Dynamic path planning method and device and storage medium

Abstract

The invention discloses a dynamic path planning method and device and a storage medium, and relates to the technical field of intelligent mobile and the technical field of digital medical treatment. Comprising the steps that in response to a path planning instruction of the intelligent mobile device, path planning information is determined, and the path planning information comprises a current position, a previous moment position, a pre-arriving end point position and surrounding environment information; based on the path planning information, determining a driving path net-shaped model in real time, the driving path net-shaped model being composed of a plurality of driving paths; determining a current sub-node of the current position node on each driving path, determining node attribute information based on the current sub-node, determining path attribute information of each driving path based on the current position node and the current sub-node, and determining the current position node according to the node attribute information and the path attribute information of each driving path. Determining the transition probability from the current position node to the current child node on each driving path; and selecting a target driving path based on the transition probability.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent mobile technology and digital medical technology, and in particular, to a dynamic path planning method, apparatus, and storage medium. Background Art

[0002] With the continuous development of technology, mobile robots and autonomous driving technologies are becoming more and more popular. For example, intelligent mobile robots are introduced in hospitals to navigate patients. To ensure the stable operation of intelligent mobile devices such as mobile robots and autonomous driving vehicles, it is necessary to reasonably plan their driving paths.

[0003] Currently, the driving path on the map is usually used as the static driving path of intelligent mobile devices. However, this static driving path cannot avoid unexpected situations on the road, resulting in a low determination accuracy of the driving path and potential driving risks. Summary of the Invention

[0004] The present invention provides a dynamic path planning method, apparatus, and storage medium, mainly capable of improving the planning accuracy of the driving path and enhancing the driving safety of intelligent mobile devices.

[0005] According to a first aspect of the present invention, a dynamic path planning method is provided, including:

[0006] In response to a path planning instruction of an intelligent mobile device, determining path planning information of the intelligent mobile device, where the path planning information includes a current position, a position at the previous moment, a pre-arrival end position, and surrounding environment information;

[0007] Based on the path planning information, determining a driving path network model for the intelligent mobile device in real time, where the driving path network model is composed of multiple driving paths, and each driving path includes the same position node at the previous moment, current position node, and pre-arrival end position node, and at least one position node between the current position node and the pre-arrival end position node that is not completely the same;

[0008] Determining a current sub-node of the current position node on each driving path, and determining node attribute information based on the current sub-node on each driving path. Based on the current position node and the current sub-node on each driving path, determining path attribute information of each driving path. According to the node attribute information and the path attribute information of each driving path, determining a transfer probability from the current position node to the current sub-node on each driving path;

[0009] Based on the transfer probability, selecting a target driving path for the intelligent mobile device in each driving path.

[0010] Optionally, determining node attribute information based on the current sub-nodes on each of the driving paths, determining path attribute information for each of the driving paths based on the current position node and the current sub-nodes on each of the driving paths, and determining the transition probability from the current position node to the current sub-nodes on each of the driving paths according to the node attribute information and the path attribute information of each of the driving paths, includes:

[0011] Counting the total number of nodes of the current sub-nodes on each of the driving paths as the node attribute information, and taking the path lengths between the current position node and the current sub-nodes on each of the driving paths as the path attribute information for each of the driving paths;

[0012] According to the node attribute information and the path attribute information of each of the driving paths, determining the tendency information of each of the driving paths respectively, and determining the total tendency information based on the path attribute information of each of the driving paths and the total number of position nodes in the driving path network model;

[0013] Determining the tendency probability from the current position node to the current sub-nodes on each of the driving paths according to the ratio of the tendency information of each of the driving paths to the total tendency information;

[0014] Based on the tendency probability corresponding to each of the driving paths, determining the transition probability from the current position node to the current sub-nodes on each of the driving paths.

[0015] Optionally, the determining the transition probability from the current position node to the current sub-nodes on each of the driving paths based on the tendency probability corresponding to each of the driving paths, includes:

[0016] Respectively determining the ratio of the path attribute information of each of the driving paths to the total path attribute information commonly corresponding to each of the driving paths as the correction parameter of the tendency probability corresponding to each of the driving paths;

[0017] Based on the correction parameter, correcting the tendency probability corresponding to each of the driving paths to obtain the transition probability from the current position node to the current sub-nodes on each of the driving paths.

[0018] Optionally, the selecting a target driving path for the intelligent mobile device in each of the driving paths based on the transition probability, includes:

[0019] Take any child node in each of the current child nodes as a target child node, and take the transition probability from the current position node to the target child node as the target transition probability. Determine the low transition probabilities that are less than the target transition probability among each of the transition probabilities.

[0020] Determine the path selection probabilities from the current position node to different current child nodes among the low transition probabilities and the target transition probability. Based on the path selection probabilities, select a target driving path for the intelligent mobile device in each of the driving paths.

[0021] Optionally, the determining the path selection probabilities from the current position node to different current child nodes among the low transition probabilities and the target transition probability includes:

[0022] In the case where there is a sudden obstacle in the surrounding environment information of the intelligent mobile device, determine the approaching rate of the sudden obstacle and the maximum safe speed of the intelligent mobile device. Determine the ratio of the approaching rate to the maximum safe speed as the speed safety margin, and based on the speed safety margin, determine the environmental risk coefficient.

[0023] Based on the environmental risk coefficient, the probability value of the low transition probability, and the probability value of the target transition probability, determine weight coefficients for the low transition probability and the target transition probability respectively.

[0024] Add up the weight coefficients to obtain the total weight coefficient, and randomly generate a random number between zero and the total weight coefficient.

[0025] Sort the low transition probability and the target transition probability in ascending order of the weight coefficient, calculate the cumulative weight coefficient of each sorted transition probability, and determine the target cumulative weight coefficient that is preferentially greater than the random number among each of the cumulative weight coefficients. Determine the transition probability corresponding to the target cumulative weight coefficient as the path selection probability from the current position node to different current child nodes.

[0026] Optionally, the real-time determination of the driving path network model for the intelligent mobile device based on the path planning information includes:

[0027] Determine the constraint conditions during the driving process of the intelligent mobile device, where the constraint conditions include at least one of the dynamic performance constraint of the intelligent mobile device, the passenger comfort constraint, and the road driving rule constraint.

[0028] Based on the path planning information, set multiple candidate driving paths for the intelligent mobile device in real time.

[0029] Based on each of the said constraint conditions, determine multiple driving paths for the intelligent mobile device in each of the said candidate driving paths, and construct a driving path network model based on each of the said driving paths.

[0030] Optionally, after selecting a target driving path for the intelligent mobile device in each of the said driving paths based on the said transition probability, the method further includes:

[0031] Determine the current state parameters of the intelligent mobile device and the driving path attribute information of the target driving path, where the current state parameters include wheelbase, heading, current speed, and current acceleration, and the driving path attribute information includes path curvature, lane driving rule information, and road adhesion coefficient;

[0032] Use the current sub-node in the target driving path as the preview point, determine the angle between the preview point and the heading, and determine the preview distance between the current position node of the intelligent mobile device and the preview point;

[0033] Based on the angle, the wheelbase, and the preview distance, determine the lateral control amount of the intelligent mobile device;

[0034] Determine the expected preview time for the intelligent mobile device to travel from the current position node to the preview point, and determine the speed limit of the intelligent mobile device based on the driving path attribute information;

[0035] Based on the preview distance, the expected preview time, and the speed limit, determine the target speed of the intelligent mobile device, and use a preset control algorithm to determine the longitudinal control amount of the intelligent mobile device based on the difference between the current speed and the target speed;

[0036] Based on the lateral control amount and the longitudinal control amount, perform driving control on the intelligent mobile device on the target driving path.

[0037] According to the second aspect of the present invention, there is provided a dynamic path planning device, including:

[0038] An information determination unit, configured to determine the path planning information of the intelligent mobile device in response to a path planning instruction of the intelligent mobile device, where the path planning information includes the current position, the position at the previous moment, the position to reach the end point, and the surrounding environment information;

[0039] A model determination unit, configured to determine a driving path network model for the intelligent mobile device in real time based on the path planning information, where the driving path network model is composed of multiple driving paths, and each driving path includes the same previous moment position node, current position node, and pre-arrival end position node, and at least one position node between the current position node and the pre-arrival end position node that is not completely the same;

[0040] A probability determination unit, configured to determine the current child node of the current position node on each driving path, determine node attribute information based on the current child node on each driving path, determine path attribute information of each driving path based on the current position node and the current child node on each driving path respectively, and determine the transition probability from the current position node to the current child node on each driving path according to the node attribute information and the path attribute information of each driving path;

[0041] A path selection unit, configured to select a target driving path for the intelligent mobile device from each driving path based on the transition probability.

[0042] According to a third aspect of the present invention, there is provided a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above dynamic path planning method is implemented.

[0043] According to a fourth aspect of the present invention, there is provided a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, the above dynamic path planning method is implemented.

[0044] According to a dynamic path planning method, device, and storage medium provided by the present invention, compared with the current method of using the driving path on the map as the static driving path of the intelligent mobile device, the present invention determines a driving path network model including multiple driving paths for the intelligent mobile device in real time based on the collected path planning information, calculates the transition probability from the current position node to the current child node on each driving path according to the node attribute information determined by the current child node on each driving path and the path attribute information of each driving path, and finally dynamically selects a target driving path for the intelligent mobile device according to the transition probability. By dynamically planning the driving path of the intelligent mobile device, it can respond to changes in the driving environment in real time, avoid the occurrence of emergencies, thereby improving the planning accuracy of the driving path and further improving driving safety. Description of the Drawings

[0045] The accompanying drawings described herein are used to provide a further understanding of the present invention and form a part of this application. The illustrative embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0046] Figure 1 A flowchart of a dynamic path planning method provided by an embodiment of the present invention is shown;

[0047] Figure 2 A schematic diagram of a driving path network model provided by an embodiment of the present invention is shown;

[0048] Figure 3 A flowchart of another dynamic path planning method provided by an embodiment of the present invention is shown;

[0049] Figure 4 A schematic structural diagram of a dynamic path planning device provided by an embodiment of the present invention is shown;

[0050] Figure 5 A schematic structural diagram of another dynamic path planning device provided by an embodiment of the present invention is shown;

[0051] Figure 6 A schematic physical structure diagram of a computer device provided by an embodiment of the present invention is shown. Detailed Embodiments

[0052] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other.

[0053] Currently, the method of using the driving path on the map as the static driving path of the intelligent mobile device cannot avoid unexpected situations on the road, resulting in a low accuracy in determining the driving path and there will be driving risks.

[0054] To solve the above problems, an embodiment of the present invention provides a dynamic path planning method, as Figure 1 shown, the method includes:

[0055] 101. In response to a path planning instruction of the intelligent mobile device, determine the path planning information of the intelligent mobile device, where the path planning information includes the current position, the position at the previous moment, the position of the pre-arrival destination, and the surrounding environment information.

[0056] Among them, the intelligent mobile device can be an intelligent mobile robot or an autonomous driving vehicle. The intelligent mobile robot can be applied to navigation tasks, disinfection tasks, and medicine delivery tasks in hospitals. For example, the intelligent mobile robot can automatically deliver medicines and instruments to the designated positions, reducing the workload of medical staff. The surrounding environment information includes weather information, road conditions information, obstacle information, traffic signal information, etc. around the intelligent mobile device.

[0057] For the embodiments of the present invention, in order to perform dynamic driving path planning for the intelligent mobile device, first, it is necessary to use data acquisition devices such as sensors to collect path planning information such as the current position information, the position information at the previous moment, the pre-arrival end position, and the surrounding environment information of the intelligent mobile device in real time. Then, based on the above path planning information, a driving path is planned for the intelligent mobile device in real time, so as to realize dynamic path planning, avoid the influence of events such as sudden obstacles, and improve driving safety.

[0058] For example, in the material transportation scenario in the infectious disease area of the hospital, when transporting sampling reagents for the isolation ward, it is necessary to avoid the contaminated area and plan a disinfection path. The intelligent mobile device responsible for material transportation formulates a one-way driving path of "clean area → semi-contaminated area → contaminated area" based on the path planning information. After the transportation is completed, a path back to the disinfection station is automatically generated, and the ultraviolet lamp disinfection program is linked. Another example is in the navigation scenario of a remote surgical robot. It is necessary to plan an optimal driving path for the surgical robot between the computer room and the operating room based on the path planning information of the surgical robot, and at the same time realize the function of avoiding medical staff in real time, so as to be able to shorten the operation time and improve the running stability of the surgical robot.

[0059] For example, in the autonomous driving scenario, based on path planning information such as the surrounding environment information and the real-time position of the road on which the autonomous driving vehicle travels, an optimal driving path can be dynamically planned for the autonomous driving vehicle in real time for obstacles such as sudden obstacles, so as to ensure the stable and safe operation of the autonomous driving vehicle.

[0060] 102. Based on the path planning information, determine a driving path network model for the intelligent mobile device in real time. Among them, the driving path network model is composed of multiple driving paths. Each driving path includes the same position node at the previous moment, the current position node, and the pre-arrival end position node, and at least one position node between the current position node and the pre-arrival end position node that is not completely the same.

[0061] For the embodiments of the present invention, for example, as Figure 2The figure shows one of the driving path network models. In this driving path network model, point A is the current position node of the intelligent mobile device, point O is the position node of the intelligent mobile device at the previous moment, point F is the pre - reached key position node of the intelligent mobile device, and B1, B2, B3, C1, C2, C3, D1, D2, D3, E1, E2 are the position nodes between the current position node and the pre - reached end position node respectively. In the driving path network model, for example, O→A→B1→C1→D1→E1→F is one of the driving paths, and O→A→B1→C2→D2→E1→F is also one of the driving paths. Multiple driving paths can be determined in the driving path network model.

[0062] 103. Determine the current child node of the current position node on each driving path, and determine the node attribute information based on the current child node on each driving path. Determine the path attribute information of each driving path based on the current position node and the current child node on each driving path respectively. According to the node attribute information and the path attribute information of each driving path, determine the transition probability from the current position node to the current child node on each driving path.

[0063] Among them, the position node at the previous moment corresponding to the current position node on each driving path is used as its corresponding current parent node, and the position node at the next moment is used as its corresponding current child node. For example, Figure 2 if point A is the current position node, then point O is the current parent node corresponding to point A, and B1, B2, B3 are the current child nodes corresponding to point A respectively; the node attribute information can be the total number of nodes of the current child node corresponding to the current position node on each driving path; the path attribute information can be the path length between the current position node and the current child node on each driving path respectively.

[0064] For the embodiments of the present invention, according to the total number of nodes and the path length between the current position node and each of its corresponding current child nodes, calculate the transition probability of the current position node transferring to each current child node respectively. Finally, select the final driving path of the intelligent mobile device according to the transition probability. Through the transition probability, the embodiments of the present invention can evaluate the driving profiles of different driving paths, so as to select the optimal path and ensure the selection accuracy of the path.

[0065] 104. Based on the transition probability, select a target driving path for the intelligent mobile device in each driving path.

[0066] For the embodiments of the present invention, among the transition probabilities from the current position node to the current child nodes on each driving path, the driving path corresponding to the maximum probability is determined as the target driving path of the intelligent mobile device. Thus, through the path planning information collected in real time, the embodiments of the present invention perform dynamic planning of the driving path for the intelligent mobile device by probability, can respond to environmental changes in real time, such as traffic signals, pedestrians, sudden obstacles, etc., and thus can accurately plan the driving route for the intelligent mobile device and improve driving safety.

[0067] According to a dynamic path planning method provided by the present invention, compared with the current method of taking the driving path on the map as the static driving path of the intelligent mobile device, the present invention determines in real time a driving path network model including multiple driving paths for the intelligent mobile device through the collected path planning information, and calculates the transition probabilities from the current position node to the current child nodes on each driving path according to the node attribute information determined by the current child nodes on each driving path and the path attribute information of each driving path, and finally dynamically selects the target driving path for the intelligent mobile device according to the transition probabilities. By performing dynamic planning of the driving path of the intelligent mobile device, it can respond to changes in the driving environment in real time, avoid the occurrence of sudden situations, and thus can improve the planning accuracy of the driving path and further improve driving safety.

[0068] Further, in order to better illustrate the above process of classifying data, as a refinement and extension of the above embodiments, the embodiments of the present invention provide another dynamic path planning method, as Figure 3 shown, the method includes:

[0069] 201. In response to a path planning instruction of the intelligent mobile device, determine the path planning information of the intelligent mobile device, where the path planning information includes the current position, the position at the previous moment, the position of the pre-arrival end point, and the surrounding environment information.

[0070] Specifically, when receiving a path planning instruction of the intelligent mobile device, use a data collection device such as a sensor to collect the path planning information of the intelligent mobile device in real time.

[0071] 202. Based on the path planning information, determine in real time a driving path network model for the intelligent mobile device, where the driving path network model is composed of multiple driving paths, and each driving path includes the same position node at the previous moment, the current position node, and the position node of the pre-arrival end point, and at least one position node between the current position node and the position node of the pre-arrival end point that are not completely the same.

[0072] For the embodiments of the present invention, after the path planning information is collected, it is necessary to determine a driving path network model for the intelligent mobile device. Based on this, step 202 specifically includes: determining the constraint conditions during the driving process of the intelligent mobile device, where the constraint conditions include at least one of the dynamic performance constraints of the intelligent mobile device, passenger comfort constraints, and road driving rule constraints; based on the path planning information, setting multiple candidate driving paths for the intelligent mobile device in real time; based on each of the constraint conditions, determining multiple driving paths for the intelligent mobile device in each of the candidate driving paths, and constructing a driving path network model based on each of the driving paths.

[0073] Among them, the dynamic performance constraints include the maximum speed, maximum acceleration, maximum steering angle, etc. of the intelligent mobile device to ensure that the intelligent mobile device is physically executable; the passenger comfort constraints refer to the acceleration change rate, path curvature, etc. to ensure the riding experience; the road driving rule constraints include traffic signals, lane rules, speed limits, etc. to ensure that the path is legal.

[0074] Specifically, a candidate path planning model can be trained and constructed in advance through a sample data set. The specific construction method includes: obtaining an initial path planning model and a sample data set, where the sample data set includes the path planning information of sample intelligent mobile devices with path labels; dividing the sample data set into training data and test data, training the initial path planning model with the training data, and testing the trained initial path planning model with the test data. Finally, the initial path planning model that meets the test conditions is used as the candidate path planning model. Among them, meeting the test conditions means that the number of training times reaches the requirement or the test accuracy reaches the requirement, etc. Further, the path planning information is input into the candidate path planning model for path planning, and the candidate path planning model can output multiple candidate driving paths between the current position and the pre-reached end position. The multiple candidate driving paths determined above are only set to avoid obstacles and adapt to the driving environment, and do not consider the constraint conditions such as the dynamic performance of the intelligent mobile device, the comfort of passengers, and road driving rules. Based on this, it is necessary to select the driving paths that meet the constraint conditions in each candidate driving path. For example, to meet the passenger comfort requirements, select the driving paths with a path curvature less than a preset threshold (where the preset threshold is set according to actual needs) in each candidate driving path; to meet the road driving rule constraints, select the paths that can pass, the paths that drive along the lane lines, etc. in each candidate driving path; to meet the dynamic performance constraints, select the paths that the intelligent mobile device can drive within the maximum speed, maximum acceleration, and maximum steering angle in each candidate driving path. Further, after determining each driving path, merge the same position nodes in each driving path, and thus a driving path network model constructed by multiple driving paths can be obtained.

[0075] 203. Determine the current child nodes of the current position node on each driving path, count the total number of nodes of the current child nodes on each driving path as node attribute information, and use the path lengths between the current position node and the current child nodes on each driving path as the path attribute information of each driving path respectively.

[0076] 204. According to the node attribute information and the path attribute information of each driving path, determine the tendency information of each driving path respectively, and determine the total tendency information based on the path attribute information of each driving path and the total number of position nodes in the driving path network model.

[0077] 205. Determine the tendency probability from the current position node to the current child nodes on each driving path according to the ratio of the tendency information of each driving path to the total tendency information.

[0078] For the embodiments of the present invention, the tendency probability P from the current position node to the current child nodes on each driving path can be determined according to the following formula c (x→y i )

[0079]

[0080] where x is the current position node, y i is the i-th current child node, n is the total number of nodes of the current child nodes, where n can also be the total number of the current child nodes and the current parent node, c i x is the path length between the current position node and the i-th current child node y i , f is the total number of nodes in the driving path network model, j is each node identifier in the driving path network model, q j is the total number of nodes of the parent node and the child node corresponding to node j, n + c i x is the tendency information, is the total tendency information. It should be noted that the above formula is only a schematic way of calculating the tendency probability, and the embodiments of the present invention are not limited to the above way to calculate the tendency probability.

[0081] 206. Determine the transition probability from the current position node to the current child nodes on each driving path based on the tendency probability corresponding to each driving path.

[0082] For the embodiments of the present invention, after determining the inclination probability corresponding to each driving path, it is necessary to determine the transition probability based on the inclination probability. Based on this, step 206 specifically includes: respectively determining the ratio of the path attribute information of each driving path to the total path attribute information commonly corresponding to each driving path as the correction parameter of the inclination probability corresponding to each driving path; based on the correction parameter, correcting the inclination probability corresponding to each driving path to obtain the transition probability from the current position node to the current sub-node on each driving path.

[0083] Specifically, the transition probability P(x→y from the current position node to the current sub-node on each driving path can be determined according to the following formula i ):

[0084]

[0085] where is the correction parameter.

[0086] In another embodiment of the present invention, the transition probability P(x→y can also be determined according to the following formula i ):

[0087]

[0088] 207. Based on the transition probability, select a target driving path for the intelligent mobile device in each driving path.

[0089] For the embodiments of the present invention, after determining the transition probabilities from the current position node to the current child nodes on each driving path, it is necessary to select a target driving path according to the transition probabilities. Based on this, step 207 specifically includes: taking any child node in each of the current child nodes as a target child node respectively, and taking the transition probability from the current position node to the target child node as the target transition probability, and determining the low transition probabilities that are less than the target transition probability among each of the transition probabilities; determining the path selection probabilities from the current position node to different current child nodes among the low transition probabilities and the target transition probability, and based on the path selection probabilities, selecting a target driving path for the intelligent mobile device in each of the driving paths. Among them, the method for determining the path selection probability among the low transition probability and the target transition probability includes: in the case that there is a sudden obstacle in the surrounding environment information of the intelligent mobile device, determining the approaching rate of the sudden obstacle and the maximum safe speed of the intelligent mobile device, determining the speed safety margin as the ratio of the approaching rate to the maximum safe speed, and determining the environmental risk coefficient based on the speed safety margin; determining the weight coefficients for the low transition probability and the target transition probability respectively based on the environmental risk coefficient, the probability value of the low transition probability, and the probability value of the target transition probability; adding up the weight coefficients to obtain the total weight coefficient, and randomly generating a random number between zero and the total weight coefficient; sorting the low transition probability and the target transition probability in ascending order of the weight coefficients, calculating the cumulative weight coefficients of each sorted transition probability, and determining the target cumulative weight coefficient that is preferentially greater than the random number among each of the cumulative weight coefficients, and determining the transition probability corresponding to the target cumulative weight coefficient as the path selection probability from the current position node to different current child nodes.

[0090] Among them, the sudden obstacles include people, animals, vehicles, etc. that suddenly appear around the intelligent mobile device.

[0091] Specifically, for example, in the driving path network model, if the current child nodes corresponding to the current position node are: B1, B2, and B3 respectively, and the transition probabilities from the current position node to the current child nodes B1, B2, and B3 are 1 / 6, 1 / 3, and 1 / 2 respectively. Taking the current child node B2 as the target child node as an example, since the transition probability less than that of the current child node B2 is the transition probability corresponding to the current child node B1, the transition probability corresponding to the current child node B1 is taken as the low transition probability. Then, a transition probability is selected from the target transition probability 1 / 3 corresponding to the current child node B2 and the low transition probability 1 / 6 corresponding to the current child node B1 as the path selection probability from the current position node to the current child node B2. Thus, the path selection probabilities from the current position node to each current child node can be determined in the above manner. Further, the method of selecting a probability between the target transition probability and the low transition probability is as follows: First, determine the environmental risk coefficient α according to the following formula:

[0092]

[0093] where β is the reference coefficient, v obs is the approaching speed of the sudden obstacle, v max is the maximum safe speed of the intelligent mobile device. Further, determine the weight coefficient according to the following formula:

[0094]

[0095] where k is the identifier of the low transition probability or the target transition probability, N is the total number of the low transition probability and the target transition probability, P k is the probability value of the low transition probability or the target transition probability, α is the environmental risk coefficient, ω kis the weight coefficient for the low transition probability or the target transition probability. Further, for example, if the low transition probabilities A, B, and C are 1 / 7, 1 / 6, and 1 / 8 respectively, the target transition probability D is 1 / 4, the weight coefficients corresponding to the respective low transition probabilities are 0.3, 0.2, and 0.1, the weight coefficient corresponding to the target transition probability is 0.4, the total weight coefficient is 1, the random number selected between 0 and 1 is 0.5, and the sorting result obtained by sorting the respective transition probabilities according to the magnitude of the weight coefficients is: low transition probability C (0.1), low transition probability B (0.2), low transition probability A (0.3), target transition probability D (0.4), the cumulative weight coefficient corresponding to the low transition probability C is 0.1, the cumulative weight coefficient corresponding to the low transition probability B is 0.3, and the cumulative weight coefficient corresponding to the low transition probability C is 0.6. Thus, it can be seen that the cumulative weight coefficient corresponding to the low transition probability C first exceeds the random number 0.5. Therefore, the probability value 1 / 8 corresponding to the low transition probability C is determined as the path selection probability from the current position node to the target current child node. Thus, the path selection probability from the current position node to each current child node can be determined in the above manner. Finally, the maximum path selection probability is determined among the path selection probabilities corresponding to each current child node, and the path corresponding to the maximum path selection probability is determined as the target driving path for the intelligent mobile device to pre-drive. It should be noted that the above example is only illustrative and does not limit the embodiments of the present application.

[0096] Further, after determining the target driving path of the intelligent mobile device, it is also necessary to control the intelligent mobile device to drive on the target driving path. Based on this, the method includes: determining the current state parameters of the intelligent mobile device and the driving path attribute information of the target driving path, where the current state parameters include wheelbase, heading, current speed, and current acceleration, and the driving path attribute information includes path curvature, lane driving rule information, and road adhesion coefficient; using the current child node in the target driving path as the preview point, determining the angle between the preview point and the heading, and determining the preview distance between the current position node of the intelligent mobile device and the preview point; determining the lateral control amount of the intelligent mobile device based on the angle, the wheelbase, and the preview distance; determining the expected preview time for the intelligent mobile device to travel from the current position node to the preview point, and determining the speed limit value of the intelligent mobile device based on the driving path attribute information; determining the target speed of the intelligent mobile device based on the preview distance, the expected preview time, and the speed limit value, and using a preset control algorithm to determine the longitudinal control amount of the intelligent mobile device based on the difference between the current speed and the target speed; and performing driving control on the intelligent mobile device on the target driving path based on the lateral control amount and the longitudinal control amount.

[0097] Specifically, the lateral control amount δ of the intelligent mobile device is determined according to the following formula:

[0098]

[0099] Where L is the wheelbase of the intelligent mobile device, γ is the angle between the preview point and the course, and d is the preview distance between the current position node of the intelligent mobile device and the preview point. The lateral control amount in the embodiments of the present invention is the amount used to adjust the lateral movement of the intelligent mobile device, such as the steering wheel angle, etc. Further, the ratio of the preview distance to the expected preview time is determined as the target speed of the intelligent mobile device, and it is necessary to ensure that the target speed is less than or equal to the speed limit of the intelligent mobile device. Then, the absolute value of the difference between the current speed and the target speed of the intelligent mobile device is determined as the difference e v , and the longitudinal control amount a of the intelligent mobile device is determined according to the following formula cmd :

[0100]

[0101] Where K p is the proportional gain for controlling the response speed, K i is the integral gain for eliminating the steady-state error, and K d is the derivative gain for suppressing overshoot. The longitudinal control amount in the embodiments of the present invention refers to the control parameter of the intelligent mobile device in the driving direction, such as acceleration, etc. Finally, based on the lateral control amount and the longitudinal control amount, a control command is generated, and the intelligent mobile device is controlled to drive based on the control command.

[0102] According to another dynamic path planning method provided by the present invention, compared with the current method of using the driving path on the map as the static driving path of the intelligent mobile device, the present invention determines a driving path network model including multiple driving paths for the intelligent mobile device in real time through the collected path planning information, and calculates the transfer probability from the current position node to the current sub-node on each driving path according to the node attribute information determined by the current sub-node on each driving path and the path attribute information of each driving path. Finally, the target driving path is dynamically selected for the intelligent mobile device according to the transfer probability. By dynamically planning the driving path of the intelligent mobile device, it can respond to changes in the driving environment in real time, avoid the occurrence of emergencies, thereby improving the planning accuracy of the driving path and further improving the driving safety.

[0103] Further, as a Figure 1 specific implementation, the embodiments of the present invention provide a dynamic path planning device, as shown in Figure 4 , the device includes: an information determination unit 31, a model determination unit 32, a probability determination unit 33, and a path selection unit 34.

[0104] The information determination unit 31 can be configured to determine the path planning information of the intelligent mobile device in response to a path planning instruction of the intelligent mobile device, where the path planning information includes the current position, the position at the previous moment, the position of the destination to be reached, and the surrounding environment information.

[0105] The model determination unit 32 can be configured to determine a driving path network model for the intelligent mobile device in real time based on the path planning information, where the driving path network model is composed of multiple driving paths, and each driving path includes the same position node at the previous moment, the current position node, and the position node of the destination to be reached, and at least one position node between the current position node and the position node of the destination to be reached that is not completely the same.

[0106] The probability determination unit 33 can be configured to determine the current child node of the current position node on each driving path, determine node attribute information based on the current child nodes on each driving path, determine path attribute information of each driving path based on the current position node and the current child nodes on each driving path, and determine the transition probability from the current position node to the current child nodes on each driving path according to the node attribute information and the path attribute information of each driving path.

[0107] The path selection unit 34 can be configured to select a target driving path for the intelligent mobile device from each driving path based on the transition probability.

[0108] In a specific application scenario, in order to determine the transition probability from the current position node to the current child nodes on each driving path, as Figure 5 shown, the probability determination unit 33 includes a statistics module 331 and a first determination module 332.

[0109] The statistics module 331 can be configured to count the total number of nodes of the current child nodes on each driving path as the node attribute information, and use the path lengths between the current position node and the current child nodes on each driving path as the path attribute information of each driving path respectively.

[0110] The first determination module 332 can be configured to determine the tendency information of each driving path respectively according to the node attribute information and the path attribute information of each driving path, and determine the total tendency information based on the path attribute information of each driving path and the total number of position nodes in the driving path network model.

[0111] The first determination module 332 may specifically be configured to determine the tendency probability from the current position node to the current sub-node on each of the driving paths according to the ratio of the tendency information of each driving path to the total tendency information.

[0112] The first determination module 332 may specifically be configured to determine the transition probability from the current position node to the current sub-node on each of the driving paths based on the tendency probability corresponding to each driving path.

[0113] In a specific application scenario, in order to determine the transition probability from the current position node to the current sub-node on each driving path, the first determination module 332 may specifically be configured to respectively determine the ratio of the path attribute information of each driving path to the total path attribute information jointly corresponding to each driving path as the correction parameter of the tendency probability corresponding to each driving path; based on the correction parameter, correct the tendency probability corresponding to each driving path to obtain the transition probability from the current position node to the current sub-node on each driving path.

[0114] In a specific application scenario, in order to select a target driving path for the intelligent mobile device in each driving path, the path selection unit 34 includes a second determination module 341 and a selection module 342.

[0115] The second determination module 341 may be configured to respectively use any sub-node in each of the current sub-nodes as a target sub-node, and use the transition probability from the current position node to the target sub-node as the target transition probability, and determine the low transition probability less than the target transition probability among each of the transition probabilities.

[0116] The selection module 342 may be configured to determine the path selection probability from the current position node to different current sub-nodes among the low transition probability and the target transition probability, and based on the path selection probability, select a target driving path for the intelligent mobile device in each driving path.

[0117] In a specific application scenario, in order to determine the path selection probability from the current position node to different current child nodes between the low transition probability and the target transition probability, the selection module 342 can specifically be used to determine the approaching rate of the sudden obstacle and the maximum safe speed of the intelligent mobile device when there is a sudden obstacle in the surrounding environment information of the intelligent mobile device, determine the ratio of the approaching rate to the maximum safe speed as the speed safety margin, and determine the environmental risk coefficient based on the speed safety margin; determine the weight coefficients for the low transition probability and the target transition probability respectively based on the environmental risk coefficient, the probability value of the low transition probability, and the probability value of the target transition probability; add up the weight coefficients to obtain the total weight coefficient, and randomly generate a random number between zero and the total weight coefficient; sort the low transition probability and the target transition probability in ascending order of the weight coefficients, calculate the cumulative weight coefficients of each sorted transition probability, and determine the target cumulative weight coefficient that is preferentially greater than the random number in each cumulative weight coefficient, and determine the transition probability corresponding to the target cumulative weight coefficient as the path selection probability from the current position node to different current child nodes.

[0118] In a specific application scenario, in order to determine the driving path network model for the intelligent mobile device in real time, the model determination unit 32 includes a third determination module 321 and a setting module 322.

[0119] The third determination module 321 can be used to determine the constraint conditions during the driving process of the intelligent mobile device, where the constraint conditions include at least one of the dynamic performance constraint of the intelligent mobile device, the passenger comfort constraint, and the road driving rule constraint.

[0120] The setting module 322 can be used to set multiple candidate driving paths for the intelligent mobile device in real time based on the path planning information.

[0121] The third determination module 321 can also be used to determine multiple driving paths for the intelligent mobile device in each candidate driving path based on each constraint condition, and construct a driving path network model based on the driving paths.

[0122] In a specific application scenario, in order to perform driving control on the intelligent mobile device on the target driving path, the device further includes a driving control unit 35.

[0123] The driving control unit 35 can be used to determine the current state parameters of the intelligent mobile device and the driving path attribute information of the target driving path. Among them, the current state parameters include wheelbase, heading, current speed, and current acceleration, and the driving path attribute information includes path curvature, lane driving rule information, and road adhesion coefficient; taking the current sub-node in the target driving path as the preview point, determining the angle between the preview point and the heading, and determining the preview distance between the current position node of the intelligent mobile device and the preview point; based on the angle, the wheelbase, and the preview distance, determining the lateral control amount of the intelligent mobile device; determining the expected preview time for the intelligent mobile device to move from the current position node to the preview point, and based on the driving path attribute information, determining the speed limit value of the intelligent mobile device; based on the preview distance, the expected preview time, and the speed limit value, determining the target speed of the intelligent mobile device, and based on the difference between the current speed and the target speed, using a preset control algorithm to determine the longitudinal control amount of the intelligent mobile device; based on the lateral control amount and the longitudinal control amount, performing driving control on the intelligent mobile device on the target driving path.

[0124] It should be noted that for other corresponding descriptions of each functional module involved in the dynamic path planning device provided in the embodiment of the present invention, reference can be made to Figure 1 the corresponding description of the method shown, which will not be elaborated here.

[0125] Based on the above as Figure 1For the method described above, correspondingly, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the following steps are implemented: In response to a path planning instruction of an intelligent mobile device, determine the path planning information of the intelligent mobile device, where the path planning information includes the current position, the position at the previous moment, the position of the pre-arrival destination, and the surrounding environment information; Based on the path planning information, determine a driving path network model for the intelligent mobile device in real time, where the driving path network model is composed of multiple driving paths, and each driving path includes the same position node at the previous moment, the current position node, and the position node of the pre-arrival destination, and at least one position node between the current position node and the position node of the pre-arrival destination that are not completely the same; Determine the current sub-node of the current position node on each driving path, and determine the node attribute information based on the current sub-nodes on each driving path. Determine the path attribute information of each driving path based on the current position node and the current sub-nodes on each driving path respectively. According to the node attribute information and the path attribute information of each driving path, determine the transition probability from the current position node to the current sub-nodes on each driving path; Based on the transition probability, select a target driving path for the intelligent mobile device in each driving path.

[0126] Based on the above as Figure 1 shown in the method and as Figure 4 shown in the device embodiment, an embodiment of the present invention further provides a physical structure diagram of a computer device, as Figure 6As shown in the figure, the computer device includes: a processor 41, a memory 42, and a computer program stored in the memory 42 and executable on the processor. Both the memory 42 and the processor 41 are provided on a bus 43. When the processor 41 executes the program, the following steps are implemented: In response to a path planning instruction of an intelligent mobile device, determine the path planning information of the intelligent mobile device, where the path planning information includes the current position, the position at the previous moment, the position of the pre-arrival end point, and the surrounding environment information; Based on the path planning information, determine a driving path network model for the intelligent mobile device in real time, where the driving path network model is composed of multiple driving paths, and each driving path includes the same position node at the previous moment, the current position node, and the position node of the pre-arrival end point, and at least one position node between the current position node and the position node of the pre-arrival end point that are not completely the same; Determine the current sub-node of the current position node on each driving path, and determine the node attribute information based on the current sub-node on each driving path. Determine the path attribute information of each driving path based on the current position node and the current sub-node on each driving path respectively. According to the node attribute information and the path attribute information of each driving path, determine the transition probability from the current position node to the current sub-node on each driving path; Based on the transition probability, select a target driving path for the intelligent mobile device on each driving path.

[0127] Through the technical solution of the present invention, the present invention determines a driving path network model including multiple driving paths for an intelligent mobile device in real time through the collected path planning information, and calculates the transition probability from the current position node to the current sub-node on each driving path according to the node attribute information determined by the current sub-node on each driving path and the path attribute information of each driving path. Finally, the target driving path is dynamically selected for the intelligent mobile device according to the transition probability. By dynamically planning the driving path of the intelligent mobile device, it can respond to changes in the driving environment in real time, avoid the occurrence of emergencies, thereby improving the planning accuracy of the driving path and further improving the driving safety.

[0128] Obviously, those skilled in the art should understand that the various modules or steps of the present invention described above can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Optionally, they can be implemented by program codes executable by the computing device. Thus, they can be stored in a storage device and executed by the computing device. And in some cases, the steps shown or described can be executed in a sequence different from that here, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module for implementation. In this way, the present invention is not limited to any specific combination of hardware and software.

[0129] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A dynamic path planning method, characterized in that, Including: In response to a path planning instruction of an intelligent mobile device, determining path planning information of the intelligent mobile device, where the path planning information includes a current position, a position at the previous moment, a position of a pre-arrival destination, and surrounding environment information; Based on the path planning information, determining a driving path network model for the intelligent mobile device in real time, where the driving path network model is composed of multiple driving paths, and each driving path includes the same position node at the previous moment, the current position node, and the position node of the pre-arrival destination, and at least one position node between the current position node and the position node of the pre-arrival destination that are not completely the same; Determining a current child node of the current position node on each driving path, and determining node attribute information based on the current child nodes on each driving path, determining path attribute information of each driving path based on the current position node and the current child nodes on each driving path respectively, and determining a transfer probability from the current position node to the current child nodes on each driving path according to the node attribute information and the path attribute information of each driving path; Based on the transfer probability, selecting a target driving path for the intelligent mobile device in each driving path.

2. The method according to claim 1, characterized in that, The determining node attribute information based on the current child nodes on each driving path, determining path attribute information of each driving path based on the current position node and the current child nodes on each driving path respectively, and determining a transfer probability from the current position node to the current child nodes on each driving path according to the node attribute information and the path attribute information of each driving path includes: Counting the total number of nodes of the current child nodes on each driving path as the node attribute information, and taking the path lengths between the current position node and the current child nodes on each driving path as the path attribute information of each driving path respectively; According to the node attribute information and the path attribute information of each driving path, determining the tendency information of each driving path respectively, and determining the total tendency information based on the path attribute information of each driving path and the total number of position nodes in the driving path network model; Determining a tendency probability from the current position node to the current child nodes on each driving path according to the ratio of the tendency information of each driving path to the total tendency information; Based on the tendency probability corresponding to each driving path, determining a transfer probability from the current position node to the current child nodes on each driving path.

3. The method according to claim 2, wherein The determining a transfer probability from the current position node to the current child nodes on each driving path based on the tendency probability corresponding to each driving path includes: Respectively determining the ratio of the path attribute information of each driving path to the total path attribute information commonly corresponding to each driving path as a correction parameter of the tendency probability corresponding to each driving path; Based on the correction parameter, correct the tendency probability corresponding to each driving path to obtain the transition probability from the current position node to the current sub-node on each driving path.

4. The method according to claim 1, wherein The selecting, for the intelligent mobile device, a target driving path in each driving path based on the transition probability includes: Taking any sub-node in each current sub-node as a target sub-node respectively, and taking the transition probability from the current position node to the target sub-node as the target transition probability, and determining a low transition probability smaller than the target transition probability among each transition probability; Determine the path selection probability from the current position node to different current sub-nodes among the low transition probability and the target transition probability, and based on the path selection probability, select a target driving path for the intelligent mobile device in each driving path.

5. The method according to claim 4, characterized in that, The determining the path selection probability from the current position node to different current sub-nodes among the low transition probability and the target transition probability includes: When there is a sudden obstacle in the surrounding environment information of the intelligent mobile device, determine the approaching rate of the sudden obstacle and the maximum safe speed of the intelligent mobile device, determine the ratio of the approaching rate to the maximum safe speed as the speed safety margin, and based on the speed safety margin, determine the environmental risk coefficient; Based on the environmental risk coefficient, the probability value of the low transition probability, and the probability value of the target transition probability, determine weight coefficients for the low transition probability and the target transition probability respectively; Add up each weight coefficient to obtain the total weight coefficient, and randomly generate a random number between zero and the total weight coefficient; Sort the low transition probability and the target transition probability in ascending order of the weight coefficient, calculate the cumulative weight coefficient of each sorted transition probability, and determine a target cumulative weight coefficient that is preferentially greater than the random number among each cumulative weight coefficient, and determine the transition probability corresponding to the target cumulative weight coefficient as the path selection probability from the current position node to different current sub-nodes.

6. The method according to claim 1, wherein The real-time determination of the driving path network model for the intelligent mobile device based on the path planning information includes: Determine the constraint conditions during the driving process of the intelligent mobile device, where the constraint conditions include at least one of the dynamic performance constraint of the intelligent mobile device, the passenger comfort constraint, and the road driving rule constraint; Based on the path planning information, set multiple candidate driving paths for the intelligent mobile device in real time; Based on each constraint condition, determine multiple driving paths for the intelligent mobile device in each candidate driving path, and construct a driving path network model based on each driving path.

7. The method according to claim 1, characterized in that, After selecting a target driving path for the intelligent mobile device in each driving path based on the transition probability, the method further includes: Determine the current state parameters of the intelligent mobile device and the driving path attribute information of the target driving path, where the current state parameters include wheelbase L, heading, current speed, and current acceleration, and the driving path attribute information includes path curvature, lane driving rule information, and road adhesion coefficient; Take the current sub-node in the target driving path as the preview point, determine the angle between the preview point and the heading, and determine the preview distance between the current position node of the intelligent mobile device and the preview point; Based on the angle, the wheelbase, and the preview distance, determine the lateral control amount of the intelligent mobile device; Determine the expected preview time for the intelligent mobile device to move from the current position node to the preview point, and based on the driving path attribute information, determine the speed limit of the intelligent mobile device; Based on the preview distance, the expected preview time, and the speed limit, determine the target speed of the intelligent mobile device, and based on the difference between the current speed and the target speed, use a preset control algorithm to determine the longitudinal control amount of the intelligent mobile device; Based on the lateral control amount and the longitudinal control amount, perform driving control on the intelligent mobile device on the target driving path.

8. A dynamic path planning device, characterized in that, Includes: An information determination unit, configured to determine the path planning information of the intelligent mobile device in response to a path planning instruction of the intelligent mobile device, where the path planning information includes the current position, the position at the previous moment, the position to reach the end point, and the surrounding environment information; A model determination unit, configured to determine a driving path network model for the intelligent mobile device in real time based on the path planning information, where the driving path network model is composed of multiple driving paths, and each driving path includes the same position node at the previous moment, current position node, and position node to reach the end point, and at least one position node between the current position node and the position node to reach the end point that is not completely the same; A probability determination unit, configured to determine the current sub-node of the current position node on each driving path, and determine the node attribute information based on the current sub-node on each driving path, determine the path attribute information of each driving path based on the current position node and the current sub-node on each driving path respectively, and determine the transition probability from the current position node to the current sub-node on each driving path according to the node attribute information and the path attribute information of each driving path; A path selection unit, configured to select a target driving path for the intelligent mobile device from each driving path based on the transition probability.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images