CN116149338A - Automatic driving control method, system and sprayer - Google Patents
Automatic driving control method, system and sprayer Download PDFInfo
- Publication number
- CN116149338A CN116149338A CN202310397866.XA CN202310397866A CN116149338A CN 116149338 A CN116149338 A CN 116149338A CN 202310397866 A CN202310397866 A CN 202310397866A CN 116149338 A CN116149338 A CN 116149338A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- information
- control
- reinforcement learning
- learning network
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000002787 reinforcement Effects 0.000 claims abstract description 52
- 230000004927 fusion Effects 0.000 claims abstract description 12
- 230000000007 visual effect Effects 0.000 claims abstract description 6
- 238000012549 training Methods 0.000 claims description 26
- 230000008569 process Effects 0.000 claims description 14
- 230000002159 abnormal effect Effects 0.000 claims description 12
- 230000007613 environmental effect Effects 0.000 claims description 11
- 238000001914 filtration Methods 0.000 claims description 9
- 238000004891 communication Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 230000010485 coping Effects 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 4
- 238000005507 spraying Methods 0.000 description 4
- 239000000575 pesticide Substances 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000000881 depressing effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000003337 fertilizer Substances 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000000994 depressogenic effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000002922 simulated annealing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
- G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0257—Control of position or course in two dimensions specially adapted to land vehicles using a radar
-
- 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)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Automation & Control Theory (AREA)
- Aviation & Aerospace Engineering (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Electromagnetism (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention provides an automatic driving control method, an automatic driving control system and a sprayer, and relates to the technical field of automatic driving, wherein the method comprises the following steps: acquiring environment information of a vehicle acquired by a visual device, first positioning information of the vehicle acquired by a GNSS device and second positioning information of the vehicle acquired by an INS device; determining an error between the first positioning information and the second positioning information, estimating the error by adopting an information fusion algorithm, and determining an optimal estimated value of the error; correcting the second positioning information according to the optimal estimated value to obtain combined navigation information; inputting the environment information and the combined navigation information into a trained reinforcement learning network, and outputting a control decision, wherein the trained reinforcement learning network is trained in a mode of combining simulated learning and reinforcement learning; and controlling the vehicle to run according to the control decision. The invention solves the problems of poor coping ability and poor vehicle safety of the automatic driving method of the vehicle in the prior art in complex road conditions.
Description
Technical Field
The invention relates to the technical field of automatic driving, in particular to an automatic driving control method, an automatic driving control system and a sprayer.
Background
With the continuous development of artificial intelligence technology, automatic driving is receiving more and more attention. For example, for plant protection vehicles such as pesticide spraying machines and spraying machines, the application of the automatic driving technology can avoid the influence of pesticide spraying, chemical fertilizer spraying and the like on the health of a driver, and can reduce the labor cost.
Currently, automatic driving for plant protection vehicles and the like mainly controls the vehicles to run according to a planned path so as to complete preset plant protection work. However, the method has high requirements on the accuracy of the planned path, and the vehicle is difficult to cope with complex road condition environments, for example, the vehicle is difficult to cope with the situation that the planned path is provided with an obstacle, so that the safety is poor.
Disclosure of Invention
The invention solves the problem of how to improve the coping capability of automatic driving of the vehicle in complex road conditions and improve the safety of the vehicle.
In order to solve the above problems, the present invention provides an automatic driving control method, an automatic driving control system and a sprayer.
In a first aspect, the present invention provides an automatic driving control method, including:
acquiring the environment information of a vehicle acquired by a visual device, the first positioning information of the vehicle acquired by a GNSS device and the second positioning information of the vehicle acquired by an INS device;
determining an error between the first positioning information and the second positioning information, estimating the error by adopting an information fusion algorithm, and determining an optimal estimated value of the error;
correcting the second positioning information according to the optimal estimated value to obtain combined navigation information;
inputting the environment information and the combined navigation information into a trained reinforcement learning network, and outputting a control decision, wherein the trained reinforcement learning network is trained in a mode of combining simulated learning and reinforcement learning;
and controlling the vehicle to run according to the control decision.
Optionally, the estimating the error by using an information fusion algorithm, and determining an optimal estimated value of the error includes:
taking an error equation of the INS device as a state equation of the integrated navigation system, and taking the error as an observed quantity to construct a state space equation of the integrated navigation system;
and determining the optimal estimated value at the current moment according to the second positioning information at the last moment and the state space equation based on a Kalman filtering algorithm.
Optionally, before the inputting the environmental information and the integrated navigation information into the trained reinforcement learning network, the method further includes:
acquiring training data, wherein the training data comprises vehicle control data of drivers under different environment information and different integrated navigation information;
pre-training the simulated learning network by adopting the training data, copying a convolution layer of the simulated learning network after the pre-training to the reinforcement learning network, and obtaining a processed reinforcement learning network;
outputting a control decision through the processed reinforcement learning network to control the vehicle to run in different environments;
when the vehicle normally runs, the obtained first vehicle control data in different environments are put into a conventional driving experience playback pool; when the vehicle is abnormal in running, obtaining second vehicle control data obtained when an operator performs teleoperation on the vehicle, and placing the second vehicle control data into an abnormal driving experience playback pool;
and randomly sampling training data from the regular driving experience playback pool and the abnormal driving experience playback pool, and training the processed reinforcement learning network to obtain the trained reinforcement learning network.
Optionally, the method further comprises:
acquiring a planned path of the vehicle;
judging whether the vehicle follows the planned path in the running process according to the integrated navigation information;
if not, outputting deviation path information to prompt an operator, and controlling the vehicle to follow the planned path by teleoperation through the operator.
Optionally, the acquiring the planned path of the vehicle includes:
acquiring a working area of the vehicle;
the planned path of the vehicle in the work area is determined based on a path planning algorithm.
Optionally, the control decision comprises at least one of a steering wheel angle, an accelerator opening, a brake opening, a clutch opening and a gear of the vehicle.
In a second aspect, the present invention provides an autopilot control system, including an environmental awareness module, a decision control module, and a bottom layer control module, where the environmental awareness module and the bottom layer control module are respectively electrically connected to the decision control module, and the decision control module is configured to implement the autopilot control method according to any one of the first aspects;
the environment sensing module is used for collecting environment information and positioning information of the vehicle;
the floor control module is used for controlling at least one of steering wheel, accelerator, brake, clutch and gear of the vehicle.
Optionally, the environment sensing module comprises a multi-path binocular camera, a GNSS device and an INS device;
the bottom layer control module comprises a steering system for driving a steering wheel of a vehicle to rotate, an accelerator system for driving an accelerator pedal of the vehicle, a brake system for driving a brake pedal of the vehicle, a clutch system for driving a clutch pedal of the vehicle and a gear shifting system for driving a gear handle of the vehicle.
Optionally, the system further comprises a teleoperation module, wherein the teleoperation module is in communication connection with the decision control module, and the teleoperation module is used for operating personnel to teleoperate the vehicle.
In a third aspect, the present invention provides a sprayer comprising an autopilot control system as defined in any one of the second aspects.
The automatic driving control method, the automatic driving control system and the sprayer provided by the invention have the beneficial effects that: the first positioning information and the second positioning information of the vehicle can be acquired in real time through a visual device mounted on the vehicle and a GNSS device and an INS device mounted on the vehicle respectively. And determining the error between the first positioning information and the second positioning information by comparison, and estimating the error by adopting an information fusion algorithm to determine the optimal estimated value of the error. And then correcting the second positioning information by using the optimal estimated value to inhibit error divergence of the second positioning information acquired by the INS device, and taking the corrected second positioning information as combined navigation information, thereby improving the positioning accuracy of the vehicle. The environment information and the combined navigation information are input into the trained reinforcement learning network, the reinforcement learning network is trained by combining simulation learning and reinforcement learning, the trained reinforcement learning network can make proper control decisions like a driver according to surrounding environments, the vehicle can be controlled to avoid obstacles according to various complex road conditions, the capability of automatic driving of the vehicle for coping with the complex road conditions is improved, and then the safety of the vehicle in the driving process is improved.
Drawings
Fig. 1 is a schematic structural diagram of an autopilot control system according to an embodiment of the present invention;
FIG. 2 is a mechanical block diagram of an autopilot control system according to one embodiment of the present invention;
FIG. 3 is another angular mechanical block diagram of an autopilot control system provided in accordance with an embodiment of the present invention;
fig. 4 is a front view of an autopilot control system according to an embodiment of the present invention;
FIG. 5 is a left side view of an autopilot control system according to one embodiment of the present invention;
fig. 6 is a schematic flow chart of an automatic driving control method according to an embodiment of the present invention;
FIG. 7 is a schematic diagram illustrating a combined navigation process of a GNSS device and an INS device according to an embodiment of the present invention;
fig. 8 is a schematic structural diagram of an autopilot control apparatus according to an embodiment of the present invention.
Reference numerals illustrate:
1. a bracket; 11. a fixed rod; 12. an adjusting rod; 21. a multi-view camera; 31. steering wheel clamp; 32. a steering wheel drive motor; 41. a first pedal; 42. a first motor; 51. a second pedal; 52 a second motor; 61. a third pedal; a third motor 62; 71. a gear shifting system; 72. a high-low gear system; 80. a case; 90. and a power supply.
Detailed Description
In order that the above objects, features and advantages of the invention will be readily understood, a more particular description of the invention will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings. While the invention is susceptible of embodiment in the drawings, it is to be understood that the invention may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided to provide a more thorough and complete understanding of the invention. It should be understood that the drawings and embodiments of the invention are for illustration purposes only and are not intended to limit the scope of the present invention.
In the coordinate system XYZ provided herein, the positive direction of the X axis represents the right direction, the negative direction of the X axis represents the left direction, the positive direction of the Y axis represents the rear direction, the negative direction of the Y axis represents the front direction, the positive direction of the Z axis represents the upper direction, and the negative direction of the Z axis represents the lower direction.
It should be understood that the various steps recited in the method embodiments of the present invention may be performed in a different order and/or performed in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the invention is not limited in this respect.
The term "including" and variations thereof as used herein are intended to be open-ended, i.e., including, but not limited to. The term "based on" is based at least in part on. The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "alternative embodiments". Related definitions of other terms will be given in the description below. It should be noted that the terms "first," "second," and the like herein are merely used for distinguishing between different devices, modules, or units and not for limiting the order or interdependence of the functions performed by such devices, modules, or units.
It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those skilled in the art will appreciate that "one or more" is intended to be construed as "one or more" unless the context clearly indicates otherwise.
The names of messages or information interacted between the devices in the embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of such messages or information.
As shown in fig. 1 to 5, an autopilot control system provided by an embodiment of the present invention includes an environment sensing module, a decision control module, and a bottom layer control module, where the environment sensing module and the bottom layer control module are respectively electrically connected to the decision control module, and the decision control module is configured to implement an autopilot control method as described below;
the environment sensing module is used for collecting environment information and positioning information of the vehicle;
the floor control module is used for controlling at least one of steering wheel, accelerator, brake, clutch and gear of the vehicle.
Specifically, the decision control module may include a decision control computer, an NVR (Network Video Recorder ) and a 4G router, and the environment sensing module includes a vision device, a GNSS device and an INS device, where the vision device may include a multi-path binocular camera, and the multi-path binocular camera may be used to collect environmental information of a farmland around the vehicle, and may be connected to the NVR through a network port, and then transmitted to the router and the decision control computer by the NVR. The GNSS device may employ GPS (GlobalPositioning System ) or beidou navigation system, etc., and the GPS and INS (Inertial Navigation System ) are used for integrated navigation, and may be directly connected to the decision control computer. The decision control computer is used for coordinating each module to receive, process, distribute and stream the data.
Optionally, the system further comprises a teleoperation module, wherein the teleoperation module is in communication connection with the decision control module, and the teleoperation module is used for operating personnel to teleoperate the vehicle.
Specifically, the teleoperation module may include a remote computer, which may be connected to the 4G router of the decision control module through a wireless communication manner, the decision control module transmits information such as an image to the remote computer, an operator may send a control instruction through a user interface of the remote computer, transmit the control instruction to the vehicle router through a wireless network, access the NVR through a network port, and transmit the control instruction to the decision control computer through an RS-232/485 serial port.
Optionally, the floor control module includes a steering system that drives the vehicle steering wheel in rotation, a throttle system that drives the vehicle throttle pedal, a brake system that drives the vehicle brake pedal, a clutch system that drives the vehicle clutch pedal, and a shift system 71 that drives the vehicle gear handle.
Specifically, the decision control module outputs a control instruction to the bottom layer control module, and the vehicle is driven by the bottom layer control module. The steering system is used for controlling the steering wheel of the vehicle to rotate according to the control instruction, the accelerator system is used for controlling the accelerator pedal of the vehicle to be stepped on or lifted up according to the control instruction, and the brake system is used for controlling the brake pedal of the vehicle to be stepped on or lifted up according to the control instruction; for a manual transmission vehicle, a clutch system for controlling the depression or lifting of a clutch pedal of the vehicle according to a control command and a gear shifting system 71 for controlling the gear of the vehicle according to the control command may be further included.
For example, the autopilot control system may include: a bracket 1 for connection with a vehicle; the steering system may include a steering wheel clamp 31 connected to the bracket 1, the steering wheel clamp 31 being for clamping a steering wheel of the vehicle and rotating the steering wheel under the drive of a steering wheel driving motor 32; the accelerator system may include a first pedal 41 for abutting against an accelerator pedal of the vehicle, the first pedal 41 depressing or releasing the accelerator pedal of the vehicle under the drive of a first motor 42; the brake system may include a second pedal 51 for abutting against a brake pedal of the vehicle, the second pedal 51 being depressed or released by the second motor 52; the clutch system may include a third pedal 61 for abutting against a clutch pedal of the vehicle, the third pedal 61 depressing or releasing the clutch pedal by the third motor 62; the shift system 71 may include a first drive mechanism for driving movement of the shift knob, the first drive mechanism driving lateral or longitudinal movement of the shift knob to effect a change in vehicle gear; the under-layer control module may also include a high-low gear system 72, the high-low gear system 72 being used to control the shift between the high and low gears of the vehicle.
The vision device may be mounted on top of the stand 1 and the GNSS device, INS device and decision control module may be mounted in a box 80 connected to the stand 1. The bracket 1 may include a fixing rod 11 and an adjusting rod 12, the adjusting rod 12 is slidably connected with the fixing rod 11, one end of the fixing rod 11 is used for being connected with a vehicle, and the adjusting rod 12 can be slid along an extension direction of the fixing rod 11 and can be fixed at an arbitrary position to increase or shorten a length of the bracket 1. The vision device may be mounted on the adjustment lever 12 and the case 80 may be mounted on the fixing lever 11. A power supply 90 for powering the various devices may also be included.
In this embodiment, the environmental information and the positioning information of the vehicle are collected by the environmental sensing module, the decision control module processes the environmental information and the positioning information, determines a control decision, and controls the vehicle by the bottom layer control module according to the control decision so as to realize automatic driving. The manual operation is not needed, the labor cost is reduced, the influence of chemical fertilizers, pesticides and the like on the health of a driver during the operation of driving the plant protection vehicle is avoided, and the safety is improved.
As shown in fig. 6, an automatic driving control method provided by an embodiment of the present invention includes:
step S100, acquiring the environment information of the vehicle acquired by the vision device, the first positioning information of the vehicle acquired by the GNSS device and the second positioning information of the vehicle acquired by the INS device.
Specifically, the vision device can include a camera, a laser radar and other sensors, and the environment information around the vehicle is collected through the vision device. The GNSS device and the INS device are used for positioning the vehicle in real time, and the GNSS device can directly determine first positioning information of the vehicle and can comprise the position and the speed of the vehicle. The accelerometer of the INS device collects acceleration of the vehicle, the gyroscope of the INS device collects angular velocity of the vehicle, and the second positioning information of the vehicle can be obtained by settling according to the acceleration and the angular velocity through an INS resolving algorithm, and a specific resolving process is the prior art and is not described herein.
Step 200, determining an error between the first positioning information and the second positioning information, estimating the error by adopting an information fusion algorithm, and determining an optimal estimated value of the error.
And step S300, correcting the second positioning information according to the optimal estimated value to obtain combined navigation information.
Specifically, the information fusion algorithm can adopt a Kalman filtering algorithm, an optimal estimated value of the error is determined through the information fusion algorithm, the second positioning information of the INS device is corrected by utilizing the optimal estimated value, the error divergence of the second positioning information acquired by the INS device can be restrained, the accuracy of the combined navigation information is improved, and the combined navigation information can comprise the real-time position and the real-time speed of the vehicle.
And step S400, inputting the environment information and the combined navigation information into a trained reinforcement learning network, and outputting a control decision, wherein the trained reinforcement learning network is trained by combining simulated learning and reinforcement learning.
Specifically, the reinforcement learning network may be trained in advance by combining the imitation learning and reinforcement learning, the environment information and the combined navigation information are input into the trained reinforcement learning network, and a control decision for the vehicle is determined, and the control decision may include a steering angle and a speed of the vehicle, and the like.
And step S500, controlling the vehicle to run according to the control decision.
Specifically, the bottom layer control module is controlled according to the control decision to control the steering wheel, the accelerator, the brake, the clutch, the gear and the like of the vehicle, so that the automatic driving of the vehicle is realized.
In this embodiment, the visual device mounted on the vehicle may collect the environmental information around the vehicle, and the GNSS device and the INS device mounted on the vehicle may collect the first positioning information and the second positioning information of the vehicle in real time, respectively. And determining the error between the first positioning information and the second positioning information by comparison, and estimating the error by adopting an information fusion algorithm to determine the optimal estimated value of the error. And then correcting the second positioning information by using the optimal estimated value to inhibit error divergence of the second positioning information acquired by the INS device, and taking the corrected second positioning information as combined navigation information, thereby improving the positioning accuracy of the vehicle. The environment information and the combined navigation information are input into the trained reinforcement learning network, the reinforcement learning network is trained by combining simulation learning and reinforcement learning, the trained reinforcement learning network can make proper control decisions like a driver according to surrounding environments, the vehicle can be controlled to avoid obstacles according to various complex road conditions, the capability of automatic driving of the vehicle for coping with the complex road conditions is improved, and then the safety of the vehicle in the driving process is improved.
Optionally, as shown in fig. 7, the estimating the error by using an information fusion algorithm, determining an optimal estimated value of the error includes:
taking an error equation of the INS device as a state equation of the integrated navigation system, and taking the error as an observed quantity to construct a state space equation of the integrated navigation system;
and determining the optimal estimated value at the current moment according to the second positioning information at the last moment and the state space equation based on a Kalman filtering algorithm.
Specifically, by comprehensively considering the requirements of the precision and the real-time performance of the integrated navigation system, attitude angle errors, speed errors, position errors, accelerometer constant drift and gyroscope constant drift can be selected as state variables. And constructing a state equation of the integrated navigation system according to the error equation of the INS device, taking the errors of the first positioning information and the second positioning information as the observed quantity of the integrated navigation system, and obtaining a discretization state space equation of the integrated navigation system.
The Kalman filtering is a filtering algorithm based on minimum variance estimation, and the state parameter updating at the current moment is completed by using the state estimation value at the previous moment and the observed quantity at the current moment. The Kalman filtering only needs to record the state parameter of the last moment, and the past history data is not needed to be saved after the state updating of the current moment is completed, so that the Kalman filtering has higher efficiency in carrying out state estimation, and the essence is a recursive algorithm. The time updating and the measurement updating form the basis of Kalman filtering, the state value of the current moment is estimated according to the measurement value of the previous moment in the time updating stage, the state value is corrected according to the measurement value of the current moment in the measurement updating stage, and the specific process is not repeated here.
In this optional embodiment, the second positioning information is corrected by using the optimal estimated value of the error between the first positioning information acquired by the GNSS device and the second positioning information acquired by the INS device, so as to realize integrated navigation, avoid the error caused by positioning the single device, and improve the positioning accuracy.
Optionally, before the inputting the environmental information and the integrated navigation information into the trained reinforcement learning network, the method further includes:
acquiring training data, wherein the training data comprises vehicle control data of drivers under different environment information and different integrated navigation information;
pre-training the simulated learning network by adopting the training data, copying a convolution layer of the simulated learning network after the pre-training to the reinforcement learning network, and obtaining a processed reinforcement learning network;
outputting a control decision through the processed reinforcement learning network to control the vehicle to run in different environments;
when the vehicle normally runs, the obtained first vehicle control data in different environments are put into a conventional driving experience playback pool; when the vehicle is abnormal in running, obtaining second vehicle control data obtained when an operator performs teleoperation on the vehicle, and placing the second vehicle control data into an abnormal driving experience playback pool;
and randomly sampling training data from the regular driving experience playback pool and the abnormal driving experience playback pool, and training the processed reinforcement learning network to obtain the trained reinforcement learning network.
Specifically, when the data in the regular driving experience playback pool and the abnormal driving experience playback pool reach the upper limit of capacity, the training degree of the data can be obtained through calculating the reinforcement learning model through an algorithm, and the data with high training score is selected to be deleted. And for experiences with low training scores and incomplete mastery of the model, the experience is kept in a corresponding experience playback pool for subsequent sampling training of the model.
In this alternative embodiment, the reinforcement learning network is trained by combining simulated learning and reinforcement learning, such that the trained reinforcement learning network is able to make decisions like a driver. And the reinforcement learning network is trained by using the data in the conventional driving experience playback pool and the abnormal driving experience playback pool, so that the coping capability of the trained reinforcement learning network for various different environments can be improved.
Optionally, the method further comprises:
acquiring a planned path of the vehicle;
judging whether the vehicle follows the planned path in the running process according to the integrated navigation information;
if not, outputting deviation path information to prompt an operator, and controlling the vehicle to follow the planned path by teleoperation through the operator.
In particular, a planned path of the vehicle is obtained, which may characterize a work route of the vehicle in the work area. Judging whether the vehicle deviates from the planned path for a long time in the driving process according to the integrated navigation information, recording the time of the vehicle deviating from the planned path, comparing the time with a preset threshold value, and when the time is smaller than the preset threshold value, indicating that the vehicle may deviate from the planned path for a short time due to obstacle avoidance and other reasons. When the time is greater than or equal to a preset threshold value, the vehicle is represented to have a long-term cheap planning path, and the abnormal running condition occurs, and at the moment, deviation path information can be output to prompt an operator, and the operator can control the vehicle to return to the planning path through teleoperation.
Inputting the environment information and the integrated navigation information into a trained reinforcement learning network, wherein the trained reinforcement learning network can identify obstacles in a vehicle driving route, and generate a control decision according to the collected environment information and positioning information, and the control decision can control the vehicle to drive according to a new driving route so as to avoid the obstacles and return to a planning path after avoiding the obstacles.
Because the farmland environment is complex, when the plant protection vehicle encounters a complex and uncertain operation area environment in the automatic driving process and cannot make an independent decision, an operator can know the operation state of the plant protection vehicle in a teleoperation mode and automatically decide whether to manually control. In the vehicle control process, operators are always endowed with the highest decision authority, so that the reliability of the whole system is enhanced. In the teleoperation control process, an operator outputs control information to control a bottom layer control module according to judgment decision of the operator by means of environment information returned by the visual device and positioning information returned by the GNSS device and the INS positioning device, so as to control a steering wheel, an accelerator, a brake and the like of the vehicle.
Optionally, the acquiring the planned path of the vehicle includes:
acquiring a working area of the vehicle;
the planned path of the vehicle in the work area is determined based on a path planning algorithm.
Specifically, the planned path of the vehicle in the working area may be determined manually by a person, or may be determined by a path planning algorithm, which may include a simulated annealing algorithm, a fuzzy logic algorithm, an ant colony algorithm, a genetic algorithm, and the like.
Optionally, the control decision comprises at least one of a steering wheel angle, an accelerator opening, a brake opening, a clutch opening and a gear of the vehicle.
As shown in fig. 8, an autopilot control apparatus provided in an embodiment of the present invention includes:
the acquisition module is used for acquiring the environment information of the vehicle acquired by the vision device, the first positioning information of the vehicle acquired by the GNSS device and the second positioning information of the vehicle acquired by the INS device;
the positioning module is used for determining errors between the first positioning information and the second positioning information, estimating the errors by adopting an information fusion algorithm and determining an optimal estimated value of the errors; correcting the second positioning information according to the optimal estimated value to obtain combined navigation information;
the decision module is used for inputting the environment information and the combined navigation information into a trained reinforcement learning network and outputting a control decision, wherein the trained reinforcement learning network is obtained by training in a mode of combining imitation learning and reinforcement learning;
and the control module is used for controlling the vehicle to run according to the control decision.
The advantage of the autopilot control apparatus of the present embodiment with respect to the prior art is the same as that of the autopilot control method described above, and is not described in detail herein.
The embodiment of the invention provides a sprayer, which comprises the automatic driving control system.
Those skilled in the art will appreciate that implementing all or part of the above-described methods in accordance with the embodiments may be accomplished by way of a computer program stored on a computer readable storage medium, which when executed may comprise the steps of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a random access Memory (RandomAccess Memory, RAM), or the like. In this application, the units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment of the present invention. In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
Although the present disclosure is disclosed above, the scope of the present disclosure is not limited thereto. Various changes and modifications may be made by one skilled in the art without departing from the spirit and scope of the disclosure, and these changes and modifications will fall within the scope of the disclosure.
Claims (10)
1. An automatic driving control method, characterized by comprising:
acquiring the environment information of a vehicle acquired by a visual device, the first positioning information of the vehicle acquired by a GNSS device and the second positioning information of the vehicle acquired by an INS device;
determining an error between the first positioning information and the second positioning information, estimating the error by adopting an information fusion algorithm, and determining an optimal estimated value of the error;
correcting the second positioning information according to the optimal estimated value to obtain combined navigation information;
inputting the environment information and the combined navigation information into a trained reinforcement learning network, and outputting a control decision, wherein the trained reinforcement learning network is trained in a mode of combining simulated learning and reinforcement learning;
and controlling the vehicle to run according to the control decision.
2. The automatic driving control method according to claim 1, wherein the estimating the error using an information fusion algorithm, determining an optimal estimated value of the error includes:
taking an error equation of the INS device as a state equation of the integrated navigation system, and taking the error as an observed quantity to construct a state space equation of the integrated navigation system;
and determining the optimal estimated value at the current moment according to the second positioning information at the last moment and the state space equation based on a Kalman filtering algorithm.
3. The automatic driving control method according to claim 1, characterized by further comprising, before the inputting the environmental information and the integrated navigation information into the trained reinforcement learning network:
acquiring training data, wherein the training data comprises vehicle control data of drivers under different environment information and different integrated navigation information;
pre-training the simulated learning network by adopting the training data, copying a convolution layer of the simulated learning network after the pre-training to the reinforcement learning network, and obtaining a processed reinforcement learning network;
outputting a control decision through the processed reinforcement learning network to control the vehicle to run in different environments;
when the vehicle normally runs, the obtained first vehicle control data in different environments are put into a conventional driving experience playback pool; when the vehicle is abnormal in running, obtaining second vehicle control data obtained when an operator performs teleoperation on the vehicle, and placing the second vehicle control data into an abnormal driving experience playback pool;
and randomly sampling training data from the regular driving experience playback pool and the abnormal driving experience playback pool, and training the processed reinforcement learning network to obtain the trained reinforcement learning network.
4. The automatic driving control method according to claim 1, characterized by further comprising:
acquiring a planned path of the vehicle;
judging whether the vehicle follows the planned path in the running process according to the integrated navigation information;
if not, outputting deviation path information to prompt an operator, and controlling the vehicle to follow the planned path by teleoperation through the operator.
5. The automatic driving control method according to claim 4, characterized in that the acquiring the planned path of the vehicle includes:
acquiring a working area of the vehicle;
the planned path of the vehicle in the work area is determined based on a path planning algorithm.
6. The automatic driving control method according to any one of claims 1 to 5, characterized in that the control decision includes at least one of a steering wheel angle, an accelerator opening, a brake opening, a clutch opening, and a gear of the vehicle.
7. An automatic driving control system, characterized by comprising an environment sensing module, a decision control module and a bottom layer control module, wherein the environment sensing module and the bottom layer control module are respectively and electrically connected with the decision control module, and the decision control module is used for realizing the automatic driving control method according to any one of claims 1 to 6;
the environment sensing module is used for collecting environment information and positioning information of the vehicle;
the floor control module is used for controlling at least one of steering wheel, accelerator, brake, clutch and gear of the vehicle.
8. The autopilot control system of claim 7 wherein the environmental awareness module includes a multi-path binocular camera, a GNSS device and an INS device;
the bottom layer control module comprises a steering system for driving a steering wheel of a vehicle to rotate, an accelerator system for driving an accelerator pedal of the vehicle, a brake system for driving a brake pedal of the vehicle, a clutch system for driving a clutch pedal of the vehicle and a gear shifting system for driving a gear handle of the vehicle.
9. The autopilot control system of claim 7 further comprising a teleoperation module in communication with the decision control module, the teleoperation module for teleoperation of the vehicle by an operator.
10. A sprayer comprising an autopilot control system as claimed in any one of claims 7 to 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310397866.XA CN116149338A (en) | 2023-04-14 | 2023-04-14 | Automatic driving control method, system and sprayer |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310397866.XA CN116149338A (en) | 2023-04-14 | 2023-04-14 | Automatic driving control method, system and sprayer |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116149338A true CN116149338A (en) | 2023-05-23 |
Family
ID=86354598
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310397866.XA Pending CN116149338A (en) | 2023-04-14 | 2023-04-14 | Automatic driving control method, system and sprayer |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116149338A (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107728175A (en) * | 2017-09-26 | 2018-02-23 | 南京航空航天大学 | The automatic driving vehicle navigation and positioning accuracy antidote merged based on GNSS and VO |
CN109521454A (en) * | 2018-12-06 | 2019-03-26 | 中北大学 | A kind of GPS/INS Combinated navigation method based on self study volume Kalman filtering |
CN110673602A (en) * | 2019-10-24 | 2020-01-10 | 驭势科技(北京)有限公司 | Reinforced learning model, vehicle automatic driving decision method and vehicle-mounted equipment |
CN111521187A (en) * | 2020-05-13 | 2020-08-11 | 北京百度网讯科技有限公司 | Combined navigation method, device, equipment and storage medium |
CN111947681A (en) * | 2020-06-24 | 2020-11-17 | 中铁第四勘察设计院集团有限公司 | Filtering correction method for GNSS and inertial navigation combined navigation position output |
CN114216459A (en) * | 2021-12-08 | 2022-03-22 | 昆山九毫米电子科技有限公司 | ELM-assisted GNSS/INS integrated navigation unmanned target vehicle positioning method |
CN114282433A (en) * | 2021-12-15 | 2022-04-05 | 中国科学院深圳先进技术研究院 | Automatic driving training method and system based on combination of simulation learning and reinforcement learning |
CN114358128A (en) * | 2021-12-06 | 2022-04-15 | 深圳先进技术研究院 | Method for training end-to-end automatic driving strategy |
-
2023
- 2023-04-14 CN CN202310397866.XA patent/CN116149338A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107728175A (en) * | 2017-09-26 | 2018-02-23 | 南京航空航天大学 | The automatic driving vehicle navigation and positioning accuracy antidote merged based on GNSS and VO |
CN109521454A (en) * | 2018-12-06 | 2019-03-26 | 中北大学 | A kind of GPS/INS Combinated navigation method based on self study volume Kalman filtering |
CN110673602A (en) * | 2019-10-24 | 2020-01-10 | 驭势科技(北京)有限公司 | Reinforced learning model, vehicle automatic driving decision method and vehicle-mounted equipment |
CN111521187A (en) * | 2020-05-13 | 2020-08-11 | 北京百度网讯科技有限公司 | Combined navigation method, device, equipment and storage medium |
CN111947681A (en) * | 2020-06-24 | 2020-11-17 | 中铁第四勘察设计院集团有限公司 | Filtering correction method for GNSS and inertial navigation combined navigation position output |
CN114358128A (en) * | 2021-12-06 | 2022-04-15 | 深圳先进技术研究院 | Method for training end-to-end automatic driving strategy |
CN114216459A (en) * | 2021-12-08 | 2022-03-22 | 昆山九毫米电子科技有限公司 | ELM-assisted GNSS/INS integrated navigation unmanned target vehicle positioning method |
CN114282433A (en) * | 2021-12-15 | 2022-04-05 | 中国科学院深圳先进技术研究院 | Automatic driving training method and system based on combination of simulation learning and reinforcement learning |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104908811B (en) | Communicating messages via a vehicle steering wheel | |
US10882522B2 (en) | Systems and methods for agent tracking | |
EP3588226B1 (en) | Method and arrangement for generating control commands for an autonomous road vehicle | |
US20190220737A1 (en) | Method of generating training data for training a neural network, method of training a neural network and using neural network for autonomous operations | |
CN111301425B (en) | Efficient optimal control using dynamic models for autonomous vehicles | |
CN109964188A (en) | Control the method and system of vehicle | |
US11110917B2 (en) | Method and apparatus for interaction aware traffic scene prediction | |
GB2610097A (en) | Vehicle control system and control method | |
JP6856575B2 (en) | Control policy learning and vehicle control method based on reinforcement learning without active search | |
US20200042656A1 (en) | Systems and methods for persistent simulation | |
US20070088469A1 (en) | Vehicle control system and method | |
CN110406530A (en) | A kind of automatic Pilot method, apparatus, equipment and vehicle | |
CA2616613A1 (en) | Guidance, navigation, and control system for a vehicle | |
US20200089244A1 (en) | Experiments method and system for autonomous vehicle control | |
CN111752274A (en) | Laser AGV path tracking control method based on reinforcement learning | |
Min et al. | Design and implementation of an intelligent vehicle system for autonomous valet parking service | |
CN110356410A (en) | Vehicle control after movement adjusting | |
CN116149338A (en) | Automatic driving control method, system and sprayer | |
Christiansen et al. | Collaborative model based development of adaptive controller settings for a load-carrying vehicle with changing loads | |
Björnberg | Shared control for vehicle teleoperation with a virtual environment interface | |
CN113960921A (en) | Visual navigation control method and system for orchard tracked vehicle | |
JP2021002397A (en) | Work management system | |
Heinrich et al. | A new control architecture for MuCAR | |
KR102262670B1 (en) | Method, apparatus and system for automatic driving of catapillar working vehicle | |
Christiansen | Co-modelling of Agricultural Robotic Systems |
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 |