Intelligent management method and management system of natural pasture using multi-machine collaboration of UAVs

An intelligent management method of a natural pasture using multi-machine collaboration of UAVs, comprising:

• obtaining pasture environment information through a UAV group;
• automatically planning, by the UAV group, a global grazing path according to the pasture environment information and a user-defined grazing time, a starting point and an end point obtained, and performing automatic driving and grazing;
• during the process of automatic driving and grazing based on the global grazing path, collecting and analyzing herd activity images, and planning local grazing paths to assist in the automatic driving and grazing;
• during the process of automatic driving and grazing based on the global grazing path, collecting and analyzing vegetation coverage image data of a grazing area to assist in the automatic driving and grazing;
• wherein, automatically planning, by the UAV group, a global grazing path according to the pasture environment information and a user-defined grazing time, a starting point and an end point obtained, and performing automatic driving and grazing comprises:
• obtaining user-defined starting point and end point areas of grazing and a grazing time;
• according to the pasture environment information and the starting point and end point areas, calling a global path planner based on an A* algorithm to plan the global grazing path, comprising:
• selecting a Manhattan distance between two points on the path as an estimate of a distance between the two points;
• obtaining terrain elevation data through sensors mounted on UAVs, establishing a digital elevation model, and adjusting a cost function of the path according to the digital elevation model:
• an actual cost function formula from a starting node to the node n is as follows:
• G(n)=G(parent)+Cost(n,parent);
• wherein:
• a parent node parent is a previous point to the node n, G(parent) is an actual cost from the starting point to the parent node, and Cost(n, parent) is an additional cost from the parent node to the node n;
• a heuristic estimated cost function H considers the path cost from the node n to a goal node goal, and the equation is as follows:
• H(n)=k1×Distance(n,parent)+k2×|Altitude(n)−Altitude(goal)|;
• the path cost function is:
• F(n)=G(n)+H(n)
• wherein:
• G(n) and H(n) are actual cost and heuristic estimated cost functions;
• the calculated path cost function will be used to select a next node to be expanded and generate the global grazing path;
• a calculation formula of the additional cost Cost(n, parent) from the parent node to the node n is as follows:
• Cost(n,parent)=k1×Distance(n,parent)+k2×|Altitude(n)−Altitude(parent)|+k3×Speed_Adjusted_Cost(n)+k4×Obstacle_Cost(n);
• wherein:
• Distance(n, parent) is the Manhattan distance between the two points;
• Altitude(n) is the elevation of node n;
• Speed_Adjusted_Cost(n) is the cost related to a flight speed of the UAV, comprising the flight speed of the UAV and a movement speed of a herd;
• Obstacle_Cost(n) is the cost related to obstacles around the node n, which is positively related to a density or distances of the obstacles around the path;
• and k1, k2, k3 and k4 are weights;
• wherein obtaining pasture environment information through a UAV group comprises:
• obtaining pasture environment image data through binocular depth cameras mounted on the UAVs;
• preprocessing the pasture environment image data and performing distortion correction and stereo correction based on a stereo matching method to obtain three-dimensional dense point cloud data;
• based on the three-dimensional dense point cloud data, constructing an octree map, dynamically adjusting accuracy and resolution, screening out the effective three-dimensional dense point cloud data, and constructing a pasture area environment map;
• and
• dividing the pasture area environment map into a plurality of areas that can be rotated for grazing, marking obstacles that affect normal flight of the UAV group in each area, and storing a vegetation coverage situation of each area in a cloud server to obtain the pasture environment information;
• wherein storing a vegetation coverage situation of each area in a cloud server comprises:
• obtaining the pasture vegetation image data through the UAV group, performing quality assessment and screening, smoothing and noise reduction, and image cropping processing on the pasture vegetation image data, calculating an area vegetation coverage degree according to the processed pasture vegetation image data and storing in the cloud server;
• wherein during the process of automatic driving and grazing based on the global grazing path, collecting and analyzing herd activity images, and planning local grazing paths to assist in the automatic driving and grazing comprises:
• during the process of automatic driving and grazing based on the global grazing path, collecting and analyzing the herd activity images to obtain a position and a movement status of a herd during an actual grazing process, according to the position and movement status of the herd during the actual grazing process, calling a local path planner based on an improved artificial potential field method to plan the local paths for UAV grazing, and driving the herd to a goal grazing area through stereo players mounted on the UAVs to achieve assistance in the automatic driving and grazing;
• wherein calling a local path planner based on an improved artificial potential field method to plan the local paths for UAV grazing comprises:
• using a Gaussian function to represent a repulsion force field, which is used to point to a direction away from the herd;
• the equation is as follows:
• where:
• Frepulsion is a vector of the repulsion force field;
• N is the number of herd individuals;
• k is a coefficient for adjusting repulsion force strength;
• α is a parameter for adjusting repulsion force attenuation;
• di is a distance from the UAV to the ith herd individual;
• (xi, yi) is the position of the ith herd individual;
• and (x, y) is the current position of the UAV;
• wherein calling a local path planner based on an improved artificial potential field method to plan the local paths for UAV grazing comprises:
• using a Gaussian function to represent an attraction force field, which is used to point to the herd individual closest to the UAV;
• the equation is as follows:
• where:
• Fattraction is a vector of the attraction force field;
• N is the number of herd individuals;
• w is a coefficient for adjusting attraction force strength;
• β is a parameter for adjusting attraction force strength attenuation;
• di is a distance from the UAV to the ith herd individual;
• (xi, yi) is the position of the ith herd individual;
• and (x, y) is the current position of the UAV;
• the improved artificial potential field method is achieved by:
• combining a traditional artificial potential field method with the global path planning method, providing overall path guidance through global path planning, and using the artificial potential field method for local obstacle avoidance to overcome limitations of the artificial potential field method;
• dynamically adjusting a shape and intensity of a potential field according to the position and movement status of the herd, and designing a potential field strategies based on the behavioral characteristics of the herd to adapt to actual grazing scenarios;
• avoiding the UAV falling into local optimality and improving robustness of path planning by introducing random disturbances, and path re-planning;
• choosing sensors to sense changes in the herd and environment in real time to dynamically update path planning;
• and
• considering dynamics and constraints of the UAV to ensure feasibility and safety of path planning;
• the calling a local path planner based on an improved artificial potential field method to plan the local paths for UAV grazing, comprises:
• using the user-defined starting point and end point position coordinates, as well as the pasture area environment map and obstacle information obtained by the UAV inspection before grazing as inputs to the path planning algorithm;
• searching a global optimal path from the starting point to the end point based on an improved A* path planning algorithm and the pasture area environment map;
• updating the pasture area environment map in real time during an actual flight process of the UAV, and planning the local path under a current situation based on the improved artificial potential field method;
• and
• planning the local paths by the UAV.