherein when the step is performed to perform a filtering analysis mechanism on the obstacle information to filter the noise step, the first computer filters the first filtering noise. The method further comprises the steps of: calculating a distance information of an obstacle and an estimated distance information; performing the variance calculation on the distance information and the estimated distance information; and determining whether the variance calculation has a significant drop oscillation state If yes, it is judged as noise. The obstacle screening method of the driving safety system according to claim 2, wherein when the step is performed to perform a filtering analysis mechanism on the obstacle information to filter the noise step, the second computer filtering noise is further executed by the vehicle computer. The method further includes the steps of: obtaining information of a first obstacle position and a second obstacle position; and when the distance between the second obstacle position and the first obstacle position meets a proportional condition; determining the same Whether the position information of the obstacle position is within a distance range; if so, the second obstacle position is twice the harmonic noise. The obstacle screening method of the driving safety system according to claim 3, wherein the step is performed to exclude a ground step by comparing the position information of the obstacle with the image information. The vehicle computer performs a filtering ground noise method according to a first distance information and a second distance information, and the method further comprises the following steps: target locking the first obstacle position; performing obstacle position tracking Estimating to generate an estimated position information; comparing the estimated position information with the next first obstacle position to obtain an absolute value; determining whether the absolute value is less than an error threshold; if yes, determining The obstacle is a noise. The obstacle screening method of the driving safety system according to any one of claims 1 to 4, wherein the vehicle-based computer system obtains a vehicle status information, an obstacle information, and the driving status information and the obstacle information. A calculation is made to obtain an obstacle speed, and when the obstacle speed exceeds a speed threshold, it is determined to be a dynamic obstacle, otherwise it is a static obstacle. The obstacle screening method of the driving safety system according to any one of claims 1 to 4, wherein the obstacle information and the image information obtained by the vehicle computer are marked with a closest obstacle in a route of the vehicle. Information, and determine whether the obstacle information is a static obstacle, and if so, perform a body identification based on the location of the obstacle information to determine whether the front target information is a vehicle. The obstacle screening method of the driving safety system according to claim 5, wherein the obstacle information and the image information obtained by the vehicle computer are marked with the closest obstacle information in the route of the vehicle, and the obstacle is determined. Whether the material information is a static obstacle, if yes, a body identification is performed according to the position of the obstacle information to determine whether the front target information is a vehicle. A driving safety system includes: a ranging module that captures more than one obstacle information; and an image capturing module that captures more than one image information and corresponds to the obstacle information; a vehicle computer is respectively connected to the distance measuring module and the image capturing module, and receives the obstacle information and the image information; and the vehicle computer performs a filtering analysis mechanism to filter the non-obstacle Reflecting noise, and filtering to generate position information of an obstacle, the vehicle computer compares the position information of the obstacle with the image information to exclude a noise generated by the ground object and thereby Get a forward target information. The driving safety system of claim 8, further comprising a brake module, a vehicle status module, and an alarm module, wherein the vehicle system is respectively connected to the brake module, the vehicle status module, and the alarm a module; wherein the alarm module provides real-time information or an alarm; the brake module receives a brake/de-oil control signal to drive the vehicle to decelerate according to the judgment result of the vehicle computer; the vehicle state module is It is used to sense the current driving state of the vehicle and provide a line of vehicle status information to the vehicle computer, so that the vehicle computer uses the driving status information for other analysis applications. The driving safety system of claim 9, wherein the computer uses the image information captured by the image capturing module, the obstacle information captured by the ranging module, and is transmitted back according to the vehicle status module. The information determines a vehicle route, and if the front target information has fallen into a warning range, the brake/de-oil control signal is immediately sent to the brake control module to cause the vehicle to automatically decelerate. In the driving safety system of claim 10, the driving status information provided by the vehicle status module includes a vehicle speed signal and a corner speed signal; the vehicle computer calculates the driving status information and the obstacle information to obtain an obstacle. The velocity of the object, when the obstacle speed exceeds a speed threshold, is determined to be a dynamic obstacle. The driving safety system of claim 11, wherein the vehicle computer is mainly composed of a microcontroller connected to a sensing fusion processor, and the microcontroller is respectively associated with the ranging module, the braking module, and the The vehicle state module and the alarm module are electrically connected, and the sensory fusion processor is electrically connected to the image capturing module and the alarm module respectively; wherein the sensory fusion processor is configured according to the obstacle material The image information indicates the closest obstacle information in the route of the vehicle and determines whether the obstacle information is a static obstacle. If yes, a body identification can be performed according to the position of the obstacle information to determine the front. Whether the target information is a vehicle, and the result of marking the warning pattern is displayed through the alarm module. The driving safety system according to any one of claims 8 to 12, wherein the distance measuring module is a millimeter wave radar, and the image capturing module is a camera. The driving safety system of claim 13 includes a display unit and/or a sound unit, through which the instant information display or warning screen is provided, and an alarm sound is emitted through the sound unit.