CN113702967B - Method for identifying and tracking guided vehicle target of ground unmanned platform and vehicle-mounted system - Google Patents
Method for identifying and tracking guided vehicle target of ground unmanned platform and vehicle-mounted system Download PDFInfo
- Publication number
- CN113702967B CN113702967B CN202111122185.XA CN202111122185A CN113702967B CN 113702967 B CN113702967 B CN 113702967B CN 202111122185 A CN202111122185 A CN 202111122185A CN 113702967 B CN113702967 B CN 113702967B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- target
- tracking
- millimeter wave
- guiding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/86—Combinations of radar systems with non-radar systems, e.g. sonar, direction finder
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/66—Radar-tracking systems; Analogous systems
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/86—Combinations of radar systems with non-radar systems, e.g. sonar, direction finder
- G01S13/865—Combination of radar systems with lidar systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Optimization (AREA)
- Mathematical Physics (AREA)
- Mathematical Analysis (AREA)
- Data Mining & Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Computational Mathematics (AREA)
- Artificial Intelligence (AREA)
- Algebra (AREA)
- Probability & Statistics with Applications (AREA)
- Traffic Control Systems (AREA)
- Radar Systems Or Details Thereof (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention relates to a guided vehicle target recognition and tracking method and a vehicle-mounted system of a ground unmanned platform, and belongs to the technical field of semi-autonomous navigation of the ground unmanned platform. The invention provides a safe and reliable guiding target identifying and tracking scheme, which can effectively realize guiding-following unmanned formation transportation, save manpower support in the material transportation process, improve the intelligent level of the formation transportation system, and simultaneously improve the safety and reliability of the unmanned autonomous formation transportation system.
Description
Technical Field
The invention belongs to the technical field of semi-autonomous navigation of ground unmanned platforms, and particularly relates to a guided vehicle target identification and tracking method and a vehicle-mounted system of the ground unmanned platform.
Background
The guiding-following system of the ground unmanned platform is a practical semi-autonomous maneuvering control mode, utilizes a perception sensor carried by the unmanned platform to carry out target identification and continuous tracking on the guiding vehicle, does not need to modify the guiding vehicle, realizes formation transportation of guiding the unmanned vehicle by the unmanned vehicle, can be applied to material transportation guarantee, can effectively reduce drivers, and simultaneously improves the intelligent level of long-distance transportation.
In the prior art, the following main problems exist in the target recognition and tracking implementation mode of the ground unmanned platform guiding-following system:
1. target identification and tracking based on visual images. On the one hand, the vision-based target recognition result needs to be converted into a Cartesian coordinate system of the vehicle body, and a coordinate conversion precision error exists, so that the precise tracking control of the following vehicle is not facilitated. On the other hand, the vision sensor is greatly affected by illumination, shielding of viewing angle, etc., and particularly when turning or passing through a tunnel in which illumination is drastically changed, the target is easily lost. In addition, the vision sensor is greatly influenced by bad weather such as rain, snow, fog and the like, and can not provide target recognition and tracking results for guiding-following tasks under all-weather conditions.
2. Target identification and tracking based on millimeter wave radar. The millimeter wave radar can effectively track a dynamic target, has low discrimination to a static vehicle, and is easy to cause failure in initializing the tracking target when the vehicle target is initialized. In addition, in the case where the target is lost, it is difficult to effectively retrieve the target.
3. Target identification and tracking based on laser radar. The laser radar can identify and track the vehicle target through geometric modeling, but as the radar point cloud becomes sparse along with the increase of the distance, the detection distance of the vehicle target is limited, and the short-distance tracking easily causes accidents in the high-speed tracking process, so that a certain safety risk exists.
4. A target tracking mode based on radio communication transmitting position information. The positioning system and the radio communication system are required to be built on the guide vehicle, and the guide vehicle is greatly improved, so that the guide vehicle has no universality and economy. Meanwhile, potential safety risks such as radio transmission delay and electromagnetic interference influence exist, the position accuracy of the guided vehicle is difficult to guarantee under the influence of the communication delay, and the control accuracy of the following vehicle is further influenced.
Disclosure of Invention
First, the technical problem to be solved
The invention aims to solve the technical problems that: and how to realize target identification and continuous tracking of the guided vehicle for autonomous vehicle following of a ground unmanned platform, thereby realizing unmanned formation transportation guarantee of semi-autonomous maneuver.
(II) technical scheme
In order to solve the technical problems, the invention provides a guided vehicle target recognition and tracking method of a ground unmanned platform, which is realized by a vehicle-mounted system, wherein the system comprises combined navigation equipment, a three-dimensional laser radar, a millimeter wave radar and a vehicle-mounted industrial personal computer, in the method, point cloud data acquired by the three-dimensional laser radar is utilized for feature extraction and guided vehicle modeling, and recognition and tracking information initialization of the guided vehicle target is realized; after the guided vehicle runs for a certain distance, after a target tracking result based on the three-dimensional laser radar point cloud data is stable, matching the tracking result of the millimeter wave radar with the tracking result of the three-dimensional laser radar, and switching to target tracking based on the millimeter wave radar; in the continuous tracking process, the gesture information provided by the integrated navigation equipment is used for merging the historical track of target tracking, and the motion trend information of the guided vehicle is calculated by using a Kalman predictor in the vehicle-mounted industrial personal computer.
Preferably, both the millimeter wave radar and the three-dimensional laser radar are horizontally installed in front of the vehicle; the inertial navigation device in the integrated navigation device is arranged in the vehicle body, and the Y-axis direction is parallel to the central axis of the advancing direction of the vehicle body; the GPS/Beidou antenna in the integrated navigation device is arranged outside the vehicle body, so that shielding is avoided; the vehicle-mounted industrial personal computer is arranged in the unmanned platform vehicle body, defines the origin of a local coordinate system of the vehicle body as the center of a rear shaft of the vehicle body, the advancing direction of the vehicle body is Y-axis positive direction, the horizontal right side vertical to the advancing direction of the vehicle body is X-axis positive direction, the horizontal right side vertical to the horizontal plane and the Z-axis positive direction along the sky of the vehicle body.
Preferably, before the motion trend information of the guided vehicle is calculated, external calibration parameters of the three-dimensional laser radar, the millimeter wave radar and the unmanned platform are obtained through joint calibration, and based on the external calibration parameters, data of the millimeter wave radar, the three-dimensional laser radar, the GPS/Beidou antenna and the inertial navigation device are unified to a vehicle body local coordinate system by utilizing a formula (1), so that reference coordinate unification of the data is realized;
in formula (1), p sensor ＝(x s ,y s ,z s ) T Representing the coordinates of data in a three-dimensional lidar coordinate system, p vehicle ＝(x v ,y v ,z v ) T Representing the coordinates of the data in the local coordinate system of the vehicle body,for rotating matrix +.>The elements in the rotation matrix and the translation matrix are external calibration parameters.
Preferably, in the method, the states of target recognition and tracking of the guided vehicle are divided into four working states of target recognition based on a three-dimensional laser radar, target tracking based on the three-dimensional laser radar, target tracking based on a millimeter wave radar and target tracking based on Kalman prediction, and the implementation flow of each working state is as follows:
step 1: starting the guide vehicle;
step 2: entering a target identification state based on a three-dimensional laser radar, guiding the vehicle target to confirm, and turning to the step 3;
step 3: based on the target tracking of the three-dimensional laser radar, the target tracking is effective, and the step 4 is switched, otherwise, the step 7 is switched;
step 4: matching with a target detection result based on the millimeter wave radar, if so, turning to step 5, otherwise, returning to step 3;
step 5: performing millimeter wave radar-based target tracking;
step 6: if the target is valid, returning to the step 5; otherwise, returning to the step 3;
step 7: performing target detection based on Kalman prediction;
step 8: judging the result of target detection based on Kalman prediction, if the result is available, returning to the step 2, otherwise, returning to the step 7;
the process is finished after the ending instruction is received, otherwise, the process is executed circularly all the time.
Preferably, in the method, according to the maneuvering process of guiding-following, guiding of the ground unmanned platform is divided into 4 sub-processes of starting guiding-following, guiding-following in a continuous process, guiding-following restarted after stopping, and guiding-following under target loss.
Preferably, the start guiding-following sub-process is a process of a ground unmanned platform autonomous vehicle tracking target identifying and tracking method under a guided vehicle start condition, and comprises the following steps:
step 1: the guiding vehicle is statically parked in a certain range in front of the unmanned platform, and a guiding vehicle target recognition and tracking program is started;
step 2: the method comprises the steps of carrying out point cloud data acquisition and feature extraction on a guided vehicle through a three-dimensional laser radar, and identifying a guided vehicle target;
step 3: the target recognition result of the guided vehicle based on the three-dimensional laser radar is stable, and the target initialization is completed;
step 4: starting continuous tracking of a target based on a laser radar, and transforming a detection result of a local coordinate system of a vehicle body into a global coordinate system according to vehicle posture information acquired by the integrated navigation equipment to form a historical tracking track;
step 5: the guiding vehicle dynamically moves, and the unmanned platform is started to guide and advance according to the local position information tracked by the target;
step 6: starting vehicle target tracking based on millimeter wave radar;
step 7: using a Hungary matching algorithm, using continuous tracking target position information of the laser radar as a basis, matching a dynamic target detection result of the millimeter wave radar with the continuous tracking target position information, and executing the step 8; otherwise, returning to the step 5;
step 8: continuously tracking a vehicle target based on millimeter wave radar;
step 9: estimating the position of the guiding vehicle at the next moment by using a Kalman estimator according to the current target tracking position;
step 10: and setting a region of interest according to the predicted position, and returning to the step 8.
Preferably, the guiding-following sub-process in the continuous process is a process of detecting the target state of the guiding vehicle in the tracking continuous process and judging the transverse and longitudinal states of the guiding vehicle, and specifically comprises the following steps:
step 1: acquiring the relative speed of the guided vehicle by utilizing a millimeter wave radar;
step 2: combining the vehicle speed information provided by the integrated navigation equipment, and calculating to obtain the speed information of the guided vehicle;
step 3: recording the local coordinate position of a guided vehicle target under each frame of data, and transforming a target tracking result under a vehicle body local coordinate system into a global coordinate system according to the vehicle posture information acquired by the integrated navigation equipment to form a history track;
step 4: continuously detecting T periods, and judging the speed state of the guided vehicle by using a Bayesian network, wherein the speed state comprises uniform speed, acceleration, deceleration, scram, parking and starting;
step 5: and predicting the transverse state of the vehicle, including straight running, left turning and right turning, by using a Kalman predictor according to the historical track information of the guided vehicle.
Preferably, the guiding-following sub-process restarted after parking is a guiding vehicle target repositioning and continuous tracking processing flow under the condition that the guiding vehicle is started after parking, and comprises the following steps:
step 1: judging whether a guiding vehicle target under the millimeter wave radar is lost, if not, executing the step 2, otherwise, executing the step 3;
step 2: setting an interested area, and when the guided vehicle moves dynamically, acquiring the dynamic information of the guided vehicle by utilizing a millimeter wave radar, and tracking the dynamic target of the guided vehicle based on the millimeter wave radar;
step 3: and repeating the flow of the ground unmanned platform autonomous vehicle tracking target identification and tracking method under the starting condition of the guided vehicle.
Preferably, the guiding-following sub-process under the condition of target loss is a target tracking process processing flow under the condition that a guiding vehicle is blocked, a target is temporarily lost, and the like, and comprises the following steps:
step 1: if the millimeter wave radar detection result exists in the region of interest, executing the step 2, otherwise executing the step 3;
step 2: continuing to track the target based on the millimeter wave radar;
step 3: judging whether a guiding vehicle target tracking result based on the three-dimensional laser radar exists in the region of interest, if so, taking the guiding vehicle target tracking result as output, executing the step 4, otherwise, executing the step 5;
step 4: the target detection result based on the millimeter wave radar is matched with the target detection result based on the three-dimensional laser radar, if the result is matched, the step 2 is executed, otherwise, the step 3 is returned;
step 5: predicting target information at the next moment by using a prediction result of a Kalman predictor according to the historical track information, and outputting the prediction information as output;
step 6: searching targets in the whole domain range by using a laser radar, identifying the targets of the guided vehicle, returning to the step 5 without identification results, and executing the step 7 otherwise;
step 7: and repeating the flow of the ground unmanned platform autonomous vehicle tracking target identification and tracking method under the starting condition of the guided vehicle.
The invention also provides a vehicle-mounted system adopted in the implementation process of the method.
(III) beneficial effects
The invention provides a safe and reliable guiding target identifying and tracking scheme, which can effectively realize guiding-following unmanned formation transportation, save manpower support in the material transportation process, improve the intelligent level of the formation transportation system, and simultaneously improve the safety and reliability of the unmanned autonomous formation transportation system.
Drawings
FIG. 1 is a state diagram of overall lead vehicle target identification and tracking in accordance with the present invention;
FIG. 2 is a chart of a guide car target movement trend prediction of the present invention;
FIG. 3 is a general flow chart of the present invention;
FIG. 4 is a state diagram of the guided vehicle target recognition and tracking of the present invention;
FIG. 5 is an exploded view of the lead-following target recognition and tracking process of the present invention.
Detailed Description
For the purposes of clarity, content, and advantages of the present invention, a detailed description of the embodiments of the present invention will be described in detail below with reference to the drawings and examples.
The existing ground unmanned platform guiding-following system is limited by the limitation of a single sensor, and has the defects in the aspects of speed, safety, tracking distance and the like. Meanwhile, the design of the guiding-following system should minimize the modification of the guiding vehicle in view of economy and versatility. Aiming at the problems, the invention provides a ground unmanned platform guided vehicle target identification and tracking method and a vehicle-mounted system based on multi-source information fusion, and particularly relates to an autonomous vehicle guided vehicle target identification and tracking method of the ground unmanned platform and a related vehicle-mounted system. Through carrying a series of perception sensors on ground unmanned platform, need not to reform transform the guide car, can carry out target identification and continuous tracking to the guide car for ground unmanned platform's autonomous vehicle follows, realizes semi-autonomous mobile unmanned formation transportation guarantee.
The carried vehicle-mounted sensor comprises a three-dimensional laser radar, a millimeter wave radar and a combined navigation system (GPS/Beidou+inertial navigation device), the three-dimensional laser radar is utilized to solve the problem of target identification of a guided vehicle and the problem of target loss during steering, the millimeter wave radar is utilized to solve the problem of long-distance target tracking in a dynamic high-speed tracking process, and meanwhile, the vehicle-mounted sensor is free from the influence of severe weather such as rain, snow and fog, and the stable tracking of the guided target under all-weather conditions is ensured. The integrated navigation system is used for carrying out historical data association on tracking information, is used for target repositioning when tracking is lost, ensures the tracking continuity and improves the system safety and reliability. The safe and effective tracking distance of the guiding vehicle target recognition and tracking system based on multi-sensor information fusion can reach 100 meters, and the guiding vehicle target recognition and tracking system based on the multi-sensor information fusion has great significance for realizing high-speed and stable tracking of the guiding-tracking system of the ground unmanned platform.
In order to achieve the aim of the invention, the method for identifying and tracking the target of the guided vehicle of the ground unmanned platform and the related hardware equipment of the vehicle-mounted system comprise combined navigation equipment (GPS/Beidou+inertial navigation), three-dimensional laser radar, millimeter wave radar, a vehicle-mounted industrial personal computer and the like. The vehicle-mounted system acquires environment sensing information through a three-dimensional laser radar and a millimeter wave radar, and acquires attitude information of the unmanned platform by utilizing the integrated navigation equipment.
According to the method, with reference to FIG. 1, point cloud data acquired by a three-dimensional laser radar is utilized to perform feature extraction and guided vehicle modeling, so that the identification and tracking information initialization of a guided vehicle target are realized; after the guided vehicle runs for a certain distance, after a target tracking result based on the three-dimensional laser radar point cloud data is stable, matching the tracking result of the millimeter wave radar with the tracking result of the three-dimensional laser radar, and switching to target tracking based on the millimeter wave radar; in the continuous tracking process, the gesture information provided by the integrated navigation equipment is used for merging the historical track of target tracking, and the motion trend information of the guided vehicle is calculated by using a Kalman predictor in the vehicle-mounted industrial personal computer. On the basis of guiding vehicle target prediction, the interested area of guiding vehicle detection at the next moment is reduced, on one hand, the real-time performance of guiding vehicle target detection is improved, on the other hand, the interference of other vehicles in irrelevant areas is filtered, when the guiding vehicle target is temporarily lost due to factors such as shielding, the system can keep the effective last moment position of guiding vehicle tracking, combines the track trend of history tracking, outputs target prediction information, and repeats the target identification and tracking process, thereby realizing safe, reliable, stable and continuous target tracking.
On the arrangement of a ground unmanned platform vehicle-mounted system, the sensing sensor comprises a millimeter wave radar and a three-dimensional laser radar, which are horizontally arranged in front of the vehicle; the inertial navigation device in the integrated navigation device is arranged in the vehicle body, and the Y-axis direction is parallel to the central axis of the advancing direction of the vehicle body; the GPS/Beidou antenna in the integrated navigation device is arranged outside the vehicle body, so that shielding is avoided; the vehicle-mounted industrial personal computer is arranged in the unmanned platform vehicle body. Defining the origin of a local coordinate system of the vehicle body as the center of a rear shaft of the vehicle body, wherein the advancing direction of the vehicle body is the positive direction of a Y shaft, the horizontal right side vertical to the advancing direction of the vehicle body is the positive direction of an X shaft, and the direction along the vehicle body to the sky is the positive direction of a Z shaft. External calibration parameters of each sensor and the unmanned platform are obtained through combined calibration, and based on the external calibration parameters, data of the millimeter wave radar, the three-dimensional laser radar, the GPS/Beidou antenna and the inertial navigation device are unified under a local coordinate system of a vehicle body by using a formula (1), so that reference coordinate unification of the data is realized, and data fusion processing is facilitated. The output information of the guiding target comprises movement trend information of the guiding vehicle in the local coordinate system of the unmanned platform vehicle body, and the information is calculated by the vehicle-mounted industrial personal computer and provided for the vehicle-mounted control system to carry out tracking control, as shown in fig. 3.
In formula (1), p sensor ＝(x s ,y s ,z s ) T Representing the coordinates of data in a three-dimensional lidar coordinate system, p vehicle ＝(x v ,y v ,z v ) T Representing the coordinates of the data in the local coordinate system of the vehicle body,for rotating matrix +.>The elements in the rotation matrix and the translation matrix are external calibration parameters.
Referring to fig. 2, the motion trend prediction of the guided vehicle in the target recognition and tracking of the guided vehicle by the unmanned platform comprises a lateral motion trend estimation and a longitudinal motion trend estimation, wherein the lateral motion trend estimation is obtained by predicting the historical tracking track position of the guided vehicle target by using a kalman predictor, and specifically comprises straight running, left turning and right turning; and estimating the longitudinal movement trend, and analyzing and obtaining according to the historical movement speed of the guided vehicle target by utilizing a Bayesian network, wherein the longitudinal movement trend comprises starting, accelerating, decelerating, uniform speed, scram and stopping states.
Referring to fig. 4, in the guiding-following process implemented by the vehicle-mounted system, the states of guiding the vehicle target recognition and tracking are divided into four working states of three-dimensional laser radar-based target recognition, three-dimensional laser radar-based target tracking, millimeter wave radar-based target tracking and kalman prediction-based target tracking according to the actual working states, and the implementation flow of each working state is as follows:
step 1: starting the guide vehicle;
step 2: entering a target identification state based on a three-dimensional laser radar, guiding the vehicle target to confirm, and turning to the step 3;
step 3: based on the target tracking of the three-dimensional laser radar, the target tracking is effective, and the step 4 is switched, otherwise, the step 7 is switched;
step 4: matching with a target detection result based on the millimeter wave radar, if so, turning to step 5, otherwise, returning to step 3;
step 5: performing millimeter wave radar-based target tracking;
step 6: if the target is valid, returning to the step 5; otherwise, returning to the step 3;
step 7: performing target detection based on Kalman prediction;
step 8: judging the result of target detection based on Kalman prediction, if the result is available, returning to the step 2, otherwise, returning to the step 7;
the process is finished after the ending instruction is received, otherwise, the process is executed circularly all the time.
Referring to fig. 5, according to the maneuvering process of guiding-following, guiding of the ground unmanned platform is divided into 4 sub-processes of starting guiding-following, guiding-following in a continuous process, guiding-following restarted after stopping, guiding-following under the condition of target loss, and target recognition and tracking of the guided vehicle in the 4 sub-processes are respectively described.
Start-guide-following sub-process: the ground unmanned platform autonomous vehicle tracking target identification and tracking method under the condition of starting the guided vehicle comprises the following steps:
step 1: the guiding vehicle is statically parked in a certain range in front of the unmanned platform, and a guiding vehicle target recognition and tracking program is started;
step 2: the method comprises the steps of carrying out point cloud data acquisition and feature extraction on a guided vehicle through a three-dimensional laser radar, and identifying a guided vehicle target;
step 3: the target recognition result of the guided vehicle based on the three-dimensional laser radar is stable, and the target initialization is completed;
step 4: based on continuous tracking start of a target of a laser radar, according to vehicle attitude information acquired by the integrated navigation equipment, converting a detection result of a vehicle body local coordinate system into a global coordinate system to form a historical tracking track;
step 5: the guiding vehicle dynamically moves, and the unmanned platform is started to guide and advance according to the local position information tracked by the target;
step 6: starting vehicle target tracking based on millimeter wave radar;
step 7: matching the dynamic target detection result of the millimeter wave radar with the continuous tracking target position information of the laser radar by using a Hungary matching algorithm, and turning to the step 8; otherwise, turning to step 5;
step 8: continuously tracking a vehicle target based on millimeter wave radar;
step 9: estimating the position of the guiding vehicle at the next moment by using a Kalman estimator according to the current target tracking position;
step 10: and setting a region of interest according to the predicted position, and returning to the step 8.
Guide-following sub-process in continuous process: the method comprises the following steps of detecting the target state of the guide vehicle in the continuous tracking process, judging the transverse and longitudinal states of the guide vehicle, and specifically comprises the following steps:
step 1: acquiring the relative speed of the guided vehicle by utilizing a millimeter wave radar;
step 2: combining the vehicle speed information provided by the integrated navigation equipment, and calculating to obtain the speed information of the guided vehicle;
step 3: recording the local coordinate position of a guided vehicle target under each frame of data, and transforming a target tracking result under a vehicle body local coordinate system into a global coordinate system according to the vehicle posture information acquired by the integrated navigation equipment to form a history track;
step 4: continuously detecting T periods, and judging the speed state of the guided vehicle by using a Bayesian network, wherein the speed state comprises uniform speed, acceleration, deceleration, scram, parking, starting and the like;
step 5: and predicting the transverse state of the vehicle, including straight running, left turning and right turning, by using a Kalman predictor according to the historical track information of the guided vehicle.
The boot-following sub-process of restarting after parking: after the guided vehicle is parked, the guided vehicle target repositioning and continuous tracking processing flow under the starting condition is as follows:
step 1: judging whether a guiding vehicle target under the millimeter wave radar is lost, if not, turning to the step 2, otherwise, turning to the step 3;
step 2: setting an interested area, and when the guided vehicle moves dynamically, acquiring the dynamic information of the guided vehicle by utilizing a millimeter wave radar, and tracking the dynamic target of the guided vehicle based on the millimeter wave radar;
step 3: and repeating the process of the ground unmanned platform autonomous vehicle tracking target identification and tracking method under the starting condition of the guided vehicle.
The boot-following sub-process under target loss: the target tracking process processing flow under the conditions that the guide vehicle is blocked, the target is temporarily lost and the like is as follows:
step 1: turning to step 2 when millimeter wave radar detection results exist in the region of interest, otherwise turning to step 3;
step 2: continuing to track the target based on the millimeter wave radar;
step 3: judging whether a guiding vehicle target tracking result based on the three-dimensional laser radar exists in the region of interest, taking the result as output, turning to the step 4, otherwise turning to the step 5;
step 4: the target detection result based on the millimeter wave radar is matched with the target detection result based on the three-dimensional laser radar, if the result is matched, the step 2 is switched to, otherwise, the step 3 is switched to;
step 5: predicting target information at the next moment by using a prediction result of a Kalman predictor according to the historical track information, and outputting the prediction information as output;
step 6: searching targets in the whole domain range by using a laser radar, identifying the targets of the guided vehicle, and turning to the step 5 without identification results, otherwise turning to the step 7;
step 7: and repeating the process of the ground unmanned platform autonomous vehicle tracking target identification and tracking method under the starting condition of the guided vehicle.
The foregoing is merely a preferred embodiment of the present invention, and it should be noted that modifications and variations could be made by those skilled in the art without departing from the technical principles of the present invention, and such modifications and variations should also be regarded as being within the scope of the invention.
Claims (9)
1. The method is characterized in that the method is realized by a vehicle-mounted system, the system comprises combined navigation equipment, a three-dimensional laser radar, a millimeter wave radar and a vehicle-mounted industrial personal computer, in the method, point cloud data acquired by the three-dimensional laser radar is utilized for feature extraction and guided vehicle modeling, and the recognition and tracking information initialization of the guided vehicle target is realized; after the guided vehicle runs for a certain distance, after a target tracking result based on the three-dimensional laser radar point cloud data is stable, matching the tracking result of the millimeter wave radar with the tracking result of the three-dimensional laser radar, and switching to target tracking based on the millimeter wave radar; in the continuous tracking process, the gesture information provided by the integrated navigation equipment is used for merging the historical track of target tracking, and the motion trend information of the guided vehicle is calculated by using a Kalman predictor in the vehicle-mounted industrial personal computer;
in the method, the states of target recognition and tracking of the guided vehicle are divided into four working states of target recognition based on a three-dimensional laser radar, target tracking based on the three-dimensional laser radar, target tracking based on a millimeter wave radar and target tracking based on Kalman prediction, and the implementation flow of each working state is as follows:
step 1: starting the guide vehicle;
step 2: entering a target identification state based on a three-dimensional laser radar, guiding the vehicle target to confirm, and turning to the step 3;
step 3: based on the target tracking of the three-dimensional laser radar, the target tracking is effective, and the step 4 is switched, otherwise, the step 7 is switched;
step 4: matching with a target detection result based on the millimeter wave radar, if so, turning to step 5, otherwise, returning to step 3;
step 5: performing millimeter wave radar-based target tracking;
step 6: if the target is valid, returning to the step 5; otherwise, returning to the step 3;
step 7: performing target detection based on Kalman prediction;
step 8: judging the result of target detection based on Kalman prediction, if the result is available, returning to the step 2, otherwise, returning to the step 7;
the process is finished after the ending instruction is received, otherwise, the process is executed circularly all the time.
2. The method of claim 1, wherein both the millimeter wave radar and the three-dimensional lidar are horizontally mounted to the front of the vehicle; the inertial navigation device in the integrated navigation device is arranged in the vehicle body, and the Y-axis direction is parallel to the central axis of the advancing direction of the vehicle body; the GPS/Beidou antenna in the integrated navigation device is arranged outside the vehicle body, so that shielding is avoided; the vehicle-mounted industrial personal computer is arranged in the unmanned platform vehicle body, defines the origin of a local coordinate system of the vehicle body as the center of a rear shaft of the vehicle body, the advancing direction of the vehicle body is Y-axis positive direction, the horizontal right side vertical to the advancing direction of the vehicle body is X-axis positive direction, the horizontal right side vertical to the horizontal plane and the Z-axis positive direction along the sky of the vehicle body.
3. The method of claim 1, wherein before the motion trend information of the guided vehicle is calculated, external calibration parameters of the three-dimensional laser radar, the millimeter wave radar and the unmanned platform are obtained through joint calibration, and based on the external calibration parameters, data of the millimeter wave radar, the three-dimensional laser radar, the GPS/Beidou antenna and the inertial navigation device are unified under a vehicle body local coordinate system by utilizing a formula (1), so that reference coordinate unification of the data is realized;
in formula (1), p sensor ＝(x s ,y s ,z s ) T Representing the coordinates of data in a three-dimensional lidar coordinate system, p vehicle ＝(x v ,y v ,z v ) T Representing the coordinates of the data in the local coordinate system of the vehicle body,for rotating matrix +.>The elements in the rotation matrix and the translation matrix are external calibration parameters.
4. The method according to claim 1, wherein the guiding of the unmanned ground platform is divided into 4 sub-processes of starting guiding-following, guiding-following in a continuous process, guiding-following restarted after stopping, guiding-following under target loss according to the maneuvering process of guiding-following.
5. The method of claim 4, wherein the launch guide-follow sub-process is a process of a ground unmanned platform autonomous vehicle tracking target recognition and tracking method under a guided vehicle launch condition, comprising the steps of:
step 11: the guiding vehicle is statically parked in a certain range in front of the unmanned platform, and a guiding vehicle target recognition and tracking program is started;
step 12: the method comprises the steps of carrying out point cloud data acquisition and feature extraction on a guided vehicle through a three-dimensional laser radar, and identifying a guided vehicle target;
step 13: the target recognition result of the guided vehicle based on the three-dimensional laser radar is stable, and the target initialization is completed;
step 14: starting continuous tracking of a target based on a laser radar, and transforming a detection result of a local coordinate system of a vehicle body into a global coordinate system according to vehicle posture information acquired by the integrated navigation equipment to form a historical tracking track;
step 15: the guiding vehicle dynamically moves, and the unmanned platform is started to guide and advance according to the local position information tracked by the target;
step 16: starting vehicle target tracking based on millimeter wave radar;
step 17: using a Hungary matching algorithm, matching the dynamic target detection result of the millimeter wave radar with the continuous tracking target position information of the laser radar as a basis, and executing step 18, wherein the matching is successful; otherwise, returning to the step 15;
step 18: continuously tracking a vehicle target based on millimeter wave radar;
step 19: estimating the position of the guiding vehicle at the next moment by using a Kalman estimator according to the current target tracking position;
step 110: based on the predicted position, the region of interest is set, and the process returns to step 18.
6. The method of claim 4, wherein the lead-following sub-process in the continuous process is a process of detecting a target state of the lead car in the tracking continuous process and judging a lateral and longitudinal state of the lead car, and specifically comprises the steps of:
step 21: acquiring the relative speed of the guided vehicle by utilizing a millimeter wave radar;
step 22: combining the vehicle speed information provided by the integrated navigation equipment, and calculating to obtain the speed information of the guided vehicle;
step 23: recording the local coordinate position of a guided vehicle target under each frame of data, and transforming a target tracking result under a vehicle body local coordinate system into a global coordinate system according to the vehicle posture information acquired by the integrated navigation equipment to form a history track;
step 24: continuously detecting T periods, and judging the speed state of the guided vehicle by using a Bayesian network, wherein the speed state comprises uniform speed, acceleration, deceleration, scram, parking and starting;
step 25: and predicting the transverse state of the vehicle, including straight running, left turning and right turning, by using a Kalman predictor according to the historical track information of the guided vehicle.
7. The method of claim 5, wherein the post-parking restart guidance-follow-up sub-process is a guidance vehicle target repositioning and follow-up process under a start condition after the guidance vehicle is parked, comprising the steps of:
step 31: judging whether the guiding vehicle target under the millimeter wave radar is lost, if not, executing the step 32, otherwise, executing the step 33;
step 32: setting an interested area, and when the guided vehicle moves dynamically, acquiring the dynamic information of the guided vehicle by utilizing a millimeter wave radar, and tracking the dynamic target of the guided vehicle based on the millimeter wave radar;
step 33: and repeating the flow of the ground unmanned platform autonomous vehicle tracking target identification and tracking method under the starting condition of the guided vehicle.
8. The method of claim 5, wherein the lead-following sub-process under target loss is a target tracking process flow in case of a blocked, short-lived target loss of a lead vehicle, comprising the steps of:
step 41: if the millimeter wave radar detection result exists in the region of interest, executing step 42, otherwise executing step 43;
step 42: continuing to track the target based on the millimeter wave radar;
step 43: judging whether a guiding vehicle target tracking result based on the three-dimensional laser radar exists in the region of interest, and if so, taking the guiding vehicle target tracking result as output, executing a step 44, otherwise, executing a step 45;
step 44: the target detection result based on the millimeter wave radar is matched with the target detection result based on the three-dimensional laser radar, if the result is matched, the step 42 is executed, otherwise, the step 43 is returned to;
step 45: predicting target information at the next moment by using a prediction result of a Kalman predictor according to the historical track information, and outputting the prediction information as output;
step 46: searching targets in the whole domain range by using a laser radar, identifying the targets of the guided vehicle, returning to the step 45 without identification results, and executing the step 47 otherwise;
step 47: and repeating the flow of the ground unmanned platform autonomous vehicle tracking target identification and tracking method under the starting condition of the guided vehicle.
9. An in-vehicle system for use in implementing a method as claimed in any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111122185.XA CN113702967B (en) | 2021-09-24 | 2021-09-24 | Method for identifying and tracking guided vehicle target of ground unmanned platform and vehicle-mounted system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111122185.XA CN113702967B (en) | 2021-09-24 | 2021-09-24 | Method for identifying and tracking guided vehicle target of ground unmanned platform and vehicle-mounted system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113702967A CN113702967A (en) | 2021-11-26 |
| CN113702967B true CN113702967B (en) | 2023-07-28 |
Family
ID=78661892
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111122185.XA Active CN113702967B (en) | 2021-09-24 | 2021-09-24 | Method for identifying and tracking guided vehicle target of ground unmanned platform and vehicle-mounted system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113702967B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609522A (en) * | 2017-09-19 | 2018-01-19 | 东华大学 | A kind of information fusion vehicle detecting system based on laser radar and machine vision |
| JP2018092483A (en) * | 2016-12-06 | 2018-06-14 | トヨタ自動車株式会社 | Object recognition device |
| CN109739266A (en) * | 2018-12-03 | 2019-05-10 | 中国北方车辆研究所 | A kind of follow-up control method of the unmanned shipping platform in ground |
| CN109829386A (en) * | 2019-01-04 | 2019-05-31 | 清华大学 | Intelligent vehicle based on Multi-source Information Fusion can traffic areas detection method |
| CN110175186A (en) * | 2019-05-15 | 2019-08-27 | 中国舰船研究设计中心 | A kind of intelligent ship environmental threat target apperception system and method |
| WO2021077287A1 (en) * | 2019-10-22 | 2021-04-29 | 华为技术有限公司 | Detection method, detection device, and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112859022B (en) * | 2021-01-22 | 2023-08-18 | 上海西井科技股份有限公司 | Method, system, equipment and storage medium for joint calibration of multiple radars and cameras |
-
2021
- 2021-09-24 CN CN202111122185.XA patent/CN113702967B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2018092483A (en) * | 2016-12-06 | 2018-06-14 | トヨタ自動車株式会社 | Object recognition device |
| CN107609522A (en) * | 2017-09-19 | 2018-01-19 | 东华大学 | A kind of information fusion vehicle detecting system based on laser radar and machine vision |
| CN109739266A (en) * | 2018-12-03 | 2019-05-10 | 中国北方车辆研究所 | A kind of follow-up control method of the unmanned shipping platform in ground |
| CN109829386A (en) * | 2019-01-04 | 2019-05-31 | 清华大学 | Intelligent vehicle based on Multi-source Information Fusion can traffic areas detection method |
| CN110175186A (en) * | 2019-05-15 | 2019-08-27 | 中国舰船研究设计中心 | A kind of intelligent ship environmental threat target apperception system and method |
| WO2021077287A1 (en) * | 2019-10-22 | 2021-04-29 | 华为技术有限公司 | Detection method, detection device, and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 基于优化DeepSort的前方车辆多目标跟踪;金立生等;《浙江大学学报(工学版)》;第55卷(第6期);第1056-1064页 * |
| 改进的交互式卡尔曼滤波对雷达数据处理技术研究;刘全周等;《机械科学与技术》;第39卷(第08期);第1248-1255页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113702967A (en) | 2021-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11768286B2 (en) | Method of determining the yaw rate of a target vehicle | |
| JP6432116B2 (en) | Vehicle position specifying device, vehicle control system, vehicle position specifying method, and vehicle position specifying program | |
| CN109358340B (en) | AGV indoor map construction method and system based on laser radar | |
| US11125567B2 (en) | Methods and systems for mapping and localization for a vehicle | |
| CN112154455B (en) | Data processing method, equipment and movable platform | |
| US11945499B2 (en) | Method and apparatus for trailer angle measurement and vehicle | |
| CN114442101B (en) | Vehicle navigation method, device, equipment and medium based on imaging millimeter wave radar | |
| CN109460035A (en) | Second level automatic obstacle avoiding system and barrier-avoiding method under a kind of unmanned boat fast state | |
| CN113124864A (en) | Water surface navigation method adopting machine vision and inertial navigation fusion | |
| US20220035036A1 (en) | Method and apparatus for positioning movable device, and movable device | |
| Babak et al. | Control of autonomous ground vehicles: A brief technical review | |
| CN114383598B (en) | Tunnel construction operation car and automatic driving system thereof | |
| CN113758482B (en) | Vehicle navigation positioning method, device, base station, system and readable storage medium | |
| CN113702967B (en) | Method for identifying and tracking guided vehicle target of ground unmanned platform and vehicle-mounted system | |
| EP4160269A1 (en) | Systems and methods for onboard analysis of sensor data for sensor fusion | |
| CN114379584B (en) | Vehicle control method, device, vehicle, equipment and storage medium | |
| US20230242148A1 (en) | Vehicle controller and vehicle control method | |
| CN114834447B (en) | Collision risk prediction method and device based on automatic parking trajectory | |
| US11353595B2 (en) | Sidelobe subtraction method in automotive radar signal processing | |
| US20230168363A1 (en) | Method to detect radar installation error for pitch angle on autonomous vehicles | |
| US20240051519A1 (en) | Vehicle Localization Based on Pose Corrections from Remote Cameras | |
| JP2023021815A (en) | Object detection device and object detection method | |
| CN113917478A (en) | Tunnel unmanned aerial vehicle positioning system and method based on double single line laser radars | |
| JP2023065112A (en) | Object detection device and object detection method | |
| CN113753079A (en) | Time alignment method used in vehicle state tracking process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |