CN-121995791-A - Behavior simulation method of mobile intelligent body applied to emergency rescue scene

CN121995791ACN 121995791 ACN121995791 ACN 121995791ACN-121995791-A

Abstract

The embodiment of the invention discloses a behavior simulation method of a mobile intelligent agent applied to an emergency rescue scene. The method comprises the steps of generating an initial simulation environment according to scene environment information and simulation configuration information, locally extracting environment data in the initial simulation environment according to intelligent agent description information corresponding to a mobile intelligent agent to obtain a real-time environment information set, generating environment state characteristics and intelligent agent state characteristics according to the real-time environment information set and intelligent agent real-time states corresponding to the mobile intelligent agent, generating behavior planning information aiming at the mobile intelligent agent according to the environment state characteristics and intelligent agent state characteristics, and performing behavior simulation on the mobile intelligent agent according to the behavior planning information and intelligent agent real-time azimuth. According to the method and the system, the robustness and the accuracy of the behavior planning of the movable robot in different emergency rescue scenes are improved.

Inventors

ZHANG ENYU
YAO HUI
LIU BING
ZHANG HENG
REN JUN
Xia Shoufei

Assignees

北京华创宇为科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260126

Claims (10)

1. The behavior simulation method of the mobile intelligent agent applied to the emergency rescue scene is characterized by comprising the following steps of: Generating an initial simulation environment according to scene environment information and simulation configuration information, wherein the scene environment information represents regional environment information corresponding to an emergency rescue scene, and the initial simulation environment contains mobile intelligent bodies which are subjected to position initialization and are to be subjected to behavior simulation; According to the intelligent agent description information corresponding to the mobile intelligent agent, carrying out local extraction of environmental data in the initial simulation environment to obtain a real-time environmental information set, wherein the intelligent agent description information comprises intelligent agent real-time azimuth and sensor specification information, and the sensor specification information represents the sensor specification of a virtual environmental sensor corresponding to the mobile intelligent agent; Generating environmental state characteristics and intelligent body state characteristics according to the real-time environment information set and intelligent body real-time states corresponding to the mobile intelligent body; generating behavior planning information for the mobile agent according to the environmental state characteristics and the agent state characteristics; and carrying out behavior simulation on the mobile intelligent body according to the behavior planning information and the intelligent body real-time azimuth.
2. The behavioral simulation method of a mobile agent for use in an emergency rescue scenario of claim 1, the method further comprising: Generating behavior evaluation information aiming at the simulation behavior; generating a behavior sample according to the behavior evaluation information, the behavior planning information, the intelligent agent description information and the scene environment information; And archiving and storing the behavior samples.
3. The method for simulating the behavior of a mobile agent applied to an emergency rescue scene according to claim 2, wherein the scene environment information comprises a scene type and a scene element information set, the scene element information comprises a scene element type, a scene element position and a scene element specification parameter, the simulation configuration information comprises an agent type, a simulation precision parameter and a simulation scale parameter, and the generating an initial simulation environment according to the scene environment information and the simulation configuration information comprises: generating a simulation space according to the simulation scale parameters; determining a sensing range according to the type of the intelligent body and the specification information of the sensor, wherein the sensing range represents the sensing range of a virtual environment sensor corresponding to the mobile intelligent body; according to the perception range, carrying out space voxelization on the simulation space to obtain a voxelized simulation space, wherein the voxelized simulation space comprises a voxel block set; Performing eight-way space segmentation on the voxelized simulation space by taking the initial loading position of the mobile intelligent agent as the center to obtain a voxel block group set, wherein each voxel block in the voxel block group corresponds to the same segmentation direction; generating a voxel block tree set according to the initial loading position and the voxel block group set, wherein the voxel block tree takes the initial loading position as a root node; And generating the initial simulation environment according to the voxel block tree set, the scene environment information and the simulation configuration information.
4. The behavioral simulation method of a mobile agent for use in an emergency rescue scene according to claim 3, wherein said generating the initial simulation environment from the set of voxel block trees, the scene environment information, and the simulation configuration information comprises: for each voxel block tree in the set of voxel block trees, performing the following processing steps: Performing breadth-first traversal on the voxel block tree to obtain a tree node group sequence, wherein tree nodes in the tree node group correspond to the same order number, and the tree nodes correspond to the voxel blocks; According to the tree node group sequence and the scene element information set, the following scene loading steps are executed: taking out a tree node group corresponding to a target order from the tree node group sequence as a target tree node group, wherein the target order is initially 1; Performing element matching on the target tree nodes in the target tree node group and the scene element information in the scene element information set according to the scene element positions included in the scene element information to obtain a matching information group; Carrying out local scene loading according to the matching information group and the simulation precision parameter to obtain a target simulation environment; ending the scene loading step in response to the target order being equal to the tree depth corresponding to the voxel block tree; Responding to the tree depth corresponding to the voxel block tree, increasing the target order, taking the target order after increasing the target order as the target order, taking a scene element information set with scene element information corresponding to the matched information set removed as a scene element information set, and re-executing the scene loading step; and generating a local simulation environment in the initial simulation environment, which corresponds to the voxel block tree in the segmentation direction, according to the obtained target simulation environment set.
5. The behavioral simulation method of a mobile agent applied to an emergency rescue scene according to claim 4, wherein the performing local extraction of environmental data in the initial simulation environment according to agent description information corresponding to the mobile agent to obtain a real-time environmental information set includes: Determining a sensor azimuth corresponding to the virtual environment sensor according to the real-time azimuth of the intelligent agent included in the intelligent agent description information; determining a voxel block tree associated with the sensor orientation in the set of voxel block trees as a candidate voxel block tree; according to the sensor azimuth and the sensor specification information, sub-tree segmentation is carried out on the candidate voxel block tree to obtain a target voxel block set; And extracting scene elements corresponding to the target voxel block in the target voxel block set under a fixed view angle according to the sensor azimuth to obtain the real-time environment information set.
6. The behavioral simulation method of a mobile agent for an emergency rescue scene according to claim 5, wherein generating an environmental state feature and an agent state feature from the real-time environmental information set and an agent real-time state corresponding to the mobile agent comprises: Determining a first environment area and a second environment area according to the real-time environment information set, wherein the first environment area represents an intersection acquisition area corresponding to the virtual environment sensor, and the second environment area represents a non-intersection acquisition area corresponding to the virtual environment sensor; Performing cross-linking feature extraction on real-time environment information corresponding to the first environment region in the real-time environment information set to obtain a first environment feature; Extracting features of real-time environment information corresponding to the second environment area in the real-time environment information set to obtain second environment features; According to the real-time azimuth of the intelligent body, performing aerial view feature projection on the first environmental feature and the second environmental feature to obtain aerial view environmental features; performing multi-scale feature extraction on the aerial view environmental features to obtain environmental state features; and performing state mapping on the real-time state of the intelligent agent to obtain the state characteristics of the intelligent agent.
7. The behavioral simulation method of a mobile agent for use in an emergency rescue scenario of claim 6, wherein the generating behavioral planning information for the mobile agent based on the environmental state characteristics and the agent state characteristics comprises: performing feature segmentation on the environmental state features to obtain a sub-environmental state feature set; carrying out parallel feature coding on the sub-environment state features in the sub-environment state feature set to obtain a coded environment state feature set; performing environment state rough classification on each sub-environment state feature in the encoded environment state feature set to obtain an environment classification type set; Generating a weight feature matrix according to the environment classification type set; According to the weight feature matrix, carrying out feature weighting on the environmental state features to obtain weighted environmental state features; And generating the behavior planning information according to the weighted environmental state characteristics, the intelligent agent state characteristics and a pre-trained behavior planning model.
8. The utility model provides a be applied to action simulation analogue means of mobile agent of emergency rescue scene which characterized in that includes: The first generation unit is configured to generate an initial simulation environment according to scene environment information and simulation configuration information, wherein the scene environment information represents regional environment information corresponding to an emergency rescue scene, and the initial simulation environment comprises mobile intelligent bodies which are subjected to position initialization and are to be subjected to behavior simulation; The environment data local extraction unit is configured to perform environment data local extraction in the initial simulation environment according to the intelligent agent description information corresponding to the mobile intelligent agent to obtain a real-time environment information set, wherein the intelligent agent description information comprises intelligent agent real-time azimuth and sensor specification information, and the sensor specification information represents the sensor specification of a virtual environment sensor corresponding to the mobile intelligent agent; The second generating unit is configured to generate environmental state characteristics and intelligent body state characteristics according to the real-time environmental information set and intelligent body real-time states corresponding to the mobile intelligent body; A third generation unit configured to generate behavior planning information for the mobile agent according to the environmental state characteristics and the agent state characteristics; and the simulation unit is configured to perform behavior simulation on the mobile intelligent body according to the behavior planning information and the intelligent body real-time azimuth.
9. An electronic device, comprising: One or more processors; A storage device having one or more programs stored thereon; when executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1 to 7.
10. A computer readable medium, characterized in that a computer program is stored thereon, wherein the computer program, when executed by a processor, implements the method according to any of claims 1 to 7.

Description

Behavior simulation method of mobile intelligent body applied to emergency rescue scene Technical Field The embodiment of the disclosure relates to the technical field of computers, in particular to the field of behavior simulation, and particularly relates to a behavior simulation method of a mobile intelligent body applied to an emergency rescue scene. Background Due to the limitation of the rescue environment, there are areas in the emergency rescue scene where rescue workers cannot directly rescue and coverage is not available, so that the movable robot is often required to perform corresponding emergency rescue (such as material transportation, environment exploration and the like). However, since the rescue environment is often severe, it is difficult to effectively plan the corresponding robot behavior in the real environment. Disclosure of Invention The disclosure is in part intended to introduce concepts in a simplified form that are further described below in the detailed description. The disclosure is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter. Some embodiments of the present disclosure propose a behavior simulation method of a mobile agent applied to an emergency rescue scene to solve the technical problems mentioned in the background section above. According to the first aspect, some embodiments of the present disclosure provide a behavior simulation method of a mobile agent applied to an emergency rescue scene, the method including generating an initial simulation environment according to scene environment information and simulation configuration information, wherein the scene environment information characterizes region environment information corresponding to the emergency rescue scene, the initial simulation environment includes mobile agents to be subjected to behavior simulation after position initialization, local extraction of environment data in the initial simulation environment according to agent description information corresponding to the mobile agents is performed to obtain a real-time environment information set, the agent description information includes an agent real-time azimuth and sensor specification information, the sensor specification information characterizes a sensor specification of a virtual environment sensor corresponding to the mobile agents, environmental state features and agent state features are generated according to the real-time environment information set and the real-time state of the mobile agents, behavior information aiming at the mobile agents is generated according to the environmental state features and the agent state features, and planning of behavior simulation is performed on the mobile agents according to the behavior information and the agent description information. In a second aspect, some embodiments of the present disclosure provide a behavior simulation device for a mobile agent applied to an emergency rescue scene, where the device includes a first generation unit configured to generate an initial simulation environment according to scene environment information and simulation configuration information, where the scene environment information characterizes region environment information corresponding to the emergency rescue scene, the initial simulation environment includes a mobile agent to be subjected to behavior simulation after being subjected to position initialization, an environment data local extraction unit configured to perform environment data local extraction in the initial simulation environment according to agent description information corresponding to the mobile agent to obtain a real-time environment information set, where the agent description information includes an agent real-time azimuth and sensor specification information, the sensor specification information characterizes a virtual environment sensor corresponding to the mobile agent, a second generation unit configured to generate an environment state feature and an agent state feature according to the real-time environment information set and an agent real-time state corresponding to the mobile agent, and a third generation unit configured to perform behavior simulation on the mobile agent according to the environment state feature and the behavior feature, and the simulation unit configured to plan the behavior feature of the mobile agent. In a third aspect, some embodiments of the present disclosure provide an electronic device comprising one or more processors, and storage means having one or more programs stored thereon, which when executed by the one or more processors, cause the one or more processors to implement the method described in any of the implementations of the first aspect. In a fourth aspect, some embodiments of the present disclosure provide a computer readable medium having a computer program stored thereon, wherein