CN-120949764-B - Autonomous navigation and obstacle avoidance system of fire-fighting robot

CN120949764BCN 120949764 BCN120949764 BCN 120949764BCN-120949764-B

Abstract

The invention discloses an autonomous navigation and obstacle avoidance system of a fire-fighting robot, and relates to the technical field of navigation and obstacle avoidance. According to the method, the accuracy and feasibility of path planning are effectively ensured by constructing the three-dimensional state unit set and eliminating the non-accessible area; the system can realize local reconstruction and continuity recovery of the path through the graph structure updating and path correcting module, and avoid delay and resource waste caused by global re-planning in the traditional system. The invention also merges the space overlap ratio calculation and jump detection method, thereby effectively improving the reliability of key frame node identification and the stability of task switching.

Inventors

DING JIE
LI RUI

Assignees

江苏优异家科技有限公司

Dates

Publication Date: 20260505
Application Date: 20250715

Claims (8)

1. The autonomous navigation and obstacle avoidance system of the fire-fighting robot is characterized by comprising the following components: The state diagram construction and obstacle perception coding module is used for dividing a working area and constructing a state transition probability diagram; the key frame extraction module is used for calculating an optimal state path in the state transition probability diagram based on the transition probability cost, extracting a task type switching point as a key frame state node and realizing path task segmentation control; The map structure updating module is used for dynamically adjusting state nodes and edge sets in the state transition probability map after obstacle detection or environmental change, and reconstructing a key frame sequence adaptation new structure; The path correction module is used for carrying out path backtracking and local reconstruction based on the pointer array after the structure of the state transition probability graph is changed or the path is interrupted, updating the key frame sequence for the second time and recovering the continuity of the path; The construction of the state transition probability map comprises the following steps: Constructing a state unit set of a three-dimensional space together according to a building structure diagram of a fire scene, actual measurement laser point cloud data and a preset grid size, and eliminating an inaccessible area and a shielding body element for each space unit; screening state units meeting carrier constraint by utilizing a pre-constructed task-carrier adaptation table based on a task type set and a carrier set to generate a state node set; In the state node set, if any pair of state nodes meet that Euclidean distance between the state nodes is smaller than a jump threshold value and a structural communication relation exists between the state units, adding a transition edge to the transition edge set, and constructing a state transition probability graph.
2. The autonomous navigation and obstacle avoidance system of the fire robot of claim 1 wherein the keyframe extraction module comprises: based on the state node set, classifying and marking each state node according to the task type set and the carrier set, and screening out state nodes representing the start point and the end point of the task; In the state transition probability diagram, taking state nodes of a screening starting point and a screening end point as two ends, and executing path planning operation by combining transition probability cost of a transition edge to obtain an ordered state node access sequence; In the state node access sequence, identifying state nodes corresponding to task type changes, screening state nodes with task segment dividing characteristics, taking the state nodes as candidate task key frame state node sets, and dividing continuous task segments; In the candidate task key frame state node set, for each candidate task segment boundary node, a transition edge pair set formed by transition edges between a previous node and a next node is identified, for each transition edge pair, the space coincidence degree of voxel sets corresponding to two transition edges is calculated, the space coincidence degree is the ratio of the number of coincident voxels to the total number of voxel union sets, state nodes with the space coincidence degree score exceeding a stability threshold are screened, and the state nodes are collected into a final task key frame state node sequence.
3. The autonomous navigational and obstacle avoidance system of the fire robot of claim 2 wherein said path planning operation comprises: In the state transition probability diagram, identifying a starting point state node and an end point state node of screening, and recording the starting point state node and the end point state node as boundary nodes of path planning; Constructing a graph structure which takes a state node as a vertex and transition probability cost as an edge weight for all the communicable paths between boundary nodes, and calculating path cost by adopting cost accumulation; And for each passable path, summing the negative logarithmic probability costs based on the transition edge, and recording a state node access sequence corresponding to the minimum total cost path.
4. The autonomous navigation and obstacle avoidance system of the fire robot of claim 3 wherein said screening status nodes featuring task segment partitioning comprises: based on the task type of each state node in the state node access sequence, identifying the position where the task type changes, and marking the corresponding state node as a candidate task segment boundary node; For each candidate task segment boundary node, evaluating task switching strength by combining jump amplitude of front and rear state nodes on transition probability cost, and screening out state nodes with variation amplitude lower than a variation threshold; collecting the reserved boundary nodes of the candidate task segments into a candidate task key frame state node set, and dividing continuous task segments between adjacent state nodes according to the positions; For candidate task key frame state nodes, identifying adjacent transfer edge sets formed by the front and rear state nodes, and extracting transfer edge pairs with cross connection relations; For each transfer edge pair, calculating the ratio of the number of overlapped voxels to the total number of adjacent voxels in the space coordinate system as a space coincidence score of the task switching point; And comparing the space coincidence degree score with a stability threshold, screening state nodes with the score higher than the stability threshold as task key frame state nodes with high confidence degree, and forming a key frame state node sequence.
5. The autonomous navigation and obstacle avoidance system of the fire robot of claim 1 wherein said dynamically adjusting the set of state nodes and edges in the state transition probability map comprises: After the environmental change information or the obstacle detection result is obtained, positioning an affected area in the state unit set, and marking the affected state node set; performing validity check on the affected state node set, if the state unit where the state node is located is blocked or is not enabled, removing the state unit from the affected state node set, and synchronously removing the associated transfer edges; Regenerating a state node based on connectivity of the state unit for the area where the accessible path newly appears due to structural shielding, and supplementing a transfer edge according to a space topological relation with the peripheral node; and re-executing path planning operation on the adjusted state transition probability diagram, adding and deleting the task key frame state node sequence, and adapting the state transition logic after structure updating.
6. The autonomous navigation and obstacle avoidance system of the fire robot of claim 1, wherein the performing path backtracking and local reconstruction based on the array of pointers, the secondarily updating the sequence of keyframes comprises: When the path structure in the state transition probability diagram is interrupted or changed, judging the range of the subsequence affected in the original state node access sequence, and identifying the start-stop state node; Based on the forward pointer array, tracing back from the affected subsequence termination node forward, and reconstructing the shortest path subsequence to the starting point; In the backtracking path, local connection and task segment recombination are executed by combining the unaffected task key frame state nodes in the original state node access sequence, and the continuity of the path structure is recovered; And outputting the state node access sequence after backtracking repair, and carrying out secondary updating on the task key frame state node sequence for supporting subsequent task segment navigation control operation.
7. A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the computer program is characterized in that the processor realizes the steps of the autonomous navigation and obstacle avoidance system of the fire-fighting robot according to any one of claims 1-6 when executing the computer program.
8. A computer readable storage medium having a computer program stored thereon, wherein the computer program when executed by a processor performs the steps of the autonomous navigation and obstacle avoidance system of a fire fighting robot according to any one of claims 1 to 6.

Description

Autonomous navigation and obstacle avoidance system of fire-fighting robot Technical Field The invention relates to the technical field of navigation and obstacle avoidance, in particular to an autonomous navigation and obstacle avoidance system of a fire-fighting robot. Background In the context of modern urban rapid propulsion, fire-fighting tasks face multiple challenges of high environmental complexity, urgent emergency response time, and high human risk. In order to effectively improve the operation efficiency and the safety in fire rescue, an intelligent fire-fighting robot is gradually paid attention as an important means for replacing manual operation. Especially in high-risk environments such as poisonous, anoxic or high-temperature environments, the use of robots with strong autonomous navigation and obstacle avoidance capabilities to assist or replace firefighters to carry out tasks such as investigation, fire extinguishment, search and rescue and the like has become a key direction for the development of intelligent fire-fighting systems. Currently, mainstream researches focus on improving navigation and obstacle avoidance capabilities of a fire-fighting robot through means of path planning, environment sensing, obstacle recognition and the like. The common path construction means comprise grid map, global shortest path search, three-dimensional environment modeling based on sensor construction and the like, and part of schemes have realized basic navigation obstacle avoidance functions, but the common path construction means still have defects in the aspects of dynamic environment adaptability, task continuity control and recovery capability after path interruption. CN102359784B discloses an autonomous navigation obstacle avoidance system and method for an indoor mobile robot, by constructing a wireless sensor network, dividing an indoor environment into a plurality of grid areas, combining an expansion algorithm and Dijkstra algorithm to realize global path planning, and the robot cooperates with a sensing node to realize positioning and path execution through a communication module. Although the scheme has certain advantages in positioning accuracy and path planning, the scheme relies on a pre-arranged static sensor network, the system deployment cost is high, and in the scene of sudden change of an environment structure or sudden movement of an obstacle, a path update mechanism is lagged, a targeted path restoration strategy is lacked, and the requirements of path reconstruction and key frame segmentation control in the dynamic environment of a fire scene cannot be met. CN118623894B discloses an autonomous navigation obstacle avoidance method under a dynamic scene of a robot, which adopts an image recognition model to process an environmental image, selects different obstacle avoidance algorithms after distinguishing dynamic and static obstacles, and dynamically adjusts a navigation path. The method has good adaptability in the aspect of processing different types of obstacles, and the obstacle avoidance capability of the robot in a complex scene is improved. However, the scheme focuses on the obstacle avoidance strategy, lacks modeling capability of the overall task path structure, is still single-segment continuous control in path control, does not relate to segment management and state transition logic control of the path task, is difficult to realize multi-target state switching in the fire-fighting task and automatic recovery requirements after the path is interrupted, and has to be improved in overall path robustness and critical path segment controllability. Disclosure of Invention The invention is provided in view of the problems existing in the prior fire-fighting robot navigation technology. Therefore, the problem to be solved by the invention is how to realize a flexible updating mechanism in an emergency environment with frequent dynamic changes, and the risk of path interruption caused by sudden obstacle changes is solved. In order to solve the technical problems, the invention provides the following technical scheme: The invention provides an autonomous navigation and obstacle avoidance system of a fire-fighting robot, which comprises a state diagram construction and obstacle perception coding module, a key frame extraction module, a diagram structure updating module and a path correction module, wherein the state diagram construction and obstacle perception coding module is used for dividing a working area and constructing a state transition probability diagram, the key frame extraction module is used for calculating an optimal state path in the state transition probability diagram based on transition probability cost, extracting a task type switching point as a key frame state node to realize path task segmentation control, the diagram structure updating module is used for dynamically adjusting state nodes and edge sets in the state transition probability diagram after obstac