Anti-collision method and system for unmanned agricultural machine operation

Abstract

The invention discloses an anti-collision method and system for unmanned agricultural machine operation, and the method comprises the steps: S1, collecting the information of a target unmanned agricultural machine and the information of a surrounding perceptual object in real time through a sensing device preset on the target unmanned agricultural machine, and transmitting the information to a cloud end; s2, based on the data collected in the step S1, the cloud side calculates the critical safety distance between the target unmanned agricultural machine and the surrounding vehicles, compares the critical safety distance with a preset safety threshold value, and judges whether a collision risk exists or not; s3, based on the data collected in the step S1, the vehicle end of the target unmanned agricultural machine analyzes the collision risk between the target unmanned agricultural machine and the surrounding perceptual objects in real time by constructing a safety field model; and S4, when any one of the cloud end or the vehicle end judges that a collision risk exists, an anti-collision processing mechanism is started immediately, and deceleration, parking or path adjustment operation is executed. According to the invention, the operation safety and efficiency of the unmanned agricultural machine in a complex farmland environment can be obviously improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned agricultural machines, and particularly relates to a collision avoidance method and system for unmanned agricultural machine operation. BACKGROUND

[0002] With the rapid development of agricultural automation technology, unmanned agricultural machines are increasingly popular in modern agriculture. Unmanned agricultural machines achieve precise operation through automatic driving technology, which can significantly improve agricultural production efficiency and reduce labor costs. However, in the scenario of multi-machine collaborative operation, different types of unmanned agricultural machines operate simultaneously, which poses a high risk of collision, posing a challenge to the safety and efficiency of operation.

[0003] Traditional collision avoidance methods mainly rely on perception technology (such as laser radar, camera, etc.) and simple algorithms to determine the safe distance between vehicles. However, in complex farmland environments, these perception devices may be affected by obstacles, weather conditions, etc., resulting in inaccurate data and limited algorithm accuracy, which cannot effectively predict and avoid potential collision risks. SUMMARY

[0004] The present application aims to provide a collision avoidance method and system for unmanned agricultural machine operation, aiming to improve the safety and efficiency of unmanned agricultural machines in complex farmland environments.

[0005] To achieve the above-mentioned purpose, the present application provides a collision avoidance method for unmanned agricultural machine operation, comprising:

[0006] S1, collecting real-time information of the target unmanned agricultural machine and surrounding perception objects through a preset sensing device on the target unmanned agricultural machine, and sending the information to the cloud;

[0007] S2, based on the data collected in step S1, the cloud calculates the critical safety distance between the target unmanned agricultural machine and the surrounding vehicles, and compares the critical safety distance with the preset safety threshold to determine whether there is a collision risk;

[0008] S3, based on the data collected in step S1, the vehicle end of the target unmanned agricultural machine constructs a safety field model to analyze the collision risk between the target unmanned agricultural machine and the surrounding perception objects in real time; wherein the safety field model includes the superposition of potential field and kinetic field, the potential field is used to represent the risk of static obstacles, and the kinetic field is used to represent the risk of dynamic obstacles, and a comprehensive risk value is generated through field strength fusion;

[0009] S4, when either the cloud or the vehicle end determines that there is a collision risk, the collision avoidance processing mechanism is immediately started to perform deceleration, parking or path adjustment operations.

[0010] Further, the critical safety distance in step S2 is calculated by formula (1):

[0011] D critical = |(v t ·t)-(v f ·t)|+L f +L t (1)

[0012] wherein, D critical is the critical safety distance, v t and L t are the speed and length of the target unmanned agricultural machine, v f and L f are the speed and length of the surrounding vehicle, and t is the time interval dynamically adjusted based on the terrain of the farmland area.

[0013] Further, the time interval t is adaptively adjusted according to the type of farmland terrain, including:

[0014] when the terrain is flat, t is 0.5 seconds;

[0015] when the terrain is hilly, t is 12 seconds;

[0016] when the terrain is slippery, t is 23 seconds;

[0017] when the terrain is densely obstructed, t is 3 seconds or more.

[0018] Further, step S2 specifically includes:

[0019] S21, identifying the farmland area where the target unmanned agricultural machine is located according to the work area number of the target unmanned agricultural machine, and detecting other unmanned agricultural machines in the farmland area where the target unmanned agricultural machine is located, and screening the nearest neighbor vehicle with the same driving direction as the target unmanned agricultural machine;

[0020] S22, combining the dynamic parameters of the target unmanned agricultural machine and the nearest neighbor vehicle, calculating the critical safety distance and performing collision risk assessment, wherein the dynamic parameters of the nearest neighbor vehicle include the position, speed and length of the nearest neighbor vehicle, and the dynamic parameters of the target unmanned agricultural machine include the direction and working state of the target unmanned agricultural machine.

[0021] Further, the method for identifying the driving direction of other unmanned agricultural machines in step S22 specifically includes:

[0022] S221, obtaining the direction of the target unmanned agricultural machine through the inertial navigation measurement data of the target unmanned agricultural machine information collected in S1;

[0023] S222, calculating the vector angle θ between the driving direction of the target unmanned agricultural machine and other unmanned agricultural machines by formula (2):

[0024]

[0025] In the formula, is a driving direction vector of the target unmanned agricultural machine acquired by the inertial navigation measurement data of the target unmanned agricultural machine collected by S1, is a relative driving direction vector of the other unmanned agricultural machine;

[0026] S223, according to θ, the relative driving direction of the target unmanned agricultural machine and the other unmanned agricultural machine is as follows:

[0027] The first case: θ≈0, indicating that the driving direction of the target unmanned agricultural machine and the other unmanned agricultural machine is basically consistent;

[0028] The second case: θ>0, indicating that the driving direction of the target unmanned agricultural machine and the other unmanned agricultural machine is different;

[0029] The third case: θ≈π, indicating that the driving direction of the target unmanned agricultural machine and the other unmanned agricultural machine is exactly opposite.

[0030] Further, the relative distance d of the target unmanned agricultural machine and the other unmanned agricultural machine in step S22 is obtained by calculation according to formula (3):

[0031]

[0032] In the formula, (x P , y P ) is the position of the target unmanned agricultural machine, and (x Q , y Q ) is the position of the other unmanned agricultural machine.

[0033] The application also provides an anti-collision system for unmanned agricultural machine operation, which comprises:

[0034] A data acquisition module is configured to acquire self information and surrounding perception object information of the target unmanned agricultural machine by a laser radar, a positioning module and an inertial navigation unit;

[0035] A cloud confirmation module is configured to calculate a critical safety distance of the target unmanned agricultural machine and surrounding vehicles, and compare the critical safety distance with a preset safety threshold to determine whether there is a collision risk;

[0036] A vehicle-end confirmation module is configured to construct a safety field model to analyze the collision risk between the target unmanned agricultural machine and surrounding perception objects in real time; wherein the safety field model comprises superposition of potential energy field and kinetic energy field, the potential energy field is used to represent the risk of static obstacles, the kinetic energy field is used to represent the risk of dynamic obstacles, and a comprehensive risk value is generated by field strength fusion;

[0037] A collision avoidance processing module is configured to start a collision avoidance processing mechanism to perform deceleration, parking or path adjustment operation when a collision risk is determined at either of the cloud terminal or the vehicle terminal.

[0038] Further, the critical safety distance of the cloud confirmation module is calculated by formula (1):

[0039] D critical = |(v t ·t)-(v f ·t)|+L f +L t (1)

[0040] In the formula, D critical is the critical safety distance, v t and L t are the speed and length of the target unmanned agricultural machine, v f and L f are the speed and length of the surrounding vehicle, and t is the time interval dynamically adjusted based on the terrain of the farmland area.

[0041] Further, the value of the time interval t is adaptively adjusted according to the type of farmland terrain, including:

[0042] When the terrain is flat, t is 0.5 seconds;

[0043] When the terrain is hilly, t is 12 seconds;

[0044] When the terrain is wet and slippery, t is 23 seconds;

[0045] When the terrain is obstacle-dense, t is 3 seconds or more.

[0046] Further, the cloud confirmation module specifically includes:

[0047] A direction filtering unit is configured to identify the farmland area where the target unmanned agricultural machine is located according to the work area number of the target unmanned agricultural machine, and detect other unmanned agricultural machines in the farmland area where the target unmanned agricultural machine is located, and screen the nearest neighbor vehicle with the same driving direction as the target unmanned agricultural machine.

[0048] A dynamic parameter adaptation unit is configured to calculate the critical safety distance and evaluate the collision risk by combining the dynamic parameters of the target unmanned agricultural machine and the nearest neighbor vehicle, wherein the dynamic parameters of the nearest neighbor vehicle include the position, speed and length of the nearest neighbor vehicle, and the dynamic parameters of the target unmanned agricultural machine include the direction and working state of the target unmanned agricultural machine.

[0049] The present application has the following advantages due to the above technical solutions:

[0050] 1. This invention uses cloud-based multi-vehicle data to calculate critical safety distances, while vehicles use safety field models to perceive environmental risks in real time. The two work together to verify and significantly reduce the probability of misjudgments and missed detections. Any terminal (cloud or vehicle) can detect a risk and trigger an emergency action (slowdown, parking, or route adjustment), improving response speed and ensuring rapid collision avoidance.

[0051] 2. The critical safety distance formula adopted by the present invention integrates vehicle speed, vehicle length, and terrain dynamic time intervals to adapt to complex terrain such as mountainous and slippery terrain. Therefore, the accuracy of safety distance calculation is improved. At the same time, the parameters are dynamically adjusted according to the operating status of the agricultural machinery (speed differences such as sowing and harvesting). Therefore, the present invention can reduce the risk of collision caused by differences in operating modes.

[0052] 3. Since the present invention uses a potential energy field to represent static obstacles (ridges, ditches) and a kinetic energy field to represent dynamic obstacles (other agricultural machinery), and generates a comprehensive risk value through field strength fusion, the accuracy of risk identification in complex scenarios is improved. It can also reduce dependence on a single sensor (such as lidar) and reduce data distortion caused by weather or occlusion.

[0053] 4. Because the present invention filters out reverse or irrelevant targets by using the angle between the driving direction vectors, it can reduce redundant calculations and significantly improve system operating efficiency. Furthermore, the present invention proactively identifies vehicles traveling along the same path, avoiding conflicts between agricultural machinery and their ranks, thereby improving the efficiency of multi-machine collaborative operations.

[0054] 5. Since the present invention automatically extends the safety response period according to the terrain type (such as t≥3 seconds in an obstacle-dense area), the present invention has enhanced adaptability to complex environments.

[0055] 6. The present invention adopts a hierarchical response strategy (deceleration → path adjustment → parking), thereby ensuring safety while maintaining operational continuity to the maximum extent and reducing downtime losses.

[0056] This invention solves the problems of poor adaptability, high misjudgment rate, and delayed response of traditional collision avoidance methods in complex farmland environments through cloud-vehicle collaboration, dynamic safety models, full-scene risk perception and other technologies, providing a new solution for the efficient and safe operation of unmanned agricultural machinery. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a flow chart of a collision avoidance method for unmanned agricultural machinery operations proposed in an embodiment of the present invention. DETAILED DESCRIPTION

[0058] In the accompanying drawings, the same or similar reference numerals are used to represent the same or similar elements or elements with the same or similar functions. The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0059] In the description of the present application, the terms "center", "longitudinal", "transverse", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, which is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the scope of protection of the present application.

[0060] Figure 1 The flow chart of the anti-collision method for unmanned agricultural machine operation proposed in the present embodiment is shown in FIG. 1, which comprises the following steps: Figure 1

[0061] Step S1, real-time collection of the target unmanned agricultural machine's own information and surrounding perceived object information through the preset sensing device on the target unmanned agricultural machine, and sending the information to the cloud.

[0062] The target unmanned agricultural machine is the unmanned agricultural machine currently performing anti-collision control. The vehicle can be any type of unmanned agricultural machine for different purposes.

[0063] The sensing device includes a laser radar module, a global navigation satellite system (GNSS) positioning module, and an inertial measurement unit (IMU, hereinafter referred to as an inertial navigation module) for data collection, obtaining the data collected by each module. These data include the target unmanned agricultural machine's own information and the position data of the vehicle's surrounding perceived objects, such as the vehicle's position, speed, heading angle, length, height, and width information, as well as the work area number where the vehicle is located. It should be noted that the work area number is based on the pre-collected farmland map data and the number is made after planning. When the unmanned agricultural machine works in different farmland areas, it will correspond to different numbers set in the map.

[0064] ​In the implementation process, the laser radar module is taken as an example for data collection. First, the laser radar module scans the surrounding environment of the target unmanned agricultural machine to generate point cloud data of the surrounding environment. Then, the point cloud data is analyzed to identify obstacles around the target unmanned agricultural machine, as well as the vehicle's contour and position information. The global navigation satellite system positioning module is installed on the target unmanned agricultural machine to obtain real-time vehicle position data. The inertial measurement unit contains an accelerometer and a gyroscope to measure the acceleration and angular velocity of the unmanned agricultural machine, helping to achieve attitude control and navigation, thereby providing high-precision position and speed information to help precise agricultural operations. This method not only accurately perceives the changes in the environment around the unmanned agricultural machine, but also provides reliable data support for collision avoidance control, ensuring the safety and efficiency of the unmanned agricultural machine in various operating environments.

[0065] It should be noted that using laser radar, positioning module or inertial navigation module alone for data collection may result in reduced data accuracy. This is because each sensor has its inherent errors and uncertainties. To solve this problem, multi-sensor fusion technology can significantly improve positioning accuracy. For example, by using an adaptive Kalman filter, the data of laser radar, global navigation satellite system positioning module and inertial measurement unit are fused, and the filter parameters are dynamically adjusted to adapt to different terrains and operating conditions, improving positioning accuracy and system robustness.

[0066] S2, based on the data collected in step S1, the cloud calculates the critical safety distance between the target unmanned agricultural machine and the surrounding vehicles, and compares the critical safety distance with the preset safety threshold to determine whether there is a collision risk.

[0067] The specific value of the preset safety threshold is obtained from the analysis of historical data.

[0068] The critical safety distance is determined according to the collected data to determine the maximum safety distance that the target unmanned agricultural machine can maintain with other unmanned agricultural machines in the current speed and position scenario.

[0069] In one embodiment, in the farmland environment, the critical safety distance calculation needs to consider the unpredictable terrain changes (such as ridges, trenches) for dynamic adjustment. The critical safety distance of this embodiment can be calculated using formula (1):

[0070] D critical = |(v t ·t) - (v f ·t)| + L f + L t (1)

[0071] In the formula, D critical is the critical safety distance, v t , Lt respectively the speed and the length of the target unmanned agricultural machine, v f , L f respectively the speed and the length of the surrounding vehicle, and t is a time interval dynamically adjusted based on the terrain of the farmland region.

[0072] In one embodiment, the complexity of farmland operation needs to adjust the time interval t according to different terrains, for example, the terrain corresponding to the farmland region is flat terrain, the characteristic is no obstacle, at this time the operation efficiency is high, then the time interval t can be set to 0.5 seconds; the terrain corresponding to the farmland region is hilly terrain, the characteristic is obvious fluctuation, then the operation needs to be slowed down, then the time interval t can be set between 12 seconds; the terrain corresponding to the farmland region is slippery terrain, the characteristic is that the ground is wet, which will increase the risk, then the time interval t can be set between 23 seconds; the terrain corresponding to the farmland region is the terrain with more obstacles, the characteristic is that there are many obstacles, which will affect the operation efficiency, then the time interval t can be set to 3 seconds or more.

[0073] Similarly, in the case of knowing the acceleration a of the target unmanned agricultural machine, the critical safety distance D critical can also be obtained by the following formula:

[0074]

[0075] The skilled person can also obtain the critical safety distance D critical in other ways.

[0076] In order to more clearly illustrate the specific implementation process of the cloud end collision avoidance calculation of the present application, the following will be described in combination with a specific cloud end collision avoidance calculation method proposed in one embodiment of the present application.

[0077] In one embodiment, step S2 specifically comprises:

[0078] S21, using the GNSS positioning module to collect the position data of the target unmanned agricultural machine, and identifying the farmland region where the target unmanned agricultural machine is located through the operation area number instead of the traditional road number. Due to the dynamic and unstructured characteristics of the farmland environment which are different from the urban roads, the present embodiment focuses on finding other unmanned agricultural machines for farmland operation in the region, rather than fixed road vehicles. Detect other unmanned agricultural machines in the farmland region where the target unmanned agricultural machine is located, and screen the nearest neighbor vehicles in the same driving direction as the target unmanned agricultural machine.

[0079] S22, combining the dynamic parameters of the target unmanned agricultural machine and the nearest neighbor vehicle, calculating the critical safety distance and performing collision risk assessment, wherein the dynamic parameters of the nearest neighbor vehicle include the position, speed and length of the nearest neighbor vehicle.

[0080] In step S22, the dynamic parameters of the target unmanned agricultural machine include the direction of the target unmanned agricultural machine, which at least affects the calculation of the critical safety distance in the following two aspects:

[0081] 1. Screening collision threats: the driving direction determines which unmanned agricultural machines are actually at risk of collision. Only vehicles with a driving direction that is substantially the same (i.e., in the same direction or with a very small included angle) need to be further evaluated for distance.

[0082] 2. Filtering irrelevant targets: if the directions are opposite or significantly different (e.g., crossing or opposite operations), the system will determine that they are not the main risk objects at the moment, effectively reducing calculation redundancy and false positives.

[0083] 3. Preventing operation conflicts: in the agricultural environment, agricultural machines may move in operation lines. By detecting the direction, it can be determined whether they will enter the same path as soon as possible, and intervention can be made in advance.

[0084] In step S22, the dynamic parameters of the target unmanned agricultural machine also include the operation state of the target unmanned agricultural machine, including seeding, fertilizing, harvesting, etc. The driving speed, operating width, and inertia of agricultural machines in different operation states are significantly different. For example, the speed is slower during seeding, the speed may be slightly faster during harvesting, and the width of the vehicle body is different during fertilizing, which will all be factors affecting the critical safety distance D critical .

[0085] In this embodiment, the critical safety distance is not only calculated based on parameters such as vehicle speed and vehicle length, but also considers the dynamic operation time interval of different farmland terrains. Therefore, the method of the present application is suitable for mountainous, plain, and slippery environments.

[0086] In step S22, the method for collision risk assessment specifically includes:

[0087] The collision risk is determined by comparing the preset safety distance D safe and the critical safety distance D critical . The difference determination formula is shown in the following formula (4):

[0088] ΔD = D safe - D critical (4)

[0089] If ΔD > 0, there is no collision risk; if ΔD < 0, there is a collision risk.

[0090] According to the identification result of the collision risk, the system timely issues operation instructions to the target unmanned agricultural machine through the cloud, such as reducing speed or stopping immediately. This mechanism is particularly designed for complex and rapidly changing projects in the agricultural environment, ensuring the safe operation and efficient operation of unmanned agricultural machines.

[0091] In one embodiment, the method of identifying the driving direction of other unmanned agricultural machines in step S22 specifically includes:

[0092] S221, the direction of the target unmanned agricultural machine is obtained through the inertial navigation measurement data of the target unmanned agricultural machine information collected by S1. In field operation, the operation direction is not only direct driving, but also may include driving around obstacles or along crop rows, therefore, the direction judgment needs to be dynamically adjusted in combination with the actual operation path of the field. The target unmanned agricultural machine provides accurate driving direction data through the inertial measurement unit. This is particularly important in field operation, because the terrain of the field is complex and variable, which may include ridges, ditches and different crop rows.

[0093] S222, in order to ensure that the unmanned agricultural machine drives along the predetermined operation path in the field, the included angle between the driving direction of the unmanned agricultural machine and other unmanned agricultural machines needs to be calculated to identify and filter out vehicles opposite to the operation direction of the target unmanned agricultural machine. The vector included angle θ between the driving direction of the target unmanned agricultural machine and other unmanned agricultural machines is calculated by using formula (2):

[0094]

[0095] In the formula, is the driving direction vector of the target unmanned agricultural machine obtained by the inertial navigation measurement data of the target unmanned agricultural machine information collected by S1, is the relative driving direction vector of other unmanned agricultural machines;

[0096] S223, according to θ, the relative driving direction of the target unmanned agricultural machine and other unmanned agricultural machines is as follows:

[0097] The first case: θ≈0, indicating that the driving direction of the target unmanned agricultural machine and other unmanned agricultural machines is basically consistent. In this case, the vehicle is suitable for operation in the same operation row, ensuring the continuity and efficiency of the operation.

[0098] The second case: θ>0, indicating that the driving direction of the target unmanned agricultural machine and other unmanned agricultural machines is different. At this time, it is necessary to further evaluate whether the vehicle is in the same operation area to decide whether to adjust the operation path or speed to avoid potential operation interference.

[0099] The third case: θ≈π, indicating that the driving direction of the target unmanned agricultural machine and other unmanned agricultural machines is exactly opposite. In field operation, this case usually needs to be filtered out to avoid operation conflict and potential collision risk, so as to ensure the safe operation of the unmanned agricultural machine. In this way, the system can intelligently identify and filter out vehicles inconsistent with the operation direction of the target unmanned agricultural machine, ensuring that the unmanned agricultural machine can complete the operation task efficiently and safely in the field.

[0100] In one embodiment, the relative distance d between the target unmanned agricultural machine and other unmanned agricultural machines in step S22 is obtained by calculating the Euclidean distance described in formula (3) to identify the closest vehicle to the target unmanned agricultural machine among the vehicles with the same operation direction.

[0101]

[0102] wherein (x P , y P ) is the position of the target unmanned agricultural machine, and (x Q , y Q ) is the position of the other unmanned agricultural machine.

[0103] S3, based on the data collected in step S1, the vehicle end of the target unmanned agricultural machine analyzes the collision risk between the target unmanned agricultural machine and the surrounding perceived objects in real time by constructing a safety field model; wherein the safety field model includes the superposition of potential energy field and kinetic energy field, the potential energy field is used to represent the risk of static obstacles, and the kinetic energy field is used to represent the risk of dynamic obstacles, and a comprehensive risk value is generated by field strength fusion.

[0104] It should be noted that the above method of judging whether there is a collision risk through the cloud is only for the vehicles operating unmanned agricultural machines, and does not take into account other obstacles in the surrounding environment. Based on this, a safety field analysis method based on vehicle-end environmental perception is proposed to predict whether there is a collision risk.

[0105] S4, when either the cloud terminal or the vehicle terminal determines that there is a collision risk, immediately start the anti-collision processing mechanism and perform the speed reduction, parking or path adjustment operation. "Speed reduction" means that in the case where stopping is not applicable, the target unmanned agricultural machine and the surrounding vehicles are controlled to reduce speed to reduce the risk of collision. "Parking" means that the surrounding vehicles and the target unmanned agricultural machine that exist collision risk are controlled to stop running to avoid collision. "Path adjustment" means path adjustment: dynamically adjust the driving path of the target unmanned agricultural machine according to real-time environmental data to avoid potential obstacles.

[0106] Step S4 triggers the anti-collision processing immediately when either the cloud terminal or the vehicle terminal confirms that there is a collision risk, so that the real-time nature of the anti-collision measures can be ensured through fast data transmission and processing, and an efficient emergency response can be formed.

[0107] In the above embodiment, the anti-collision method for unmanned agricultural machine operation of the present application also has the function of returning to the monitoring cycle. After performing the anti-collision processing, the system returns to step S1 and continues to monitor and calculate the anti-collision, ensuring continuous protection. The anti-collision method for unmanned agricultural machine operation of the present application also has the function of double confirmation mechanism. When both the cloud terminal and the vehicle terminal confirm that there is no collision risk, the system returns to the monitoring state and maintains continuous perception of the environment.

[0108] The embodiment of the present application also provides an anti-collision system for unmanned agricultural machine operation, which comprises a data acquisition module, a cloud confirmation module, a vehicle end confirmation module and an anti-collision processing module.

[0109] The data acquisition module is used for acquiring self information and surrounding perceived object information of the target unmanned agricultural machine through a laser radar, a positioning module and an inertial navigation unit.

[0110] The cloud confirmation module is used for calculating a critical safety distance of the target unmanned agricultural machine and surrounding vehicles, and comparing the critical safety distance with a preset safety threshold to determine whether there is a collision risk.

[0111] The vehicle end confirmation module is used for constructing a safety field model and analyzing collision risks between the target unmanned agricultural machine and surrounding perceived objects in real time.

[0112] The anti-collision processing module is used for starting an anti-collision processing mechanism immediately to perform deceleration, parking or path adjustment operation when it is determined at either the cloud end or the vehicle end that there is a collision risk.

[0113] In one embodiment, the critical safety distance of the cloud confirmation module is calculated by using formula (1).

[0114] In one embodiment, the cloud confirmation module specifically comprises a direction filtering unit and a dynamic parameter adaptation unit.

[0115] The direction filtering unit is used for identifying a farmland area where the target unmanned agricultural machine is located according to a work area number of the target unmanned agricultural machine, detecting other unmanned agricultural machines in the farmland area where the target unmanned agricultural machine is located, and screening a nearest neighbor vehicle with the same driving direction as the target unmanned agricultural machine.

[0116] The dynamic parameter adaptation unit is used for combining dynamic parameters of the target unmanned agricultural machine and the nearest neighbor vehicle to calculate the critical safety distance and perform collision risk assessment, wherein the dynamic parameters of the nearest neighbor vehicle include position, speed and vehicle length of the nearest neighbor vehicle, and the dynamic parameters of the target unmanned agricultural machine include direction and work state of the target unmanned agricultural machine.

[0117] In summary, the anti-collision system for unmanned agricultural machine operation of the embodiment of the present application realizes high-precision acquisition of vehicle position data by using vehicle-mounted sensor perception technology, positioning technology and inertial navigation module. Through cloud anti-collision analysis and vehicle end safety field technology analysis and prediction methods, the system can accurately calculate collision risks between unmanned agricultural machines and other perceived objects, and take anti-collision measures such as parking or deceleration in time. This system design can effectively prevent collisions between unmanned agricultural machines, improve the safety of driving vehicles and reduce maintenance and downtime costs.

[0118] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the same. Those skilled in the art should understand that the technical solutions described in the foregoing embodiments can be modified, or some technical features thereof can be replaced by equivalent ones; these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A collision avoidance method for unmanned agricultural machinery operation, characterized in that: include: S1, using the sensor equipment preset on the target unmanned agricultural machine to collect the target unmanned agricultural machine's own information and surrounding perception information in real time, and send the information to the cloud; S2: Based on the data collected in step S1, the cloud calculates the critical safety distance between the target unmanned agricultural machine and surrounding vehicles, and compares the critical safety distance with the preset safety threshold to determine whether there is a collision risk; In step S3, based on the data collected in step S1, the vehicle side of the target unmanned agricultural machine constructs a safety field model to analyze the collision risk between the target unmanned agricultural machine and surrounding sensing objects in real time. The safety field model includes the superposition of potential energy fields and kinetic energy fields. The potential energy field is used to represent the risk of static obstacles, and the kinetic energy field is used to represent the risk of dynamic obstacles. The safety field model generates a comprehensive risk value through field strength fusion. S4: When any terminal on the cloud or vehicle side determines that there is a collision risk, the anti-collision processing mechanism is immediately activated to perform deceleration, parking or path adjustment operations.

2. The anti-collision method for unmanned agricultural machinery operation according to claim 1, characterized in that: The critical safety distance in step S2 is calculated using formula (1): D critical =∣(v t ·t)-(v f ·t)∣+L f +L t (1) Where D critical is the critical safety distance, v t , L t are the speed and length of the target unmanned agricultural machine, v f , L f are the speed and length of surrounding vehicles respectively, and t is the time interval dynamically adjusted based on the terrain of the farmland area.

3. The anti-collision method for unmanned agricultural machinery operation according to claim 2, characterized in that: The value of the time interval t is adaptively adjusted according to the farmland terrain type, including: For flat terrain, t is 0.5 seconds; In hilly terrain, t is 12 seconds; On slippery terrain, t is 23 seconds; In terrain with dense obstacles, t should be 3 seconds or longer.

4. The anti-collision method for unmanned agricultural machinery operation according to any one of claim 13, wherein: Step S2 specifically includes: S21, identifying the farmland area where the target unmanned agricultural machine is located based on its operation area number, detecting other unmanned agricultural machines in the farmland area where the target unmanned agricultural machine is located, and selecting the nearest neighboring vehicles that are traveling in the same direction as the target unmanned agricultural machine; S22, combining the dynamic parameters of the target unmanned agricultural machinery and the nearest neighbor vehicle, calculating the critical safety distance and performing a collision risk assessment, wherein the dynamic parameters of the nearest neighbor vehicle include the position, speed, and length of the nearest neighbor vehicle, and the dynamic parameters of the target unmanned agricultural machinery include the direction and operating status of the target unmanned agricultural machinery.

5. The anti-collision method for unmanned agricultural machinery operation according to claim 4, characterized in that: The method for identifying the driving direction of other unmanned agricultural machines in step S22 specifically includes: S221, obtaining the direction of the target unmanned agricultural machinery through the inertial navigation measurement data of the target unmanned agricultural machinery information collected in S1; S222: Using equation (2), calculate the vector angle θ between the target unmanned agricultural machine and the driving directions of other unmanned agricultural machines: Where, The driving direction vector of the target unmanned agricultural machinery is obtained through the inertial navigation measurement data of the target unmanned agricultural machinery information collected by S1. is the relative driving direction vector of other unmanned agricultural machinery; S223: Based on θ, the relative driving directions of the target unmanned agricultural machine and other unmanned agricultural machines are determined according to the following three scenarios: The first scenario: θ≈0, indicating that the target UAV and other UAVs are moving in the same direction. The second scenario: θ>0, indicating that the target UAV and other UAVs have different driving directions; The third scenario: θ≈π indicates that the target UAV is traveling in the opposite direction to the other UAVs.

6. The anti-collision method for unmanned agricultural machinery operation according to claim 1, characterized in that: The relative distance d between the target unmanned agricultural machine and other unmanned agricultural machines in step S22 is calculated using formula (3): In the formula, (x P ,y P ) is the location of the target unmanned agricultural machine, (x Q ,y Q ) is the location of other unmanned agricultural machinery.

7. A collision avoidance system for unmanned agricultural machinery operation, characterized in that: include: The data acquisition module is used to collect the target unmanned agricultural machinery's own information and surrounding perception object information through the laser radar, positioning module and inertial navigation unit; The cloud-based confirmation module is used to calculate the critical safety distance between the target unmanned agricultural machinery and surrounding vehicles, and compare the critical safety distance with the preset safety threshold to determine whether there is a collision risk; The vehicle-side confirmation module is used to build a safety field model and analyze the collision risk between the target unmanned agricultural machinery and surrounding sensing objects in real time. The safety field model includes the superposition of potential energy fields and kinetic energy fields. The potential energy field is used to represent the risk of static obstacles, and the kinetic energy field is used to represent the risk of dynamic obstacles. The comprehensive risk value is generated through field strength fusion. The anti-collision processing module is used to immediately activate the anti-collision processing mechanism and perform deceleration, parking or path adjustment operations when any terminal on the cloud or vehicle side determines that there is a collision risk.

8. The anti-collision system for unmanned agricultural machinery operation according to claim 7, characterized in that: The critical safety distance of the cloud confirmation module is calculated using formula (1): D critical =∣(v t ·t)-(v f ·t)∣+L f +L t (1) Where D critical is the critical safety distance, v t , L t are the speed and length of the target unmanned agricultural machine, v f , L f are the speed and length of surrounding vehicles respectively, and t is the time interval dynamically adjusted based on the terrain of the farmland area.

9. The anti-collision system for unmanned agricultural machinery operation according to claim 8, characterized in that: The value of the time interval t is adaptively adjusted according to the farmland terrain type, including: For flat terrain, t is 0.5 seconds; In hilly terrain, t is 12 seconds; On slippery terrain, t is 23 seconds; In terrain with dense obstacles, t should be 3 seconds or longer.

10. The anti-collision system for unmanned agricultural machinery operation according to any one of claim 79, characterized in that: The cloud confirmation module specifically includes: A direction filtering unit is used to identify the farmland area where the target UAV is located based on its operation area number, detect other UAVs in the farmland area where the target UAV is located, and select the nearest neighbor vehicles with the same driving direction as the target UAV; The dynamic parameter adaptation unit is used to combine the dynamic parameters of the target unmanned agricultural machinery and the nearest neighbor vehicle to calculate the critical safety distance and perform collision risk assessment. The dynamic parameters of the nearest neighbor vehicle include the position, speed and length of the nearest neighbor vehicle, and the dynamic parameters of the target unmanned agricultural machinery include the direction and operating status of the target unmanned agricultural machinery.

Images