An obstacle recognition method for driverless mining trucks

Abstract

This invention provides an obstacle recognition method for unmanned mining trucks, comprising: capturing historical collision information of the unmanned mining truck from a big data platform, and determining the probability and severity of accidents at different collision locations within the same collision type; acquiring the mining truck frame, performing a safety analysis of multiple preset positions on the frame, and determining a reasonable installation position based on the accident probability and severity level; performing initial perception of the surrounding environment of the unmanned mining truck based on the obstacle recognition component at the reasonable installation position, adding new safe positions and new recognition components to supplement the initial perception results; analyzing the supplemented perception results to determine whether they meet the obstacle recognition criteria; if so, controlling the unmanned mining truck to avoid obstacles. This ensures the accuracy of obstacle recognition and the completeness of the safety recognition of the mining truck's surroundings, thereby guaranteeing the safety of the unmanned mining truck.

Description

Technical Field

[0001] This invention relates to the field of unmanned driving technology, and in particular to an obstacle recognition method for unmanned mining trucks. Background Technology

[0002] The intelligent single-vehicle design of unmanned mining trucks should enhance the perception capabilities of existing autonomous mining trucks. It should be able to identify special obstacles in the mining area, such as mining trucks and bulldozers, reducing the number of missed obstacles. The installation location of the perception system hardware should consider protective measures and reasonable installation, ensuring it is not damaged by collisions during loading, operation, or unloading, as this would undoubtedly pose a certain danger to the unmanned mining truck.

[0003] Therefore, this invention proposes an obstacle recognition method for unmanned mining trucks. Summary of the Invention

[0004] This invention provides an obstacle recognition method for unmanned mining trucks. By capturing and analyzing collision information and conducting safety analysis of the mining truck's structure, it effectively determines a reasonable installation position to monitor the entire area around the mining truck, ensuring the accuracy of obstacle recognition. Furthermore, by adding safe locations and recognition equipment, it further ensures the completeness of safety recognition of the entire area around the mining truck, thereby guaranteeing the safety of the unmanned mining truck.

[0005] This invention provides an obstacle recognition method for an unmanned mining truck, comprising:

[0006] Step 1: Capture historical collision information of unmanned mining trucks from the big data platform, and determine the probability of accidents and the severity of accidents at different collision locations within the same collision type from the historical collision information;

[0007] Step 2: Obtain the mine car frame of the unmanned mine car, perform a safety analysis of the pre-set positions of the mine car frame at multiple points, and determine a reasonable installation position based on the probability of accident occurrence and the severity level of the accident.

[0008] Step 3: Based on the obstacle recognition component installed at the reasonable location, perform a first perception of the surrounding environment of the unmanned mining truck, and supplement the first perception result by adding a new safe location and a new recognition component.

[0009] Step 4: Analyze the supplemented perception results to determine whether the obstacle recognition criteria are met. If they are met, control the unmanned mining truck to avoid the obstacle.

[0010] Preferably, determining the probability of an accident and the severity of an accident at different collision locations within the same collision type from the historical collision information includes:

[0011] Extract all first collision information of the same type of collision object from the historical collision information, and obtain the degree of collision damage corresponding to each first collision information;

[0012] A first count is performed on the number of occurrences of all first collision information for the same type of collision object; a second count is performed on the number of occurrences of the same type of collision position; and a third count is performed on the number of occurrences of all first collision information in the historical collision information.

[0013] Based on the second and third statistical results, the probability of accidents occurring at different collision locations within the same collision type is determined.

[0014] Based on all accident types involved in the same collision type, the first and third statistical results, and all collision damage levels involved, the severity of accidents at different collision locations within the same collision type is calculated.

[0015] Preferably, the process involves acquiring the mine car frame of the unmanned mining truck and performing a safety analysis of the mine car frame at multiple preset positions, including:

[0016] Based on the frame hardness database, the hardness of each frame location point on the mine car frame is determined;

[0017] Based on the architecture type database, the multiple preset positions of the mining car architecture are determined;

[0018] Associate each preset location with the hardness of the corresponding construction location point.

[0019] Preferably, a reasonable installation location is determined by combining the probability of the accident and the severity of the accident, including:

[0020] Based on the safety analysis results, the hardness of each collision location is determined, and an analysis array for the corresponding collision locations is constructed by combining the probability of an accident and the severity of the accident at each collision location.

[0021] The vibration response of the unmanned mining vehicle to the rest of the vehicle is obtained by monitoring the collision process at each collision location, and an influence array corresponding to the collision location is constructed.

[0022] Based on the analysis array and the influence array, a first reasonable set H1 of corresponding collision locations is determined;

[0023] ;

[0024] Where p1 represents the probability of an accident occurring in the analysis array at the corresponding collision location; p0 represents the probability of an accident occurring during factory testing; d1 represents the severity level of an accident in the analysis array at the corresponding collision location; d0 represents the accident level during factory testing; g1 represents the hardness of the location in the analysis array at the corresponding collision location; and g0 represents the impact resistance level during factory testing, which is less than or equal to g1. This represents the collision changes caused by the vibration response of the i1th and other parts in the influence array corresponding to the collision position. This indicates the factory condition of the corresponding collision location; n1 represents the total number of the remaining parts. This means that when p1 is greater than p0, it has reference value; otherwise, it has no reference value. This means that when d1 is greater than d0, it has reference value; otherwise, it has no reference value.

[0025] Based on the first reasonable set H1, calculate the first reasonable value H0 of the corresponding collision position;

[0026] ;

[0027] in, This represents the conversion coefficient based on probabilistic reference value; This represents the conversion factor based on the reference value of accident level; This represents a conversion factor based on the reference value of the impact severity. A conversion factor that represents the reference value of the vibration response;

[0028] When the first reasonable value is greater than the preset factory test value, the corresponding collision position is determined to be a reasonable installation position. At the same time, the maximum affected parts in each accident corresponding to each reasonable installation position are counted, and the two affected parts with the highest number of counts are determined to be reasonable installation positions.

[0029] Preferably, based on the obstacle recognition component installed at the reasonable location, the unmanned mining vehicle performs a first perception of its surrounding environment, including:

[0030] The flow of vehicles is monitored in the current moving position of the unmanned mining truck and the area where the truck is moving.

[0031] When the flow of vehicles is less than the preset flow of vehicles, the first active component on the unmanned mining truck is controlled to perform a first perception of the surrounding environment.

[0032] Otherwise, control all obstacle recognition components on the unmanned mining truck to perform initial sensing;

[0033] The first active component is a part of all obstacle recognition components.

[0034] Preferably, before controlling the first active component on the unmanned mining vehicle to perform the first perception of the surrounding environment, the method further includes:

[0035] Obtain the first layout of reasonable installation positions on the unmanned mining truck, divide it according to the mining truck frame of the unmanned mining truck, and count the number of reasonable installation positions on each divided frame;

[0036] according to Determine the number of components to be sensed in each partition architecture and sort the reasonable installation locations on each partition architecture from largest to smallest according to the first reasonable value, and lock the corresponding components as the first active components;

[0037] in, This represents the maximum reasonable value that exists in the corresponding partitioning architecture; represents the average reasonable value of the corresponding partitioned architecture; [ ] represents the rounding symbol; n2 represents the statistical result of the number of reasonable installation positions in the corresponding partitioned area; n3 represents the number of objects to be sensed in the corresponding partitioned architecture.

[0038] Preferably, based on the first perception result, a new security location and a new recognition component are added to supplement the first perception result, including:

[0039] Based on the initial perception results, a three-dimensional model of the vehicle body is constructed, and it is determined whether there are any missing locations in all directions of the vehicle body.

[0040] If it exists, then based on the missing angle corresponding to the missing location and the missing position based on the unmanned mining truck, a new safe position and a new identification component matching the new safe position are set from the missing-supplement database to supplement the first perception result.

[0041] Preferably, when analyzing the supplemented perception results to determine whether they meet the obstacle recognition criteria, the process includes:

[0042] Based on the supplemented perception structure, an all-round three-dimensional model is constructed, and several independent regions of the all-round three-dimensional model are captured in parallel to determine the possible objects in each independent region.

[0043] Perform combination analysis on all possible objects to generate several combination conditions, and analyze the correlation probability between each combination condition and the obstacle identification criteria;

[0044] The final probability is determined based on all associated probabilities. When the final probability is greater than the preset probability, the obstacle recognition standard is deemed to be met.

[0045] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0046] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0047] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0048] Figure 1 This is a flowchart of an obstacle recognition method for an unmanned mining truck according to an embodiment of the present invention. Detailed Implementation

[0049] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0050] This invention provides an obstacle recognition method for unmanned mining trucks, such as... Figure 1 As shown, it includes:

[0051] Step 1: Capture historical collision information of unmanned mining trucks from the big data platform, and determine the probability of accidents and the severity of accidents at different collision locations within the same collision type from the historical collision information;

[0052] Step 2: Obtain the mine car frame of the unmanned mine car, perform a safety analysis of the pre-set positions of the mine car frame at multiple points, and determine a reasonable installation position based on the probability of accident occurrence and the severity level of the accident.

[0053] Step 3: Based on the obstacle recognition component installed at the reasonable location, perform a first perception of the surrounding environment of the unmanned mining truck, and supplement the first perception result by adding a new safe location and a new recognition component.

[0054] Step 4: Analyze the supplemented perception results to determine whether the obstacle recognition criteria are met. If they are met, control the unmanned mining truck to avoid the obstacle.

[0055] In this embodiment, historical collision information refers to collision information involving two or more mining trucks.

[0056] In this embodiment, the same collision type is such as rear-end collision type, parallel friction type, etc.

[0057] In this embodiment, the probability of an accident refers to: for example, if 10 collisions occur at location A, and a total of 20 collisions are counted for the collision types related to location A, then the probability of an accident at location A is 50%. The severity of the accident can be obtained by calculating the average value based on the collision levels identified by the traffic police.

[0058] In this embodiment, the mining truck frame refers to the vehicle structure of the driverless mining truck itself.

[0059] In this embodiment, the safety analysis of the multi-point preset position is determined based on the hardness of different positions of the mine car itself, and the hardness is set before the mine car leaves the factory.

[0060] In this embodiment, a reasonable installation location refers to a location with a high probability of accident occurrence, a high degree of accident severity, a low hardness level, and minimal impact after a collision.

[0061] In this embodiment, the obstacle recognition component may be a camera, a laser scanning device, an ultrasonic device, etc.

[0062] In this embodiment, the first perception is to obtain surrounding information through obstacle recognition.

[0063] In this embodiment, based on the first perception result, such as the monitoring range displayed on the vehicle by the camera capturing the surrounding area, but a part of the area is not monitored, it is necessary to add a camera for that part to ensure all-round monitoring, that is, the completeness of the perception result.

[0064] In this embodiment, obstacle identification criteria refer to obstacle type, whether the obstacle has moved, etc.

[0065] If the minecart is moving and moving towards the direction where it is stopped, and the distance is less than the safe distance, then it is necessary to control the minecart to perform an evasive maneuver.

[0066] The beneficial effects of the above technical solution are: by capturing and analyzing collision information and conducting safety analysis on the mine car frame, a reasonable installation position can be effectively determined to monitor the mine car's surroundings, ensuring the accuracy of obstacle identification. Furthermore, by adding safe positions and identification equipment, the integrity of safety identification of the mine car's surroundings can be further guaranteed, thus ensuring the safety of the unmanned mine car.

[0067] This invention provides an obstacle recognition method for unmanned mining trucks, which determines the probability of accidents and the severity of accidents at different collision locations within the same collision type from historical collision information, including:

[0068] Extract all first collision information of the same type of collision object from the historical collision information, and obtain the degree of collision damage corresponding to each first collision information;

[0069] A first count is performed on the number of occurrences of all first collision information for the same type of collision object; a second count is performed on the number of occurrences of the same type of collision position; and a third count is performed on the number of occurrences of all first collision information in the historical collision information.

[0070] Based on the second and third statistical results, the probability of accidents occurring at different collision locations within the same collision type is determined.

[0071] Based on all accident types involved in the same collision type, the first and third statistical results, and all collision damage levels involved, the severity of accidents at different collision locations within the same collision type is calculated.

[0072] In this embodiment, the collision object refers to a mining car of the same model, which provides a basis for determining the degree of collision and the probability of an accident.

[0073] In this embodiment, the degree of collision damage is obtained through autonomous analysis based on the first collision information and a pre-set program.

[0074] In this embodiment, the severity of the accident is calculated by averaging the damage levels of all collisions at that location, and then optimized based on the first and third statistical results.

[0075] For example: Accident severity = ave (total collision damage) × ln(2 + m3 / m1), where m3 is the third statistical result and m1 is the first statistical result.

[0076] The beneficial effects of the above technical solution are: by extracting collision information, statistics on the number of occurrences can be performed, which facilitates subsequent calculations of probability and accident severity, providing a basis for subsequent location setting.

[0077] This invention provides an obstacle recognition method for an unmanned mining truck, which involves acquiring the truck's frame and performing a multi-point preset position safety analysis on the frame, including:

[0078] Based on the frame hardness database, the hardness of each frame location point on the mine car frame is determined;

[0079] Based on the architecture type database, the multiple preset positions of the mining car architecture are determined;

[0080] Associate each preset location with the hardness of the corresponding construction location point.

[0081] In this embodiment, the architecture hardness database is set before the mining truck leaves the factory. It includes the location points and the hardness of each location point, and the preset locations are also preset.

[0082] The beneficial effects of the above technical solution are: determining the hardness and preset position through the database provides a basis for setting up components.

[0083] This invention provides an obstacle recognition method for unmanned mining trucks, which determines a reasonable installation location by combining the probability of an accident and the severity level of the accident, including:

[0084] Based on the safety analysis results, the hardness of each collision location is determined, and an analysis array for the corresponding collision locations is constructed by combining the probability of an accident and the severity of the accident at each collision location.

[0085] The vibration response of the unmanned mining vehicle to the rest of the vehicle is obtained by monitoring the collision process at each collision location, and an influence array corresponding to the collision location is constructed.

[0086] Based on the analysis array and the influence array, a first reasonable set H1 of corresponding collision locations is determined;

[0087] ;

[0088] Where p1 represents the probability of an accident occurring in the analysis array at the corresponding collision location; p0 represents the probability of an accident occurring during factory testing; d1 represents the severity level of an accident in the analysis array at the corresponding collision location; d0 represents the accident level during factory testing; g1 represents the hardness of the location in the analysis array at the corresponding collision location; and g0 represents the impact resistance level during factory testing, which is less than or equal to g1. This represents the collision changes caused by the vibration response of the i1th and other parts in the influence array corresponding to the collision position. This indicates the factory condition of the corresponding collision location; n1 represents the total number of the remaining parts. This means that when p1 is greater than p0, it has reference value; otherwise, it has no reference value. This means that when d1 is greater than d0, it has reference value; otherwise, it has no reference value.

[0089] Based on the first reasonable set H1, calculate the first reasonable value H0 of the corresponding collision position;

[0090] ;

[0091] in, This represents the conversion coefficient based on probabilistic reference value; This represents the conversion factor based on the reference value of accident level; This represents a conversion factor based on the reference value of the impact severity. A conversion factor that represents the reference value of the vibration response;

[0092] When the first reasonable value is greater than the preset factory test value, the corresponding collision position is determined to be a reasonable installation position. At the same time, the maximum affected parts in each accident corresponding to each reasonable installation position are counted, and the two affected parts with the highest number of counts are determined to be reasonable installation positions.

[0093] Analysis array = {Position 1: Probability of accident occurrence, severity level, hardness}

[0094] Influence array = {influence of position 1 on position a1, influence of position 1 on position a2, ...}

[0095] In this embodiment, for example, after colliding with the front left of the mine car, the front bumper may fall off, and the falling windshield may cause paint to be scratched on non-collision areas of the mine car, which would have an impact.

[0096] In this embodiment, the reasonable installation position includes not only the safe position where the first reasonable value is greater than the preset factory test value and the installation positions corresponding to the first two affected parts, but also multiple preset positions.

[0097] The beneficial effects of the above technical solution are: by constructing an analysis array and an influence array at the same location, and by constructing the set and calculating reasonable values, the points that need to be monitored in the actual process can be determined, and by combining multiple preset locations, the identification can be further improved and the integrity of the identification can be guaranteed.

[0098] This invention provides an obstacle recognition method for an unmanned mining truck. Based on an obstacle recognition component installed at a suitable location, the method performs a first perception of the surrounding environment of the unmanned mining truck, including:

[0099] The flow of vehicles is monitored in the current moving position of the unmanned mining truck and the area where the truck is moving.

[0100] When the flow of vehicles is less than the preset flow of vehicles, the first active component on the unmanned mining truck is controlled to perform a first perception of the surrounding environment.

[0101] Otherwise, control all obstacle recognition components on the unmanned mining truck to perform initial sensing;

[0102] The first active component is a part of all obstacle recognition components.

[0103] Preferably, before controlling the first active component on the unmanned mining vehicle to perform the first perception of the surrounding environment, the method further includes:

[0104] Obtain the first layout of reasonable installation positions on the unmanned mining truck, divide it according to the mining truck frame of the unmanned mining truck, and count the number of reasonable installation positions on each divided frame;

[0105] according to Determine the number of components to be sensed in each partition architecture and sort the reasonable installation locations on each partition architecture from largest to smallest according to the first reasonable value, and lock the corresponding components as the first active components;

[0106] in, This represents the maximum reasonable value that exists in the corresponding partitioning architecture; represents the average reasonable value of the corresponding partitioned architecture; [ ] represents the rounding symbol; n2 represents the statistical result of the number of reasonable installation positions in the corresponding partitioned area; n3 represents the number of objects to be sensed in the corresponding partitioned architecture.

[0107] In this embodiment, the preset vehicle flow rate is pre-set.

[0108] In this embodiment, the structural divisions include, for example, the left front door, the right front door, the left rear door, the right rear door, and the water tank frame.

[0109] The beneficial effects of the above technical solution are: by counting the number of reasonable installation positions on each partitioned structure, it is easy to effectively determine the active components, facilitate the use of two modes for sensing, effectively reduce sensing loss, and ensure the rationality of monitoring and identification.

[0110] This invention provides an obstacle recognition method for unmanned mining trucks. Based on a first perception result, a new safe location and a new recognition component are added to supplement the first perception result, including:

[0111] Based on the initial perception results, a three-dimensional model of the vehicle body is constructed, and it is determined whether there are any missing locations in all directions of the vehicle body.

[0112] If it exists, then based on the missing angle corresponding to the missing location and the missing position based on the unmanned mining truck, a new safe position and a new identification component matching the new safe position are set from the missing-supplement database to supplement the first perception result.

[0113] In this embodiment, for example, if position B is missing, a new component needs to be set for position B to supplement the result.

[0114] The beneficial effect of the above technical solution is that by determining the missing locations, the perception results can be supplemented, thus ensuring the completeness of the perception.

[0115] This invention provides an obstacle recognition method for an unmanned mining truck, which analyzes the supplemented perception results to determine whether the obstacle recognition criteria are met, including:

[0116] Based on the supplemented perception structure, an all-round three-dimensional model is constructed, and several independent regions of the all-round three-dimensional model are captured in parallel to determine the possible objects in each independent region.

[0117] Perform combination analysis on all possible objects to generate several combination conditions, and analyze the correlation probability between each combination condition and the obstacle identification criteria;

[0118] The final probability is determined based on all associated probabilities. When the final probability is greater than the preset probability, the obstacle recognition standard is deemed to be met.

[0119] In this embodiment, the independent region is the aforementioned divided region, and the omnidirectional 3D model is the monitored omnidirectional scene.

[0120] In this embodiment, "potentially existing object" refers to a suspected obstacle.

[0121] Because each independent region may only detect a portion of the potential obstacles, a combined analysis is performed. For example, each two adjacent independent regions are combined to piece together the detected potential objects and generate the corresponding combination conditions, which are the outlines of the existing obstacles.

[0122] If the obstacle outline matches the standard outline set under the obstacle recognition standard, the higher the matching degree, the greater the correlation probability. The maximum probability is obtained from all correlation probabilities, and the average probability is calculated by averaging all correlation probabilities. Then, the maximum probability and the average probability are averaged to obtain the final probability.

[0123] In this embodiment, the preset probability is pre-set, and its value is generally 0.6.

[0124] The beneficial effects of the above technical solution are: by performing parallel capture of each independent area, the processing efficiency is improved, and by analyzing the possible objects in each area and performing combination analysis, the correlation probability is determined, which provides a basis for subsequent control of whether the mining truck avoids the obstacles.

[0125] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An obstacle recognition method for an unmanned mining truck, characterized in that, include: Step 1: Capture historical collision information of unmanned mining trucks from the big data platform, and determine the probability of accidents and the severity of accidents at different collision locations within the same collision type from the historical collision information; Step 2: Obtain the mine car frame of the unmanned mine car, perform a safety analysis of the pre-set positions of the mine car frame at multiple points, and determine a reasonable installation position based on the probability of accident occurrence and the severity level of the accident. Step 3: Based on the obstacle recognition component installed at the reasonable location, perform a first perception of the surrounding environment of the unmanned mining truck, and supplement the first perception result by adding a new safe location and a new recognition component. Step 4: Analyze the supplemented perception results to determine whether the obstacle recognition criteria are met. If they are met, control the unmanned mining truck to avoid the obstacle.

2. The obstacle recognition method for unmanned mining trucks according to claim 1, characterized in that, Determine the probability of accidents and the severity of accidents at different collision locations within the same collision type from the historical collision information, including: Extract all first collision information of the same type of collision object from the historical collision information, and obtain the degree of collision damage corresponding to each first collision information; A first statistic is performed on the number of occurrences of all first collision information for the same type of collision object, a second statistic is performed on the number of occurrences of the same type of collision position, and a third statistic is performed on the number of occurrences of all first collision information in the historical collision information; Based on the second and third statistical results, the probability of accidents occurring at different collision locations within the same collision type is determined. Based on all accident types involved in the same collision type, the first and third statistical results, and all collision damage levels involved, the severity of accidents at different collision locations within the same collision type is calculated.

3. The obstacle recognition method for unmanned mining trucks according to claim 1, characterized in that, Obtain the mine car frame of the unmanned mining truck and perform a safety analysis of the mine car frame at multiple preset positions, including: Based on the frame hardness database, the hardness of each frame location point on the mine car frame is determined; Based on the architecture type database, the multiple preset positions of the mining car architecture are determined; Associate each preset location with the hardness of the corresponding construction location point.

4. The obstacle recognition method for unmanned mining trucks according to claim 1, characterized in that, Based on the probability of the accident and the severity level of the accident, a reasonable installation location is determined, including: Based on the safety analysis results, the hardness of each collision location is determined, and an analysis array for the corresponding collision locations is constructed by combining the probability of an accident and the severity of the accident at each collision location. The vibration response of the unmanned mining vehicle to the rest of the vehicle is obtained by monitoring the collision process at each collision location, and an influence array corresponding to the collision location is constructed. Based on the analysis array and the influence array, a first reasonable set H1 of corresponding collision locations is determined; ； Where p1 represents the probability of an accident occurring in the analysis array at the corresponding collision location; p0 represents the probability of an accident occurring during factory testing; d1 represents the severity level of an accident in the analysis array at the corresponding collision location; d0 represents the accident level during factory testing; g1 represents the hardness of the location in the analysis array at the corresponding collision location; and g0 represents the impact resistance level during factory testing, which is less than or equal to g1. This represents the collision changes caused by the vibration response of the i1th and other parts in the influence array corresponding to the collision position. This indicates the factory condition of the corresponding collision location; n1 represents the total number of the remaining parts. This means that when p1 is greater than p0, it has reference value; otherwise, it has no reference value. This means that when d1 is greater than d0, it has reference value; otherwise, it has no reference value. Based on the first reasonable set H1, calculate the first reasonable value H0 of the corresponding collision position; ; in, This represents the conversion coefficient based on probabilistic reference value; This represents the conversion factor based on the reference value of accident level; This represents a conversion factor based on the reference value of the impact severity. A conversion factor that represents the reference value of the vibration response; When the first reasonable value is greater than the preset factory test value, the corresponding collision position is determined to be a reasonable installation position. At the same time, the maximum affected parts in each accident corresponding to each reasonable installation position are counted, and the two affected parts with the highest number of counts are determined to be reasonable installation positions.

5. The obstacle recognition method for unmanned mining trucks according to claim 1, characterized in that, Based on the obstacle recognition component installed at the aforementioned reasonable location, the unmanned mining truck performs a first perception of its surrounding environment, including: The flow of vehicles is monitored in the current moving position of the unmanned mining truck and the area where the truck is moving. When the flow of vehicles is less than the preset flow of vehicles, the first active component on the unmanned mining truck is controlled to perform a first perception of the surrounding environment. Otherwise, control all obstacle recognition components on the unmanned mining truck to perform initial sensing; The first active component is a part of all obstacle recognition components.

6. The obstacle recognition method for unmanned mining trucks according to claim 1, characterized in that, Before controlling the first active component on the driverless mining truck to perform its first perception of the surrounding environment, the method further includes: Obtain the first layout of reasonable installation positions on the unmanned mining truck, divide it according to the mining truck frame of the unmanned mining truck, and count the number of reasonable installation positions on each divided frame; according to Determine the number of components to be sensed in each partition architecture and sort the reasonable installation locations on each partition architecture from largest to smallest according to the first reasonable value, and lock the corresponding components as the first active components; in, This represents the maximum reasonable value that exists in the corresponding partitioning architecture; represents the average reasonable value of the corresponding partitioned architecture; [ ] represents the rounding symbol; n2 represents the statistical result of the number of reasonable installation positions in the corresponding partitioned area; n3 represents the number of objects to be sensed in the corresponding partitioned architecture.

7. The obstacle recognition method for unmanned mining trucks according to claim 1, characterized in that, Based on the initial perception results, new safe locations and new recognition components are added to supplement the initial perception results, including: Based on the initial perception results, a three-dimensional model of the vehicle body is constructed, and it is determined whether there are any missing locations in all directions of the vehicle body. If it exists, then based on the missing angle corresponding to the missing location and the missing position based on the unmanned mining truck, a new safe position and a new identification component matching the new safe position are set from the missing-supplement database to supplement the first perception result.

8. The obstacle recognition method for unmanned mining trucks according to claim 1, characterized in that, When analyzing the supplemented perception results to determine whether they meet the obstacle recognition criteria, the following steps are taken: Based on the supplemented perception structure, an all-round three-dimensional model is constructed, and several independent regions of the all-round three-dimensional model are captured in parallel to determine the possible objects in each independent region. Perform combination analysis on all possible objects to generate several combination conditions, and analyze the correlation probability between each combination condition and the obstacle identification criteria; The final probability is determined based on all associated probabilities. When the final probability is greater than the preset probability, the obstacle recognition standard is deemed to be met.

Images