Search

CN-121977599-A - Four-foot robot system for autonomous forest inspection

CN121977599ACN 121977599 ACN121977599 ACN 121977599ACN-121977599-A

Abstract

The invention relates to the technical field of intelligent sensing and control equipment manufacturing, and discloses a quadruped robot system for autonomous inspection of forests, which comprises a multidimensional sensing module, a terrain impedance resolving module, an energy cost mapping module and a dynamic energy gauge module, wherein the multidimensional sensing module is used for synchronously acquiring environment geometric point cloud, visual images, body gestures, joint driving motor electrical parameters and angular velocity data, the terrain impedance resolving module is used for stripping motor current gravity load components to acquire dynamic resistance moment and combining gradient and vibration frequency to generate a terrain impedance coefficient, the energy cost mapping module is used for constructing a metabolic cost map containing nonlinear energy cost based on the terrain impedance coefficient and the gradient to distinguish rigid barriers from flexible navigable vegetation, and the dynamic energy gauge module is used for planning a minimum energy consumption path and triggering return based on dynamic return energy consumption prediction.

Inventors

  • LIU TONGXIN

Assignees

  • 陕西绿动生态科技有限公司

Dates

Publication Date
20260505
Application Date
20260122

Claims (10)

  1. 1. A four-legged robot system for forest autonomous inspection, characterized in that the system includes a multidimensional sensing module, a terrain impedance resolving module, an energy cost mapping module and a dynamic energy planning module: the multidimensional sensing module is used for synchronously collecting environment geometric point cloud data, visual image data, body inertial attitude data, phase current data of each joint driving motor and joint angular velocity data; The terrain impedance calculation module is used for stripping the gravity load component from the phase current data by utilizing a Kalman filtering algorithm to obtain a dynamic resistance moment, carrying out weighted fusion on the dynamic resistance moment and the ground gradient and the vibration frequency represented by the inertial attitude data of the machine body, and outputting a terrain impedance coefficient corresponding to the current position coordinate in real time; The energy cost mapping module is used for constructing a navigation grid map and executing an energy cost assignment rule according to a terrain impedance coefficient, namely detecting an area with echo based on environmental geometric point cloud data and identified as vegetation based on visual image data, wherein the terrain impedance coefficient is lower than a preset rigidity threshold value and is set as a low damping energy cost allowing passage, and generating a metabolism cost map by non-linearly lifting the energy cost according to the product of a resistance coefficient and a slope sine value according to the area with the terrain impedance coefficient higher than the preset softness threshold value or the area with the ground slope higher than a preset angle; the dynamic energy planning module is used for planning a travel path with the minimum accumulated energy cost on the metabolism cost map by adopting a heuristic search algorithm, calculating the expected return stroke energy consumption returned to the replenishing base station along the travel path in real time, and triggering a return stroke instruction only when the difference value between the current residual energy and the expected return stroke energy consumption is smaller than a safety redundancy threshold value.
  2. 2. The four-legged robot system for autonomous forest inspection according to claim 1, wherein the energy cost mapping module generates a metabolic cost map for a single step pass of energy costs for each grid element The following quantization calculation logic is executed: , wherein, And For the preset weight coefficient, the weight coefficient is set, The reference power for the robot to walk on a flat hard road surface, For the expected time to pass through the grid cell, Is dynamic resistance moment corresponding to the topographic impedance coefficient, In order to achieve the angular velocity of the joint, The gravitational potential energy influencing factor is used to determine, Is the quality of the whole machine, The acceleration of the gravity is that, The dynamic energy planning module is used for inputting single-step traffic energy cost as weight into a cost function of an A-type algorithm so as to search a topological path with the lowest global total power consumption.
  3. 3. The four-foot robot system for autonomous forest inspection according to claim 1, wherein the terrain impedance resolving module comprises a gravity compensation subunit, a disturbance observation subunit, a characteristic fusion subunit and a feature fusion subunit, wherein the gravity compensation subunit is used for calculating static moment components of each joint generated by self gravity in real time based on fuselage inertia attitude data and a robot connecting rod dynamics model, the disturbance observation subunit is used for mapping phase current data into total electromagnetic moment and subtracting the static moment components and the inertia moment components to obtain dynamic resistance moment representing surface medium interaction, the feature fusion subunit is used for carrying out time domain alignment on the amplitude change rate of the dynamic resistance moment and high-frequency vibration components in fuselage inertia attitude data, determining a crushed stone pavement when low resistance moment and high vibration frequency characteristics are detected, determining a soft mud biogas pavement when high resistance moment and low vibration frequency characteristics are detected, and generating a terrain impedance coefficient according to the static moment.
  4. 4. The four-foot robot system for autonomous forest inspection according to claim 1, wherein the assignment of the energy cost mapping module to the low damping energy cost follows the following logic, when the geometrical obstacle exists in front of the environmental geometrical point cloud data representation and the visual image data representation obstacle texture features conform to shrubs or high grasses, the system tentatively presumes the area as a potential passing area, and if the instantaneous contact resistance fed back by the terrain impedance resolving module is smaller than a preset flexible crossing threshold value when the four-foot robot contacts the area, the passing state of the area in the navigation grid map is corrected from unreachable to reachable, and the energy cost of the area is set as a preset multiple of the reference land leveling cost, and the preset multiple is in positive correlation with vegetation density.
  5. 5. The four-legged robot system for autonomous forest inspection according to claim 1, wherein the dynamic energy planning module comprises a dynamic backhaul prediction subunit, the subunit is used for reversely searching a minimum energy consumption path from a current coordinate point to a nearest replenishment base station based on a currently generated metabolic cost map at a preset frequency during the process of performing an inspection task by the robot, integrating and summing energy costs of all grid units on the minimum energy consumption path to obtain expected backhaul energy consumption, monitoring a real-time state of charge output by the battery management system, and interrupting the task and controlling the robot to return along the minimum energy consumption path when a difference value of residual energy corresponding to the real-time state of charge minus the expected backhaul energy consumption is lower than a safety redundancy threshold.
  6. 6. The four-foot robot system for autonomous forest inspection according to claim 1, wherein the multidimensional sensing module further comprises a visual topology positioning subunit for taking over the positioning function when satellite positioning signals are lost, wherein the visual topology positioning subunit is used for extracting trunk textures, rock contours and forest window illumination distribution characteristics in the surrounding environment as visual road signs to construct a topology node map, and when the multidimensional sensing module detects that satellite signal precision factors are lower than preset standards, the visual topology positioning subunit is used for carrying out feature matching on currently acquired visual characteristics and key frames in a historical database, and calculating a relative pose transformation matrix to correct odometer drift until high-precision satellite signals are reacquired or a replenishment base station is reached.
  7. 7. The four-legged robot system for autonomous forest inspection according to claim 1, further comprising an electric power data base station deployed in a forest area, wherein the electric power data base station comprises a photovoltaic power generation module, a storage battery pack, a wireless charging interface and a network relay module, the four-legged robot further comprises a power receiving module which is adaptive to the wireless charging interface, and when the four-legged robot executes a return instruction and reaches a docking range of the electric power data base station, the four-legged robot performs non-contact energy supplementing by aligning the wireless charging interface through visual servo, and simultaneously uploads inspection data through the network relay module.
  8. 8. The four-legged robot system for autonomous forest inspection according to claim 3, wherein the feature fusion subunit further comprises slip detection logic for determining that the robot is currently in a high slip rate state when it is detected that the phase current data is rising sharply and the acceleration integral displacement in the inertial attitude data of the fuselage is not changing in a corresponding proportion, and the terrain impedance calculation module is responsive to the high slip rate state for adjusting the terrain impedance coefficient of the current position to infinity, marking the corresponding region as a high-risk energy consumption trap region in the metabolic cost map, and prohibiting the subsequent path planning to traverse the region.
  9. 9. A quadruped robot system for autonomous forest inspection according to claim 1, characterized in that the dynamic energy planning module, when planning the path, processes for the terrain gradient follow the anisotropic cost rule that for the same grid area the energy cost in the gradient up direction is set to a preset multiple of the energy cost in the gradient down direction, and when planning the path involves traversing the gradient, the dynamic energy planning module introduces a lateral stability penalty factor that increases non-linearly with the tangent of the gradient angle.
  10. 10. The four-foot robot system for autonomous forest inspection according to claim 1, further comprising a data management platform, wherein the data management platform comprises a task scheduling module and a data analysis module, the task scheduling module is used for dividing a plurality of inspection subareas according to geographical information of a forest region and updating a global energy consumption topological network according to a metabolic cost map returned by each four-foot robot, the data analysis module is used for receiving data acquired by the four-foot robot, identifying animal and plant traces by using a deep learning algorithm, and correlating an identification result with a topographic impedance coefficient of an acquisition position to generate a comprehensive forestry electronic map comprising biodistribution and topographic accessibility features.

Description

Four-foot robot system for autonomous forest inspection Technical Field The invention relates to a four-foot robot system for autonomous forest inspection, and belongs to the technical field of intelligent sensing and control equipment manufacturing. Background The current large-area unstructured forest environment automatic data acquisition and inspection are industry development trends, traditional forestry investigation relies on manual hiking inspection or fixed-point arrangement of infrared cameras, manual inspection is high in labor intensity and low in efficiency, personal safety risks exist in the face of complex mountain terrains or extreme climates, limited data acquisition of the coverage range of the infrared cameras is lagged, manual regular maintenance is required, and real-time global perception requirements of forest dynamic changes are difficult to meet. A four-legged robot foot type mobile platform gradually introduces field inspection, an existing mainstream scheme is used along a general mobile robot navigation control framework, the aim is to improve the mechanical reliability and modularized maintenance efficiency of a robot body, the requirements of field operation are met through hardware structure optimization, for example, chinese patent application publication No. CN120697872A discloses a four-legged robot, a hoop type connecting component is adopted to achieve rapid disassembly and assembly of a joint motor and a limb mechanism, the problem of complicated assembly is solved by utilizing precise matching of a positioning groove and a bulge, the equipment assembly efficiency and the maintenance convenience are improved, the four-legged robot foot type mobile platform gradually introduces field inspection, the existing mainstream scheme is used along the general mobile robot navigation control framework, a laser radar is used for combining a global navigation satellite system to sense an environment and planning path, a geometric grid map is built through core logic, objects higher than the ground are regarded as rigid barriers, a planning route is based on the principle of the geometric distance, the state of cruising is triggered according to battery voltage or a preset percentage threshold value, and the state of cruising is triggered, and the adaptability defect exists in a forest unstructured high-impedance environment is achieved based on geometric information and a static threshold control strategy. Therefore, how to break through the limitation of single geometric vision perception, improve the physical property ability of the robot to identify the environment medium, and construct a path planning and dynamic energy management system based on a physical acting mechanism, so that the invention is the technical problem to be solved. Disclosure of Invention In order to solve the problems in the background technology, the technical scheme of the invention is as follows, a four-foot robot system for autonomous forest inspection, which comprises a multidimensional sensing module, a terrain impedance resolving module, an energy cost mapping module and a dynamic energy planning module: the multidimensional sensing module is used for synchronously collecting environment geometric point cloud data, visual image data, body inertial attitude data, phase current data of each joint driving motor and joint angular velocity data; The terrain impedance calculation module is used for stripping the gravity load component from the phase current data by utilizing a Kalman filtering algorithm to obtain a dynamic resistance moment, carrying out weighted fusion on the dynamic resistance moment and the ground gradient and the vibration frequency represented by the inertial attitude data of the machine body, and outputting a terrain impedance coefficient corresponding to the current position coordinate in real time; The energy cost mapping module is used for constructing a navigation grid map and executing an energy cost assignment rule according to a terrain impedance coefficient, wherein echo is detected based on environmental geometric point cloud data, but vegetation is identified based on visual image data, low damping energy cost allowing passage is given to a region with the terrain impedance coefficient lower than a preset rigidity threshold, and the energy cost is non-linearly promoted according to the product of a resistance coefficient and a slope sine value for a region with the terrain impedance coefficient higher than a preset softness threshold or a region with the ground slope higher than a preset angle, so that a metabolism cost map is generated; the dynamic energy planning module is used for planning a travel path with the minimum accumulated energy cost on the metabolism cost map by adopting a heuristic search algorithm, calculating the expected return stroke energy consumption returned to the replenishing base station along the travel path in real time, and triggering a return str