CN115723751A

CN115723751A - Virtual track detection system and method thereof

Info

Publication number: CN115723751A
Application number: CN202211395625.3A
Authority: CN
Inventors: 陈益成; 张耘菱; 郭宜钧
Original assignee: Automotive Research and Testing Center
Current assignee: Automotive Research and Testing Center
Priority date: 2022-11-08
Filing date: 2022-11-08
Publication date: 2023-03-03

Abstract

The invention provides a virtual track detection system and a method thereof, belonging to the field of vehicle detection. When the vehicle runs on the track, the image acquisition device acquires the front road image, and the processor identifies the track pattern from the front road image and judges whether the running path of the vehicle is a straight line or a non-straight line. If the driving path is a straight line, calculating a linear equation of the driving path and outputting the linear equation to a power control end of the vehicle-mounted system. If the driving path is not straight, the position of the positioning element is detected through an inductor, and a curve curvature, a turning speed and a correction angle of the vehicle are calculated according to the position of the positioning element so as to enable the power control end to correct the driving speed and the course angle of the vehicle. The invention combines the image and the magnetic conduction type virtual track detection, is not influenced by the external environment and can reduce the cost.

Description

Virtual track detection system and method thereof

Technical Field

The invention relates to the technical field of vehicle detection, in particular to a virtual track detection system and a method thereof.

Background

The self-driving bus is an intelligent traffic technology which is actively researched and developed by current developers, adopts an unmanned technology, and can realize full-automatic and efficient boarding transfer service. Since the self-driving bus runs on a general road on which no physical rails are laid, how to advance along a virtual track, how to avoid a collision with pedestrians and vehicles, and the like are problems to be overcome.

Currently, a navigation technology combining GPS (global positioning system) navigation and map data packets is mainly applied to the self-driving bus. The navigation technique obtains a fuzzy GPS location based on the bus location, and then further pinpoints using a map data packet. However, GPS may be affected by environment or weather, and thus, signal transmission may be unstable. Another navigation technique is to use lidar synchronous positioning and mapping (lidar slam) or virtual synchronous positioning and mapping (visor slam); however, the navigation technique is affected by the environment change to the positioning accuracy, so that the base map needs to be recorded again in a fixed time when the navigation technique is used outdoors, which increases the operation cost and the occupied database space.

In addition, a navigation technique is to use a camera to capture the tracks drawn on the ground of the road ahead. As shown in fig. 1, when the vehicle 30 is traveling straight along the track, its camera field of view is sufficient to capture the track several meters ahead; as shown in fig. 2, when the vehicle 30 enters a curve, the field of view of the camera may not be sufficient to capture the track, and at this time, the vehicle 30 cannot advance along the track, and can only stop or derail. In addition, as shown in fig. 3, when the curvature of the curve is too large, although the field of view of the camera is sufficient to capture the track, the on-board system may misinterpret the track as a stop line or road edge so that the vehicle 30 does not continue to follow the track.

Disclosure of Invention

An object of the present invention is to provide a virtual rail detection system and a method thereof, which combine an image rail and a magnetic rail, and move along a rail pattern on a linear section, and use the rail pattern with a positioning element on a non-linear section, so as to avoid the occurrence of misidentification such as identifying the rail pattern as a stop line or a curb.

Another object of the present invention is to provide a virtual rail detection system and method thereof, which can achieve fully automatic driving of a vehicle without driving monitoring and GPS navigation.

To achieve the above object, the present invention provides a virtual rail detection system, comprising: a plurality of positioning elements, laid along the curve of a track; the system comprises at least one image acquisition device, a data acquisition device and a data processing device, wherein the image acquisition device is arranged on a vehicle and is used for acquiring a road image in front of the vehicle, and the road image in front comprises a track pattern of a track; the sensor is arranged on the vehicle and used for detecting the position of the positioning element; at least one processor, set in a vehicle system of the vehicle, connected with the image acquisition device and the sensor, the processor receives the front road image and identifies the track pattern, the processor judges the running path of the vehicle as a straight line or a non-straight line according to the track pattern, if the running path is a straight line, a linear equation of the running path is calculated and output to a power control end of the vehicle system, so that the power control end controls the vehicle according to the linear equation, if the running path is a non-straight line, a curve curvature is calculated according to the position of the positioning element, a curve passing speed and a correction angle of the vehicle are calculated according to the curve passing speed and the correction angle, and the curve passing speed and the correction angle are output to the power control end, so that the power control end controls the vehicle according to the curve passing speed and the correction angle.

According to an embodiment of the invention, the processor receives the front road image, searches a plurality of track feature points therein, and identifies the track pattern according to the track feature points.

According to the embodiment of the invention, the processor takes the head center of the vehicle as the origin, the head center is overlapped with the center of the track, a linear equation is calculated according to the track pattern, and then the running path is judged to be a straight line or a non-straight line.

According to the embodiment of the invention, the processor calculates the curvature of the curve according to the position of the positioning element and an image equation, calculates the over-bending speed of the vehicle according to the curvature of the curve, and calculates the correction angle of the vehicle according to the curvature of the curve and the over-bending speed.

According to the embodiment of the invention, the power control end comprises a transverse control system which is used for controlling the angle of a steering wheel of the vehicle according to the correction angle, and controlling an accelerator and a brake of the vehicle according to the over-bending speed, so that the vehicle runs according to the over-bending speed and the correction angle.

According to an embodiment of the present invention, the positioning element is a magnetic positioning element, and the inductor is a magnetic inductor.

The invention also provides a virtual track detection method, which is applied to a track, wherein a track pattern is drawn on the track, a plurality of positioning elements are laid along the curve of the track, and when a vehicle runs on the track, the virtual track detection method comprises the following steps: collecting the image of the road in front of the vehicle through at least one image collecting device; receiving the front road image and identifying a track pattern through at least one processor, and judging whether the driving path of the vehicle is a straight line or a non-straight line according to the track pattern; if the driving path is a straight line, calculating a linear equation of the driving path and outputting the linear equation to a power control end of the vehicle-mounted system so that the power control end can control the vehicle according to the linear equation, if the driving path is a non-straight line, detecting the position of each positioning element through an inductor, calculating the curvature of a curve according to the position of each positioning element, calculating the bending speed and the correction angle of the vehicle according to the curvature speed and the correction angle, and outputting the bending speed and the correction angle to the power control end so that the power control end can control the vehicle according to the bending speed and the correction angle.

According to an embodiment of the present invention, after receiving the front road image, the at least one processor searches a plurality of track feature points therein, and identifies the track pattern according to each track feature point.

According to an embodiment of the present invention, the step of searching for the track feature point includes: selecting an interested range on the front road image; a point having a set of color peaks is found from the range of interest, and the set of color peaks corresponds to a color value of the trajectory feature point.

According to an embodiment of the invention, the step of calculating the curvature of the curve from the positions of the positioning elements further comprises: the processor identifies a situation according to the track pattern; according to the position of the triggered first positioning element in the positioning elements, the curve curvature is calculated by using an image equation in cooperation with the track pattern identified in the front road image.

According to the specific embodiment provided by the invention, the invention discloses the following technical effects: the image track and the magnetic track are combined, the track pattern advances along the track pattern on the linear section, and the track pattern is matched with the positioning element to be used together on the non-linear section, so that the condition that the track pattern is identified as a stop line or a curb and the like which are identified by errors is avoided.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.

FIG. 1 is a schematic view of the field of view of a camera while a vehicle is traveling on a straight road segment;

FIG. 2 is a schematic view of a first embodiment of the field of view of the camera while the vehicle is traveling on a curved road segment;

FIG. 3 is a schematic view of a second embodiment of the field of view of the camera while the vehicle is traveling on a curved road segment;

FIG. 4 is a schematic view of a vehicle according to the present invention traveling on a road along a virtual track;

FIG. 5 is a block diagram of a virtual rail detection system according to the present invention;

FIG. 6 is a flowchart illustrating a virtual track detection method according to the present invention.

Description of the symbols:

10-virtual track detection system, 12-vehicle, 14-image acquisition device, 16-sensor, 17-vehicle system, 18-processor, 19-power control end, 191-transverse control system, 192-accelerator, 194-brake, 196-steering wheel, 20-road, 22-track pattern, 24-positioning element, 30-vehicle, and S10-S26-step flow.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.

It will be understood that the terms "comprises" and "comprising," when used in this specification and appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting. As used in the specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in this specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items and includes such combinations.

In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.

Please refer to fig. 4 and 5 together, wherein fig. 4 is a schematic view illustrating a vehicle 12 traveling on a road 20 along a virtual track according to the present invention, and fig. 5 is a block diagram illustrating a virtual track detection system 10 according to the present invention. The virtual rail detection system 10 includes a vehicle end and a rail end on which the vehicle 12 travels. The vehicle 12 is an autonomous vehicle that travels along a virtual track, which has no physical rails and is traditionally a lined track. The road surface of the road 20 is provided with at least one track pattern 22 as a virtual track, and different sections may have different track patterns 22, for example, the track patterns 22 of a straight line section, an entrance/exit section, a curved road section, etc. may be different. A plurality of locating elements 24 are provided in the curved portion of the roadway 20 for locating the curved portion.

The vehicle 12 is provided with at least one image capturing device 14, at least one sensor 16, and an onboard system 17. The on-board system 17 has at least one processor 18 and a power control end 19 in signal communication with each other, wherein the power control end 19 has control elements such as a throttle 192, a brake 194 and a steering wheel 196 for controlling the speed and direction of the vehicle 12. The image capture device 14 and the sensor 16 are in signal communication with a processor 18. An image capturing device 14 and a sensor 16 are disposed at the front end of the vehicle 12, wherein the image capturing device 14 is a camera or a video camera for capturing images of the road ahead of the vehicle 12, and the sensor 16 is used for detecting the positioning element 24. The processor 18 is disposed in an onboard system 17 of the vehicle 12, receives the front road image collected by the image collecting device 14, and identifies the track pattern 22 of the road 20 from the front road image. The processor 18 determines whether the travel path of the vehicle 12 is linear or non-linear based on the track pattern 22, and switches between a linear track mode or a curve detection mode accordingly.

If the path of travel of the vehicle 12 is straight, the processor 18 initiates the straight track mode. The processor 18 calculates a linear equation of the traveling path and outputs the linear equation to a power control end 19 of the vehicle-mounted system 17, so that the power control end 19 can control the vehicle 12 according to the linear equation. The following equation (1) is a linear equation, which is a binary quadratic error equation (errorEquation), and the values of the coefficients a, b, and c are calculated using three simultaneous equations.

Wherein

Further, since one lane has two lane lines, left and right, the vehicle coordinate system positions (x) of the feature points of all the left and right lane lines _i ,y _i ) The values of a, b and c can be calculated by substituting the above formula (1) to obtain the left lane line equation and the right lane line equation respectively. After fitting the left lane line equation and the right lane line equation, the reliability analysis and the logic analysis are combined to eliminate error information (such as characteristic points of non-lane lines with errors eliminated), and then a lane model can be established. Finally, the power supply control terminal 19 controls the vehicle 12 according to the lane model.

If the travel path is not straight, the processor 18 switches to the curve detection mode. The processor 18 will calculate a curve curvature based on the position of the positioning element 24. Since the curve curvature is larger, the speed of passing the curve needs to be adjusted down so as not to turn over or derail, and therefore a speed of passing the curve and a correction angle of the vehicle 12 need to be calculated further according to the curve curvature. The corrected angle is the yaw angle of the vehicle 12, i.e., the steering angle of the steering wheel, to control the heading angle of the vehicle 12. Then, the processor 18 outputs the over-bending speed and the corrected angle of the vehicle 12 to the power control end 19, so that the power control end 19 controls the vehicle 12 according to the over-bending speed and the corrected angle.

In one embodiment, the sensor 16 is a magnetic sensor for detecting the position of the positioning element 24, and the positioning element 24 may be a magnetic pin or other magnetic element.

The positioning element 24 is disposed along the center line of the road of the curve. The spacing distance of each positioning element 24 need not be the same. For example, if the curvature of the curve is larger, the track pattern 22 is not within the field of view of the image capturing device 14 (as shown in fig. 2). While the sensor 16 can detect the locating element 24 on a curve, if the next locating element 24 is detected to be 3 degrees to the left and the vehicle 12 is traveling too fast, the inertia of the vehicle will likely cause the vehicle 12 to have too much time to deflect and the vehicle 12 will derail. Therefore, when the curvature of the curve is large, not only is the spacing distance between the positioning elements 24 shortened and the density increased, but the speed of the vehicle 12 passing through the curve is controlled to decrease to avoid the vehicle 12 having no time to deflect when the positioning elements 24 are detected.

In one embodiment, the power control end 19 further includes a lateral control system 191 that couples the throttle 192, the brake 194, and the steering wheel 196. The lateral control system 191 is signally connected to the processor 18 and may control the angle of the steering wheel 196 of the vehicle 12 based on the corrected angle. In addition, the lateral control system 191 may also control the throttle 192 and brake 194 of the vehicle 12 based on the speed of the over-curve to allow the vehicle 12 to travel safely.

The application method of the virtual track detection system of the present invention is described in detail below. Please refer to fig. 6, which is a flowchart illustrating a virtual track detection method according to the present invention. When the vehicle 12 is traveling on the track, as described in step S10, the image of the road ahead of the vehicle 12 is captured by the at least one image capturing device 14, and the image of the road ahead includes other vehicles, signs, pedestrians, track patterns, and the like on the road. In step S12, the at least one processor 18 receives the front road image and identifies the track pattern. Further, the processor 18 identifies the track pattern from the front road image by filtering non-track objects such as pedestrians, road trees, signs, etc. from the front road image by using an image processing algorithm, finds out a plurality of track feature points that can represent the track, and connects the track feature points to form the track pattern. Next, in step S14, the processor 18 determines whether the travel path of the vehicle 12 is a straight line or a non-straight line according to the track pattern.

If the travel path is a straight line, in steps S16 to S18, the processor 18 starts a straight track mode, calculates a linear equation of the travel path, and outputs the linear equation to a power control end 19 of the on-board system 17, so that the power control end 19 controls the vehicle 12 according to the linear equation.

If the travel path is not straight, the processor 18 initiates a curve detection mode to detect the position of the positioning element 24 via the sensor 16, calculates a curve curvature based on the position of the positioning element 24, and calculates a turn speed and a correction angle of the vehicle 12 using the curve curvature, as described in steps S20-S24. The processor 18 then calculates the speed of the vehicle 12 to negotiate a curve to avoid derailment in time of the curve due to too fast a speed of the vehicle. Finally, the processor 18 calculates the corrected angle of the vehicle 12 based on the curve curvature and the speed of the curve passing. Then, as shown in step S26, the processor 18 outputs the over-bending speed and the corrected angle to the power control end 19, so that the power control end 19 controls the vehicle 12 according to the over-bending speed and the corrected angle, such as controlling the throttle 192 to be released, controlling the brake 194 to decelerate, and controlling the steering wheel 196 to turn, etc.

In one embodiment, the processor 18 finds a plurality of track feature points representing the track in step S12 by framing the interest area in the front road image, such as framing the ground information. Then, the portion of the image having a set of color peaks is found from the interested range, and the color peaks are matched with the color values of the orbit feature points. Taking the lane lines as an example, the lane lines are white, the RGB values on the color list are close (250, 250, 250), and the ground is black, but since in the daytime and in the case of a taillight, the normal condition is not detected as being completely black, the RGB values on the color list are close (10, 10, 10). The peak value is identified by utilizing the color in the image, and the numerical value close to the color of the track characteristic point is found out, so that the track characteristic point can be found out.

In one embodiment, when the processor 18 determines in step S14 that the travel path is a straight line or a non-straight line, the processor calculates a linear equation from the track pattern 22 with the head center of the vehicle 12 as the origin and the head center overlapping the center of the track, and determines that the travel path is a straight line or a non-straight line. If the travel path is not a straight line, the linear equation is obviously not a straight line equation, but a special path equation.

In one embodiment, the track pattern 22 on the road 20 is provided with a special graphic representation in order to enable the processor 18 to determine whether the travel path is straight or not in step S14. Thus, the processor 18 only recognizes the particular graphic representation and knows that it is going to enter a curve or to enter or exit from a station next, rather than going straight. This eliminates the need to calculate a linear equation.

In one embodiment, the processor 18 further includes the following step in step S22: the processor 18 first identifies a situation, such as a curve situation or an entrance/exit situation, based on the track pattern 22, and then calculates a curve curvature using an image equation according to the position of the triggered first positioning element 24 and the line or shape of the track pattern 22 identified by the front road image, so as to correct the vehicle heading angle. The image equation is a linear equation for calculating the lane by using the least squares method. Calculating linear equations of the left lane line and the right lane line, fitting to obtain a linear equation of the center line (i.e. the track pattern) of the lane

Wherein the coefficient c is the curvature of the curve, b is a straight line, and a is a constant term. The values of a, b and c can be obtained after differentiating the linear equation, and further the curvature of the curve can be obtained.

Because the invention can draw the track pattern with special graphic representation in special situations and match with the positioning element, the station can be accurately accessed. For example, special drawings are drawn at the entrance and exit of the station, and dense positioning elements are laid, such as one positioning element every 10 cm. When the special graphic representation enters the visual field range of the image acquisition device, the vehicle can predict that the vehicle is about to enter the station; when the sensor detects the first positioning element, the vehicle can judge that the vehicle enters the station at present and is about to arrive at the parking position.

The present invention can be used in conjunction with image tracks and magnetic tracks to guide the travel of the vehicle 12. In a straight section, the vehicle 12 follows the track pattern. When the path changes (non-linear), the image navigation and the magnetic navigation are combined, and the position of the positioning element is detected while the image of the front road is detected to give a curved preview, for example, whether the track pattern of the image of the front road is a straight line or not, whether a special diagram exists or not is viewed in advance, so as to correct the track of the driving path in real time and plan the heading angle of the vehicle. In particular, when the road junction is free of track patterns, the vehicle can still be driven automatically by the positioning elements laid down. However, in the case of an excessively large curvature of a curve or a complex road situation, driving monitoring is required and a GPS signal is matched in the prior art, otherwise, a risk of erroneous detection occurs, for example, the curve is mistakenly recognized as a stop line or a curb and filtered. The invention can realize full automatic driving, does not need driving monitoring and GPS navigation, and can solve the problem of road detection error only by laying denser positioning elements at the curve or the intersection without a track pattern.

In summary, the present invention provides a virtual track detection system and a method thereof, which combine an image track and a magnetic track to follow a track pattern on a linear road section. When special situations such as a curve and a station are met, the vehicle is switched to a curve detection mode at a proper time, and the image of the road in front of the vehicle is matched for detecting the curve, so that the vehicle can still stably drive by itself and move forward even if the path change is too large, and the curve cannot be mistakenly identified as a stop line or a road edge.

The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.

The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.

Claims

1. A virtual rail detection system, comprising:

a plurality of positioning elements, laid along the curve of a track;

the image acquisition device is arranged on a vehicle and used for acquiring a front road image of the vehicle, wherein the front road image comprises a track pattern of the track;

the sensor is arranged on the vehicle and used for detecting the position of each positioning element;

at least one processor, which is arranged in a vehicle-mounted system of the vehicle and is connected with the image acquisition device and the sensor, receives the front road image and identifies the track pattern, the at least one processor judges the driving path of the vehicle to be a straight line or a non-straight line according to the track pattern,

if the driving path is a straight line, calculating a linear equation of the driving path and outputting the linear equation to a power control end of the vehicle-mounted system for the power control end to control the vehicle according to the linear equation,

if the driving path is non-linear, calculating a curve curvature according to the position of each positioning element, calculating a bending speed and a correction angle of the vehicle according to the curve curvature, and outputting the bending speed and the correction angle to the power control end so that the power control end can control the vehicle according to the bending speed and the correction angle.

2. The virtual rail detection system of claim 1, wherein the at least one processor receives the front road image, searches for a plurality of rail feature points therein, and identifies the rail pattern according to each rail feature point.

3. The virtual rail detection system of claim 1, wherein the at least one processor uses a head center of the vehicle as an origin, and the head center overlaps with a center of the rail, calculates the linear equation according to the rail pattern, and determines whether the travel path is a straight line or a non-straight line.

4. The virtual rail detection system of claim 1, wherein the at least one processor calculates the curve curvature according to the position of each positioning element and an image equation, calculates the vehicle turning speed according to the curve curvature, and calculates the vehicle correction angle according to the curve curvature and the turning speed.

5. The virtual rail detection system of claim 1, wherein the power control end comprises a lateral control system for controlling the steering wheel angle of the vehicle according to the correction angle and controlling the throttle and brake of the vehicle according to the over-bending speed, so that the vehicle can run according to the over-bending speed and the correction angle.

6. The virtual rail detection system of claim 1, wherein each positioning element is a magnetic positioning element and the at least one sensor is a magnetic sensor.

7. A virtual track detection method is applied to a track, a track pattern is drawn on the track, and a plurality of positioning elements are laid along the curve of the track, and is characterized in that when a vehicle runs on the track, the virtual track detection method comprises the following steps:

collecting the front road image of the vehicle through at least one image collecting device;

receiving the front road image and identifying the track pattern through at least one processor, and judging that the driving path of the vehicle is a straight line or a non-straight line according to the track pattern;

if the driving path is a straight line, calculating a linear equation of the driving path, and outputting the linear equation to a power control end of the vehicle-mounted system so that the power control end can control the vehicle according to the linear equation;

if the driving path is non-linear, the position of each positioning element is detected through an inductor, a curve curvature is calculated according to the position of each positioning element, a bending speed and a correction angle of the vehicle are calculated according to the curve curvature, and the bending speed and the correction angle are output to the power control end, so that the power control end can control the vehicle according to the bending speed and the correction angle.

8. The method as claimed in claim 7, wherein the at least one processor searches for a plurality of track feature points after receiving the front road image, and identifies the track pattern according to each track feature point.

9. The virtual track detection method of claim 8, wherein the step of searching for the plurality of track feature points comprises:

framing an interesting range on the front road image;

a point with a set of color peaks is found from the interested range, and the set of color peaks matches a color value of the orbit feature point.

10. The method as claimed in claim 7, wherein the at least one processor calculates the linear equation according to the track pattern by using a head center of the vehicle as an origin and the head center overlaps with a center of the track, and determines whether the driving path is a straight line or a non-straight line.

11. The virtual rail detection method of claim 7, wherein the at least one processor calculates the curve curvature according to the position of each positioning element and an image equation, calculates the vehicle turning speed according to the curve curvature, and calculates the vehicle correction angle according to the curve curvature and the turning speed.

12. The method as claimed in claim 7, wherein the power control end comprises a lateral control system for controlling a steering wheel angle of the vehicle according to the correction angle and controlling a brake and an accelerator of the vehicle according to the over-bending speed, so that the vehicle can travel according to the over-bending speed and the correction angle.

13. The virtual rail detection method of claim 7, wherein the step of calculating the curvature of the curve based on the position of each positioning element further comprises:

the at least one processor identifies a situation according to the track pattern;

and calculating the curvature of the curve by using an image equation according to the position of the triggered first positioning element in each positioning element and the track pattern identified in the front road image.