CN-122018560-A - Machine dog formation collaborative operation method and system in disaster environment

CN122018560ACN 122018560 ACN122018560 ACN 122018560ACN-122018560-A

Abstract

The invention relates to the field of robot navigation, in particular to a robot dog formation collaborative operation method and system in a disaster environment. The method comprises the steps of obtaining future intention information of a pilot dog and a local environment model perceived by the pilot dog, fusing and constructing a prospective environment model, calculating a link quality score of the pilot dog on the intention of the pilot dog based on the model, constructing a potential field, solving a gradient to obtain a communication gradient driving force, switching the operation mode of the pilot dog according to the operation state of the pilot dog, fusing the communication gradient driving force, communication attraction and obstacle repulsion pointing to the pilot dog in the corresponding mode, calculating navigation resultant force and controlling cooperative operation of the pilot dog. According to the invention, through a prospective communication quality gradient field, the relay dog pre-judges and avoids a communication blind area, and the problem that formation communication is easy to interrupt in a disaster environment is solved by combining with the mode switching adaptation requirement, so that the communication continuity and robustness are improved, and a support is provided for the efficient collaborative operation of the machine dog.

Inventors

ZHAI HUILIN
ZHANG LINGLING
SHI HONGRUI
WANG FAQING

Assignees

博浩科技有限公司
杭州乾朗九章科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260413

Claims (10)

1. The machine dog formation collaborative operation method under the disaster environment is characterized by comprising the following steps: acquiring future intention information broadcast by a pilot robot dog and a local environment model perceived by the relay robot dog, and fusing to construct a prospective environment model; Calculating a link quality score of the relay machine dog on future intention information of the pilot machine dog at the current position based on the prospective environment model, and constructing a potential field of the link quality score; Calculating the gradient of the potential field of the link quality score to obtain a communication gradient impetus; Switching an operation mode according to the operation state of the pilot robot dog, under the operation mode, fusing the communication gradient pushing force, the communication attraction based on the current position of the pilot robot dog and the obstacle repulsive force, calculating to obtain a final navigation resultant force, and controlling the relay robot dog to execute cooperative operation.
2. The method for collaborative operation of a machine dog formation in a disaster environment according to claim 1, wherein the obtaining future intention information broadcast by a pilot machine dog comprises: a probabilistic path cone broadcast by a pilot robot dog is obtained, the probabilistic path cone being comprised of a central path and a set of time-extended variance parameters.
3. The method for collaborative work on a machine dog formation in a disaster environment according to claim 1, wherein the fusing builds a prospective environmental model further comprising: the relay robot dog builds a three-dimensional occupied grid map in real time by using a sensor; and carrying out semantic attribution on the voxels in the three-dimensional occupied grid map, obtaining the material category of each voxel, constructing a semantic voxel map, and recording the semantic voxel map as a prospective environment model.
4. A method for collaborative work on a machine dog formation in a disaster environment according to claim 3, wherein the method for obtaining semantic attribution comprises: fusing coordinates of the voxels, laser radar reflection intensity and color information, extracting geometric features, physical features and visual features, and forming feature vectors; The feature vectors are input into a pre-trained random forest classifier, and the material class of each voxel is output.
5. The method for collaborative operation of a machine dog formation in a disaster environment according to claim 1, wherein calculating a link quality score for future intent information of a lead machine dog at a current location for the relay machine dog comprises: Performing ray tracing from a current location of the relay robot dog to a plurality of future path points in the pilot robot dog future intent information in the prospective environment model; for any obstacle voxel through which the ray passes, calculating linear path loss according to the material type and the penetration thickness of the obstacle voxel; And calculating the link quality of the relay robot dog current position to the future path point based on the linear path loss, and fusing the link quality of all the future path points to obtain the link quality score.
6. The method for collaborative operation of a machine dog formation in a disaster environment according to claim 5, wherein the equation for calculating the linear path loss is: ; ; Wherein, the Representing the first pass of the ray The linear path loss of the individual obstacles, And Respectively represent materials Fixed loss and loss per meter, N represents the relay dog position To future point of path The number of obstacles on the ray in between, Representing the first pass of the ray Penetration thickness of individual obstacles; Indicating relay dog position For future path points Is used for the linear link quality of (a), Is the power of the transmission in a linear fashion, Is based on linear path loss.
7. A method of collaborative work on a machine dog formation in a disaster environment according to claim 1, wherein the computing a gradient of a potential field of the link quality score to derive a communication gradient impulse includes: estimating the gradient of the link quality score potential field by calling the calculation of the link quality score for a plurality of times at the current position of the relay robot dog and the virtual sampling points in the close vicinity of the current position of the relay robot dog by adopting a finite difference method ; The calculation formula of the communication gradient pushing force is as follows: ; Wherein, the For the driving force of the communication gradient, Is a gradient gain coefficient.
8. The method for collaborative operation of a machine dog formation in a disaster environment according to claim 1, wherein the switching of operation modes according to an operation state of a pilot machine dog comprises: if the pilot robot dog is in a mobile exploration state, switching the relay robot dog to a dynamic following mode; If the pilot robot dog is in a resident operation state, the relay robot dog is switched to a static anchoring mode.
9. The method for collaborative operation of a machine dog formation in a disaster environment according to claim 8, wherein the final navigational total force is calculated as: ; Wherein, the To dynamically follow the resultant force, also referred to as the final navigation resultant force, To pilot the communication attraction generated by the current position of the dog, As a result of the repulsive force of the obstacle, For the base station repulsive force, 、、 For the corresponding gain factor(s), For the driving force of the communication gradient, Is a gradient gain coefficient.
10. A machine dog formation co-operating system in a disaster environment, comprising a processor and a memory, the memory storing computer program instructions which, when executed by the processor, implement a machine dog formation co-operating method in a disaster environment according to any one of claims 1-9.

Description

Machine dog formation collaborative operation method and system in disaster environment Technical Field The invention relates to the field of robot navigation, in particular to a robot dog formation collaborative operation method and system in a disaster environment. Background In disaster environments such as earthquake, fire disaster or building collapse, collaborative search and rescue are performed by utilizing multiple robots, particularly high-mobility robot dog formation, so that the method has become a key technology for guaranteeing the safety of rescue workers and improving the search and rescue efficiency. In a typical pilot-relay operation mode, the pilot machine dog goes deep into the disaster site to perform autonomous exploration, and the relay machine dog at the rear is responsible for following and constructing a dynamic communication link so as to ensure that video and sensor data of the pilot dog can be returned to the rear command center in real time. To achieve this goal, the navigation algorithm of the relay robot dog is critical. Currently, navigation strategies based on the artificial potential field method (ARTIFICIAL POTENTIAL FIELD, APF) are commonly employed in the industry. The method sets a target point, such as a pilot dog, as an attractive force source, sets an obstacle as a repulsive force source, and the relay robot dog moves along the resultant force direction of the potential field. However, conventional APF methods have serious drawbacks in a complex, unstructured environment, such as a disaster site. APF is a reactive algorithm that makes decisions based solely on the attraction and repulsion forces at the current moment, lacking a prospective prediction of future environmental changes. This strategy is very easy for relay machine dogs to sink into locally optimal solution traps. In disaster sites, the trap is embodied in that the relay robot dog may be navigated to a physically safe and well signaled location, but the location is topologically a dead communication corner. When the pilot dog continues to go deep, detours around a load-bearing wall or a massive obstacle, the communication link between the relay dog and the pilot dog is interrupted instantaneously. At this time, if the relay robot dog wants to reestablish the connection, the relay robot dog needs to go backward and reselect points greatly, which wastes valuable rescue time seriously and may even cause the pilot dog to lose connection. Therefore, how to solve the problem of local optimal trap of the relay robot dog in the navigation process, so that the relay robot dog can not only follow the pilot dog, but also avoid future communication dead angles in a prospective manner, and realize stable and continuous communication relay is a technical problem to be solved in the field. Disclosure of Invention The invention provides a robot dog formation collaborative operation method and a robot dog formation collaborative operation system in a disaster environment, which aim to solve the technical problems that the traditional artificial potential field method is easy to fall into communication dead angles and local optimal solutions. In a first aspect, the invention provides a cooperative work method for machine dog formation in a disaster environment, which adopts the following technical scheme: A machine dog formation collaborative operation method under a disaster environment comprises the following steps: acquiring future intention information broadcast by a pilot robot dog and a local environment model perceived by the relay robot dog, and fusing to construct a prospective environment model; Calculating a link quality score of the relay machine dog on future intention information of the pilot machine dog at the current position based on the prospective environment model, and constructing a potential field of the link quality score; Calculating the gradient of the potential field of the link quality score to obtain a communication gradient impetus; Switching an operation mode according to the operation state of the pilot robot dog, under the operation mode, fusing the communication gradient pushing force, the communication attraction based on the current position of the pilot robot dog and the obstacle repulsive force, calculating to obtain a final navigation resultant force, and controlling the relay robot dog to execute cooperative operation. According to the invention, by introducing a communication gradient pushing force based on the future intention of the pilot dog, the traditional gravitation which only depends on the current position is replaced, so that the relay dog obtains the capability of predicting the future communication quality. The relay dog can proactively avoid upcoming communication dead angles, thoroughly solve the problem that the traditional APF method is easy to sink into a local optimal solution trap, and remarkably improve communication continuity and navigation robustness of disast