A multi-dimensional cooperative enhancement-based MDC-AFI adaptive feedback fusion intelligent path planning method and system

Abstract

The application discloses a kind of MDC-AFI adaptive feedback fusion intelligent path planning method and system based on multidimensional synergistic enhancement, belong to robot path planning technical field, comprising: obtaining the initial state information of intelligent robot;Based on initial state information, calling dynamic adaptive global path planner searches global path, iteratively calculates and obtains global path node, generates global optimization path;Global optimization path is decomposed to obtain several local target points and is used as the initial reference input of local path planner, and the parameters of local path planner are adaptively adjusted through MDC-AFI framework, and iteratively generate dynamic optimization local path;When the degree of local path deviating from global optimization path exceeds the preset tolerance range, it is judged whether it needs to re-optimize global path, and the latest global path data is fed back to local path planner for feedback fusion until intelligent robot reaches target point.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of robot path planning, and particularly relates to a MDC-AFI adaptive feedback fusion intelligent path planning method and system based on multi-dimensional cooperative enhancement. BACKGROUND

[0002] In the field of modern robot autonomous navigation and path planning, improving path planning efficiency and reliability is the key for intelligent robots to achieve efficient operation in complex dynamic environments. In application scenarios such as logistics and warehousing, and unmanned driving, intelligent robots need to consider both global path planning and local path obstacle avoidance to cope with dynamic obstacle distribution and real-time environmental changes. Traditional path planning methods mainly include graph-based path search algorithms and sampling-based probabilistic path planning algorithms. The former, such as the A* algorithm, is widely used for its simplicity and efficiency, but it has poor real-time performance and flexibility in handling dynamic environments. The latter, such as the RRT algorithm, uses random sampling to find paths in high-dimensional space, which is suitable for solving complex problems, but has shortcomings in path smoothness and global optimality. The ability of feature fusion between global and local path planning modules is the key to improving path planning performance.

[0003] To improve the adaptability of path planning algorithms to complex environments, researchers have proposed dynamic path optimization models, cost map fusion algorithms, and reinforcement learning-driven path planning methods, which optimize paths by integrating environmental perception, path evaluation, and dynamic characteristics. However, existing technologies have significant shortcomings in feature fusion between global and local path planning modules. For example, the decomposition and reconstruction process of global and local paths lacks comprehensive consideration of real-time environmental information, resulting in weak adaptability to dynamic environments. Therefore, there is an urgent need for an intelligent path planning method that can efficiently and dynamically fuse global and local paths in complex dynamic environments while considering path smoothness and dynamic constraints. SUMMARY

[0004] The present application aims to solve the shortcomings of the prior art and provides the following solution:

[0005] A MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional cooperative enhancement, comprising the following steps:

[0006] S1. Obtain the initial state information of the intelligent robot, including the initial coordinates, attitude information, obstacle distribution layer information, and target point coordinates;

[0007] S2. Based on the initial state information, call the dynamic adaptive global path planner to search for a global path, iteratively calculate the global path nodes, and generate a globally optimized path;

[0008] S3. decompose the global optimization path to obtain a plurality of local target points and take them as initial reference inputs of a local path planner, adaptively adjust parameters of the local path planner through a multi-dimensional feasibility dynamic evaluator under the MDC-AFI framework and distribution information of dynamic obstacles in a local environment, iteratively generate a dynamically optimized local path according to dynamic constraint conditions of the intelligent robot and real-time environmental information;

[0009] S4. when the local path deviates from the global optimization path to a degree exceeding a preset tolerance range, the multi-dimensional feasibility dynamic evaluator judges whether the global path needs to be re-optimized, and feeds back the latest global path data to the local path planner for feedback fusion;

[0010] S5. repeat the S1 to the S4 until the intelligent robot reaches the target point and completes the target navigation task.

[0011] Preferably, in the S1, the method for obtaining the attitude information comprises:

[0012] The formula for calculating the quaternion by using the accelerometer and the magnetometer data in the IMU is:

[0013]

[0014] wherein θ represents a rotation angle around the unit vector u=[u X , u Y , u Z ], w represents a real part of the quaternion, x, y and z represent imaginary parts of the quaternion, and respectively represent components of rotation around the X, Y and Z axes;

[0015] Substitute the initial attitude into the formula of the quaternion to obtain the initial quaternion q0=[w0, x0, y0, z0]:

[0016] Calculate the angular increment Δq by using the angular velocity of the gyroscope, update the quaternion, and obtain the updated quaternion:

[0017]

[0018] wherein t0 represents an initial time, and Δt represents an update time step;

[0019] Solve the attitude angle by using the updated quaternion q'=[w', x', y', z'], and obtain the attitude information:

[0020]

[0021] Pitch=sin -1 (2(w'y'-x'z'))), |Pitch|≤90°,

[0022]

[0023] Where Roll represents the roll angle around the X-axis, Pitch represents the pitch angle around the Y-axis, and Yaw represents the yaw angle around the Z-axis;

[0024] The rotation matrix R is obtained by calculating the quaternion q = [w, x, y, z].

[0025]

[0026] The corresponding rotation matrix R is calculated based on the attitude information from each sensor. ε ;

[0027] Set the local coordinate system of each sensor Mapped to global coordinate system This provides a unified benchmark for subsequent algorithms, ensuring that data is correctly transferred and processed between different coordinate systems.

[0028]

[0029] Where ε represents the index of each sensor in the sensor group;

[0030] Reversible transformation, that is, converting global coordinates to local coordinates:

[0031]

[0032] in, This represents the local coordinates of the sensor with index ε.

[0033] Preferably, the method for optimizing the global path includes:

[0034] A dynamic adaptive cost heuristic function h′(n) is introduced, which combines multi-dimensional information such as obstacle distribution density, dynamic obstacle velocity direction and Euclidean distance to the target point to optimize path search efficiency;

[0035] A path smoothing algorithm is adopted to optimize the objective function and process the initial global path, reduce path polylines, and improve the smoothness and controllability of the path.

[0036] By using hierarchical grid modeling, higher weights are assigned to highly complex regions and lower weights to low-complexity regions, thereby improving the globality and efficiency of path search.

[0037] Preferably, the method for optimizing path search efficiency includes:

[0038] The computation time cost h time (n), which is the shortest time estimate from node n to target point G:

[0039]

[0040] Where ||ng|| represents the Euclidean distance from node n to target point G, and v represents the linear velocity of the intelligent robot;

[0041] Calculate the spatial distance cost h distance (n), which is the Euclidean distance from node n to target point G:

[0042]

[0043] Among them, (x n y n (x) represents the coordinates of node n. g y g () represents the coordinates of the target point G;

[0044] Calculate the dynamic environment cost h dynamic (n), which represents the impact of dynamic obstacles on path planning:

[0045]

[0046] Where k represents the number of dynamic obstacles, d i θ represents the distance from node n to the i-th dynamic obstacle. i This represents the angle between the direction from node n to target point G and the velocity direction of the dynamic obstacle;

[0047] Based on the time cost h time (n), the spatial distance cost h distance (n) and the dynamic environment cost h dynamic (n) Construct a dynamic adaptive cost heuristic function h′(n):

[0048] h'(n)=αh time (n)+βh distance (n)+γh dynamic (n),

[0049] Where α, β, and γ represent h respectively. time (n), h distance (n), h dynamin The weight parameters of (n) are given, and α+β+γ=1;

[0050] The dynamic adaptive cost heuristic function h′(n) can be expressed as the product of the weight vector and the cost vector:

[0051] h′(n)=W·H(n),

[0052] Where W = [α, β, γ] represents the weight vector, and H(n) = [htime (n), h distance (n), h dynamic (n)] T represents a cost vector;

[0053] defining a dynamic environment parameter vector and dynamically self-adapting the weight vector W according to the environment parameter vector E(t):

[0054] W = W0+ K·E(t),

[0055]

[0056] E(t) = [p obs , d obs , c env ],

[0057]

[0058] wherein W0represents an initial weight vector, K represents a gain matrix, p obs represents a dynamic obstacle density, d obs represents a nearest distance of a node to a dynamic obstacle, c env represents an environment openness, represents a dynamic density function, represents an area of an environment region , p = [x, y] T represents a two-dimensional coordinate of a dynamic obstacle, represents a two-dimensional real number space, i represents an index of a dynamic obstacle set, P node represents a node position coordinate to be evaluated, v i represents a linear velocity vector of a dynamic obstacle with index i at time t, P i (t) represents a coordinate of a dynamic obstacle with index i at time t, represents an obstacle set, ||·||2represents an Euclidean norm, p max represents a maximum obstacle density that can be reached in a unit area.

[0059] Preferably, the method of generating a dynamically optimized local path comprises:

[0060] decomposing the path into local target points with reference to a global path optimized by an improved dynamic self-adaptive cost heuristic function h'(n);

[0061] calling a local dynamic obstacle avoidance planner in a complex environment to obtain the local path according to the distribution environment of dynamic obstacles and the motion state of the intelligent robot, using a multi-dimensional constraint matrix potential field optimization algorithm MDC-MPFO to evaluate the influence of local path planning, and making dynamic obstacle avoidance adjustment in time according to the evaluation result.

[0062] While dynamically avoiding obstacles, the path planning is constrained by the dynamic characteristics of the intelligent robot, using the discrete point curvature κ. i Assess the smoothness of the path to ensure its smoothness and dynamic feasibility.

[0063] Preferably, the method for obtaining the local path includes:

[0064] Matrixed path point coordinate set P and obstacle coordinate set O:

[0065]

[0066] Where p represents the number of local path points to be evaluated, and m represents the number of obstacles;

[0067] Calculate path point P i =(x i y i ) to target point G = (x g y g The gravitational field strength F att，i :

[0068]

[0069] Where, d goal k represents the gravitational range of the target point. att This is the gravitational field gain coefficient;

[0070] Based on the gravitational field strength F att，i Establish the gravitational field matrix F att :

[0071]

[0072] Among them, F att，xn F represents the x-axis component of the gravitational field strength at each path point to be evaluated. att，yn This represents the y-axis component of the gravitational field strength at each path point to be evaluated.

[0073] Calculate path point P i =(x i y i To the nearest obstacle O j =(x oj y oj The repulsive field strength F rep，i :

[0074]

[0075] Where, d safe k represents the safe distance. repd represents the repulsive field gain coefficient. obs,i Represents path point P i =(x i y i ) to obstacle O j =(x oj y oj The closest distance;

[0076] Based on the gravitational field strength F rep，i Establish the repulsive field matrix F rep :

[0077]

[0078] Among them, F rep，xp F represents the x-axis component of the repulsive field intensity from each path point to be evaluated to the nearest obstacle. rep，yp This represents the y-axis component of the repulsive field intensity from each path point to be evaluated to the nearest obstacle;

[0079] According to the superposition theorem, based on the gravitational field matrix F att and the repulsive field matrix F rep Constructing the resultant force field F total :

[0080]

[0081] Based on the updated path point positions of the local path planning using the resultant force field, a dynamically optimized local path is generated:

[0082]

[0083] Among them, P i，new P represents the updated coordinates of the local path points. i，current This represents the coordinates of the local path point before the update. F represents the path update step size. total，i This represents the resultant force field intensity at each local path point.

[0084] Preferred methods for evaluating the impact of local path planning using the Multidimensional Constraint Matrix Potential Field Optimization (MDC-MPFO) algorithm include: path feasibility assessment and path cost assessment.

[0085] The methods for assessing the feasibility of the proposed route include:

[0086] like The local path planning is unreliable and needs to be replanned.

[0087] like The local path, based on feasibility assessment, drives the intelligent robot's motion module to reach the local target point;

[0088] The methods for evaluating the cost of the path include:

[0089] Take the resultant force field F at each path point total The sum of the modulus lengths is J force :

[0090]

[0091] Among them, ||F total，i ‖ represents the resultant force field F total The modulus is p, where p represents the number of local path points to be evaluated.

[0092] Calculate first-order smoothness:

[0093]

[0094] Among them, J smooth，1st P represents the first-order smoothness value. i Let P represent the i-th local path point. i-1 This represents the (i-1)th local path point;

[0095] Calculate second-order smoothness:

[0096]

[0097] Among them, J smooth，2nd P represents the second-order smoothness value. i+1 This represents the (i+1)th local path point;

[0098] Weighted fusion, calculating the cost of path planning quality assessment J evaluate :

[0099] J evaluate =J force +λ1·J smooth，1st +λ2·J smooth，2nd ,

[0100] Where λ1 represents the first-order smoothness penalty coefficient and λ2 represents the second-order smoothness penalty coefficient.

[0101] Preferred methods for constrained path planning that incorporate the dynamic characteristics of intelligent robots include:

[0102] According to the dynamic model of the intelligent robot, the local path needs to satisfy three constraints, including:

[0103] Joint vector form constraint of velocity and angular velocity:

[0104]

[0105] Where Ψ represents the joint vector of velocity and angular velocity, X limrepresents the joint velocity-angular velocity constraint, v represents the linear velocity of the intelligent robot, ω represents the angular velocity, v min represents the minimum linear velocity, v max represents the maximum linear velocity, ω min represents the minimum angular velocity, ω max represents the maximum angular velocity;

[0106] Linear-angular acceleration vector form constraint:

[0107]

[0108] wherein, represents the joint linear-angular acceleration vector, V lim represents the joint linear-angular acceleration constraint, a represents the linear acceleration, β represents the angular acceleration, a min represents the minimum linear acceleration, a max represents the maximum linear acceleration, β min represents the minimum angular acceleration, β max represents the maximum angular acceleration;

[0109] Turning radius vector form constraint:

[0110]

[0111] wherein, R(Ψ) represents a nonlinear function of the turning radius, R min represents the minimum turning radius of the robot model.

[0112] Preferably, the discrete point curvature κ i is used to evaluate the path smoothness, wherein the discrete point curvature κ i is calculated by the following method:

[0113] Three consecutive local path points and

[0114] The area S of the curvature triangle is calculated by the determinant method ki :

[0115]

[0116] The point pair distance between three points is calculated:

[0117]

[0118] wherein, d i-1,i represents the distance between and , d i,i+1 represents the distance between and , di-1,i+1 denotes the distance between ;

[0119] Take the maximum point distance between three points as the total chord length L:

[0120] L = max {d i，i+1 , d i-1，i+1}; i-1，i i，i+1 i-1，i+1 ;

[0121] Substitute to solve the discrete point curvature κ i :

[0122]

[0123] When the local path point is dense, the discrete point curvature κ i is simplified as:

[0124]

[0125] Preferably, the method of feedback fusion comprises:

[0126] Based on the dynamic adaptive cost heuristic function h'(n) optimization of the global path decomposition of the local target point, the initial position of the intelligent robot is updated after completing each local target point;

[0127] When the local path deviates from the global path by more than a set threshold, the multi-dimensional feasibility dynamic evaluation module under the MDC-AFI framework judges whether to trigger the global path planner to re-optimize the global path, and dynamically updates the data fed back by the global path planner to the local path planning module;

[0128] Set the path switching priority, when the local path is seriously affected by dynamic obstacles, take the global path point as the priority reference level point, ensure the global optimality and safety of path planning, and prevent the robot from falling into local optimization.

[0129] Preferably, the method of judging whether to trigger the global path planner to re-optimize the global path comprises:

[0130] Calculate the Euclidean distance and point projection distance between the local path point and the global path point;

[0131] The Euclidean distance δ euclidean is:

[0132]

[0133] Wherein, is the point coordinate of the local path, is the point coordinate of the global path;

[0134] The point projection distance δprojection is:

[0135]

[0136] wherein, and denote two end points of a certain section in the global path;

[0137] According to the complexity of the planning scene, combined with the Euclidean distance delta euclidean and the point projection distance delta euclidean , the comprehensive deviation data is obtained by weighted fusion according to the weight parameters sigma, psi:

[0138] delta fused-linear = sigma delta euclidean + psi delta projection

[0139]

[0140] Wherein, sigma + psi = 1, delta fused-linear denotes linear fusion deviation data, and delta fused-nonliner denotes nonlinear fusion deviation data, sigma denotes the fusion weight of the Euclidean distance, and psi denotes the fusion weight of the point projection distance.

[0141] When it is detected that the fusion deviation data delta fused-linear and delta fused-nonliner are greater than the deviation threshold delta max , the multi-dimensional feasibility dynamic evaluator under the MDC-AFI framework triggers the global path planning to re-optimize the global path, fits the feedback data of the local path planner to obtain a more reliable global path, and then takes the obtained global path as the input of the local path planner, and iteratively optimizes.

[0142] The application also provides an MDC-AFI adaptive feedback fusion intelligent path planning system based on multi-dimensional cooperative enhancement, the system applies the method described in any one of the above, and is characterized in that it comprises: an initial information acquisition module, a global path optimization module, a local path optimization module and a feedback fusion module.

[0143] The initial information acquisition module is used for acquiring the initial state information of the intelligent robot, and the initial state information comprises: initial coordinates, attitude information, obstacle distribution layer information and target point coordinates.

[0144] The global path optimization module searches the global path based on the initial state information, calls a dynamic adaptive global path planner, iteratively calculates the global path node, and generates a global optimization path.

[0145] The local path optimization module is used for decomposing the global optimization path to obtain a plurality of local target points and taking the local target points as initial reference inputs of a local path planner, adaptively adjusting parameters of the local path planner through a multi-dimensional feasibility dynamic evaluator under an MDC-AFI framework and distribution information of dynamic obstacles in a local environment, and iteratively generating a dynamically optimized local path according to dynamic constraint conditions of the intelligent robot and real-time environment information.

[0146] The feedback fusion module judges whether the global path needs to be re-optimized when the local path deviates from the global optimization path by a degree exceeding a preset tolerance range, and feeds back the latest global path data to the local path planner for feedback fusion until the intelligent robot reaches the target point and completes the target navigation task.

[0147] Compared with the prior art, the present application has the following beneficial effects:

[0148] (1) The present application uses multi-dimensional cooperative enhancement technology and an adaptive feedback mechanism to real-time adjust global and local modules of path planning, so as to effectively generate an efficient and safe path in a complex dynamic environment, especially in a dense obstacle area, and to realize efficient and high-performance path planning.

[0149] (2) The present application proposes a multi-dimensional feature fusion strategy to solve the problem that path dynamic adjustment in a dense obstacle area falls into local optimization, and to realize real-time sensing and processing of dynamic environment change parameters, comprehensive optimization of dynamic constraints of multi-target paths, and dynamic balance between global and local targets.

[0150] (3) The present application introduces a dynamic adaptive cost heuristic function to overcome the problem of insufficient path continuity and stability caused by the limitation of the cost function in a complex dynamic environment, realizes efficient feedback adjustment and reconstruction between global and local path planning, and corrects path deviation in a dynamic environment with high precision, so as to improve the smoothness and robustness of the path.

[0151] (4) The present application proposes a multi-dimensional constraint matrix potential field optimization algorithm to solve the problem that path planner fails to plan due to low obstacle recognition accuracy or path adjustment lag, and still effectively identifies and avoids obstacles in a dense obstacle area, and significantly enhances the obstacle avoidance performance of autonomous navigation. BRIEF DESCRIPTION OF DRAWINGS

[0152] In order to more clearly illustrate the technical solutions of the present application, the following briefly introduces the drawings needed in the embodiments. Obviously, the drawings described below only show some of the embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0153] Figure 1 The method flowchart of the embodiment of the present application;

[0154] Figure 2 The method flowchart of the embodiment of the present application;

[0155] Figure 3 The comparison chart of using quaternion or rotation matrix to solve the initial data collected from IMU and directly integrating the gyroscope data, wherein a is the change of roll angle with time in the conventional scene, b is the change of roll angle with time in the Gimbal Lock scene, c is the change of pitch angle with time in the conventional scene, d is the change of pitch angle with time in the Gimbal Lock scene, e is the change of yaw angle with time in the conventional scene, and f is the change of yaw angle with time in the Gimbal Lock scene;

[0156] Figure 4 The schematic diagram of evaluating and adjusting the local path planning by the multi-dimensional constraint matrix potential field optimization algorithm of the embodiment of the present application, wherein a is the overall effect of global path planning and optimization of the transportation task, b is the overall effect of dynamic avoidance of the dynamic obstacle on the left side of the warehouse during local path tracking, c is the overall effect of dynamic avoidance of the dynamic obstacle on the right side of the warehouse during local path tracking, and d is the overall effect of global optimization path and dynamic obstacle avoidance during local path tracking;

[0157] Figure 5 The comparison chart of introducing Kalman filtering algorithm to calculate the path turning change and other algorithms (such as weighted average method and simple average method);

[0158] Figure 6 The comparison chart of using discrete point curvature to evaluate the path smoothness and the smoothness analysis method based on second derivative integral of the embodiment of the present application, wherein a is the comparison chart of the effects of three filtering algorithms for obtaining path turning angle, and b is the error comparison chart of the three filtering algorithms for obtaining path turning angle. DETAILED DESCRIPTION

[0159] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort are within the protection scope of the present application.

[0160] In order to make the above objectives, characteristics and advantages of the present application more apparent, the present application will be described in further detail below with reference to the accompanying drawings and specific embodiments.

[0161] Embodiment one

[0162] In this embodiment, as shown in Figure 1 、 Figure 2 , a method for intelligent path planning based on multi-dimensional cooperative enhancement MDC-AFI adaptive feedback fusion, comprising the following steps:

[0163] S1. Obtain the initial state information of the intelligent robot, and the initial state information includes: initial coordinates, attitude information, obstacle distribution layer information and target point coordinates.

[0164] In S1, the method for obtaining the attitude information includes: using the formula for calculating the quaternion by using the accelerometer and magnetometer data in the IMU:

[0165]

[0166] wherein, θ represents the rotation angle around the unit vector u=[u X , u Y , u Z ], w represents the real part of the quaternion, x, y, z represent the imaginary parts of the quaternion, and respectively represent the components of rotation around the X, Y and Z axes; the initial attitude is substituted into the formula of the quaternion to obtain the initial quaternion q0=[w0, x0, y0, z0];

[0167] The angular increment Δq is calculated by the gyro angular velocity, the quaternion is updated, and the updated quaternion is obtained:

[0168]

[0169] wherein, t0 represents the initial time, and Δt represents the update time step;

[0170] The attitude angle is solved by using the updated quaternion q′=[w′, x′, y′, z′], and the attitude information is obtained:

[0171]

[0172] Pitch=sin -1(2(w′y′-x′z′)), |Pitch|≤90°,

[0173]

[0174] Where Roll represents the roll angle around the X-axis, Pitch represents the pitch angle around the Y-axis, and Yaw represents the yaw angle around the Z-axis;

[0175] The rotation matrix R is obtained by calculating the quaternion q = [w, x, y, z].

[0176]

[0177] The corresponding rotation matrix R is calculated based on the attitude information from each sensor. ε ;

[0178] Set the local coordinate system of each sensor Mapped to global coordinate system This provides a unified benchmark for subsequent algorithms, ensuring that data is correctly transferred and processed between different coordinate systems.

[0179]

[0180] Where ε represents the index of each sensor in the sensor group;

[0181] Reversible transformation, that is, converting global coordinates to local coordinates:

[0182]

[0183] in, This represents the local coordinates of the sensor with index ε.

[0184] S2. Based on the initial state information, call the dynamic adaptive global path planner to iteratively calculate and optimize the global path, obtain the global path nodes, and generate the globally optimized path.

[0185] Methods to optimize global paths include:

[0186] A dynamic adaptive cost heuristic function h′(n) is introduced, which combines multi-dimensional information such as obstacle distribution density, dynamic obstacle velocity direction and Euclidean distance to the target point to optimize path search efficiency.

[0187] Methods to optimize path search efficiency include: calculating the time cost h. time (n), which is the shortest time estimate from node n to target point G:

[0188]

[0189] where ||n-g|| represents the Euclidean distance from the node n to the target point G, and v represents the linear speed of the intelligent robot;

[0190] calculating the spatial distance cost h distance (n), i.e., the Euclidean distance from the node n to the target point G:

[0191]

[0192] where (x n , y n ) represents the coordinates of the node n, and (x g , y g ) represents the coordinates of the target point G;

[0193] calculating the dynamic environment cost h dynamic (n), i.e., the influence of the dynamic obstacle on the path planning:

[0194]

[0195] where k represents the number of dynamic obstacles, d i represents the distance from the node n to the i-th dynamic obstacle, and θ i represents the included angle between the direction from the node n to the target point G and the speed direction of the dynamic obstacle;

[0196] constructing a dynamic adaptive cost heuristic function h'(n) based on the time cost h time (n), the spatial distance cost h distance (n), and the dynamic environment cost h dynamic (n):

[0197] h'(n) = ah time (n) + bh distance (n) + gh dynamic (n),

[0198] where a, b, g respectively represent the weight parameters of h time (n), h distance (n), and h dynamic (n), and a+b+g=1;

[0199] expressing the dynamic adaptive cost heuristic function h'(n) as the product of the weight vector and the cost vector:

[0200] h'(n) = W·H(n),

[0201] where W = [a, b, g] represents the weight vector, and h(n) = [h time (n), h distance (n), h dynamic(n)] T representing a cost vector;

[0202] defining a dynamic environment parameter vector and dynamically adapting the weight vector W according to the environment parameter vector E:

[0203] W = W0+ K·E(t),

[0204]

[0205] E(t) = [ρ obs , d obs , c env ],

[0206]

[0207] wherein W0 represents an initial weight vector, K represents a gain matrix, ρ obs represents a dynamic obstacle density, d obs represents a nearest distance of a node to a dynamic obstacle, c env represents an environment openness, represents a dynamic density function, represents an area of an environment region , p = [x, y] T represents a two-dimensional coordinate of a dynamic obstacle, represents a two-dimensional real space, i represents an index of a dynamic obstacle set, P node represents a node position coordinate to be evaluated, v i represents a linear velocity vector of a dynamic obstacle with index i at time t, P i (t) represents a coordinate of a dynamic obstacle with index i at time t, represents an obstacle set, ||·||2 represents an Euclidean norm, ρ max represents a maximum obstacle density that can be reached in a unit area.

[0208] A path smoothing algorithm (such as Bezier curve fitting, cubic spline interpolation, etc.) is used to optimize the objective function to process the initial global path, reduce the path polyline, and improve the smoothness and controllability of the path.

[0209] A hierarchical grid modeling is used to set a higher weight for a high complexity region and a lower weight for a low complexity region, thereby improving the globality and efficiency of path search.

[0210] S3. The global optimization path is decomposed to obtain several local target points, and these local target points are used as the initial reference inputs of the local path planner. Through the multi-dimensional feasibility dynamic evaluator under the MDC-AFI framework and the distribution information of dynamic obstacles in the local environment, the parameters of the local path planner are adaptively adjusted. Based on the dynamic constraints of the intelligent robot and the real-time environmental information, a dynamically optimized local path is iteratively generated.

[0211] Methods for generating dynamically optimized local paths include:

[0212] The path is decomposed into local target points, taking into account the global path optimized by the improved dynamic adaptive cost heuristic function h′(n).

[0213] The local dynamic obstacle avoidance planner in complex environments is invoked to obtain local paths based on the distribution of dynamic obstacles and the motion state of the intelligent robot. The multidimensional constraint matrix potential field optimization algorithm MDC-MPFO is used to evaluate the impact of local path planning, and dynamic obstacle avoidance adjustments are made in a timely manner based on the evaluation results.

[0214] Methods for obtaining local paths include: matrixing the set of path point coordinates P and the set of obstacle coordinates O.

[0215]

[0216] Where p represents the number of local path points to be evaluated, and m represents the number of obstacles;

[0217] Calculate path point P i =(x i y i ) to target point G = (x g y g The gravitational field strength F att，i :

[0218]

[0219] Where, d goal k represents the gravitational range of the target point. att This is the gravitational field gain coefficient;

[0220] Based on the gravitational field strength F att，i Establish the gravitational field matrix F att :

[0221]

[0222] Among them, F att，xn F represents the x-axis component of the gravitational field strength at each path point to be evaluated. att，yn This represents the y-axis component of the gravitational field strength at each path point to be evaluated.

[0223] Calculate the path point P i = (x i , y i ) to the repulsive force field strength F j of the nearest obstacle O oj = (x oj , y rep，i ) :

[0224]

[0225] Wherein, d safe represents the safety distance, k rep represents the repulsive force field gain coefficient, d obs,i represents the nearest distance between the path point P i = (x i , y i ) and the obstacle O j = (x oj , y oj ) ;

[0226] The repulsive force field matrix F rep，i is established based on the attractive force field strength F rep :

[0227]

[0228] Wherein, F rep，xp represents the x-axis component of the repulsive force field strength of each to-be-evaluated path point to the nearest obstacle, and F rep，yp represents the y-axis component of the repulsive force field strength of each to-be-evaluated path point to the nearest obstacle;

[0229] According to the superposition theorem, the resultant force field F att is constructed based on the attractive force field matrix F rep and the repulsive force field matrix F total :

[0230]

[0231] The path point position of the local path planning is updated based on the resultant force field, and the dynamically optimized local path is generated:

[0232]

[0233] Wherein, P i，new represents the coordinate of the local path point after updating, P i，current represents the coordinate of the local path point before updating, represents the path updating step, and F total，i represents the resultant force field strength at each local path point

[0234] The method for evaluating the influence of local path planning by using the multi-dimensional constraint matrix potential field optimization algorithm MDC-MPFO includes path feasibility evaluation and path generation value evaluation.

[0235] If The local path planning is unreliable, and needs to be re-planned.

[0236] If The local path passes the feasibility evaluation, drives the intelligent robot motion module to the local target point.

[0237] The method for path generation value evaluation includes:

[0238] Taking the sum of the lengths of the force fields F total of all path points as J force :

[0239]

[0240] Wherein, ||F total，i || represents the length of the force field F total , and p represents the number of local path points to be evaluated.

[0241] Calculate the first-order smoothness:

[0242]

[0243] Wherein, J smooth，1st represents the first-order smoothness value, P i represents the i-th local path point, and P i-1 represents the i-1-th local path point.

[0244] Calculate the second-order smoothness:

[0245]

[0246] Wherein, J smooth，2nd represents the second-order smoothness value, and P i+1 represents the i+1-th local path point.

[0247] Weighted fusion, calculate the path planning quality evaluation cost J evaluate :

[0248] J evaluate = J force +λ1·J smooth，1st +λ2·J sniith，2nd ,

[0249] Wherein, λ1 represents the first-order smoothness penalty coefficient, and λ2 represents the second-order smoothness penalty coefficient.

[0250] While dynamically avoiding obstacles, the path planning is constrained by the dynamic characteristics of the intelligent robot, using the discrete point curvature κ. i Assess the smoothness of the path to ensure its smoothness and dynamic feasibility.

[0251] Methods for constrained path planning based on the dynamic characteristics of intelligent robots include: according to the dynamic model of the intelligent robot, the local path needs to satisfy three constraints, including:

[0252] Joint vector form constraint of velocity and angular velocity:

[0253]

[0254] Where Ψ represents the joint vector of velocity and angular velocity, X lim This represents a combined constraint of velocity and angular velocity, where v represents the linear velocity of the intelligent robot and ω represents the angular velocity. min V represents the minimum linear velocity. max ω represents the maximum linear velocity. min ω represents the minimum angular velocity. max Indicates the maximum angular velocity;

[0255] Linear angular acceleration vector form constraint:

[0256]

[0257] in, V represents the joint vector of linear and angular accelerations. lim This represents the combined constraint of linear and angular acceleration, where 'a' represents linear acceleration and 'β' represents angular acceleration. min a represents the minimum linear acceleration. max β represents the maximum linear acceleration. min β represents the minimum angular acceleration. max Indicates the maximum angular acceleration;

[0258] Turning radius vector form constraint:

[0259]

[0260] Where R(Ψ) represents a nonlinear function of the turning radius, R min This represents the minimum turning radius of the robot model.

[0261] Using discrete point curvature κ i The path smoothness is evaluated, where the curvature κ at discrete points is used. i The calculation methods include:

[0262] Select three consecutive local path points and

[0263] Calculate the area S of the curvature triangle by determinant method κi :

[0264]

[0265] Calculate the point pair distance between three points:

[0266]

[0267] Where d i-1,i represents the distance between and d represents the distance between and d represents the distance between and d i,i+1 represents the distance between and d represents the distance between and d represents the distance between and d i-1,i+1 represents the distance between and d represents the distance between and d represents the distance between and d

[0268] Take the maximum point pair distance between three points as the total chord length L:

[0269] L = max{d i-1，i , d i，i+1 , d i-1，i+1};

[0270] Substitute to solve the discrete point curvature κ i :

[0271]

[0272] When the local path point is dense, simplify the discrete point curvature κ i :

[0273]

[0274] S4. When the local path deviates from the global optimization path to the extent that exceeds the preset tolerance range, the multi-dimensional feasibility dynamic evaluator judges whether it is necessary to re-optimize the global path, and feeds back the latest global path data to the local path planner for feedback fusion.

[0275] The method of feedback fusion includes:

[0276] The local target point of the global path decomposition based on dynamic adaptive cost heuristic function h'(n) optimization, after completing each local target point, update the initial position of the intelligent robot.

[0277] When the local path deviates from the global path by more than a set threshold, the multi-dimensional feasibility dynamic evaluation module under the MDC-AFI framework judges whether to trigger the global path planner to re-optimize the global path, and dynamically updates the data fed back by the global path planner to the local path planning module.

[0278] The method for determining whether to trigger the global path planner to re-optimize the global path comprises: calculating a Euclidean distance and a point projection distance between the local path point and the global path point;

[0279] Euclidean distance δ euclidean is:

[0280]

[0281] wherein, is the point coordinate of the local path, is the point coordinate of the global path;

[0282] Point projection distance δ projection is:

[0283]

[0284] wherein, and represent two end points of a certain segment in the global path;

[0285] According to the complexity of the planning scene, the Euclidean distance δ euclidean and the point projection distance δ euclidean are weighted and fused according to the weight parameters σ and ψ to obtain comprehensive deviation data:

[0286] δ fused-linear = σ δ euclidean + ψ δ projection ,

[0287]

[0288] wherein, σ + ψ = 1, δ fused-linear represents linear fusion deviation data, δ fused-nonliner represents nonlinear fusion deviation data, σ represents the fusion weight of the Euclidean distance, and ψ represents the fusion weight of the point projection distance; the path turning change amount Δθ local and Δθ global are calculated to dynamically and adaptively adjust σ and ψ: the local path point state is in the form of a vector including the position and direction angle information of the point; the control input vector is including the linear velocity, angular velocity and acceleration information; a state transition equation is established to describe how the local path point is transferred from the j-1 time to the j time:

[0289]

[0290] where A represents a state transition matrix, which describes the influence of steering changes on the direction angle; B represents a control input matrix, which describes the influence of inputs on the system state; v j represents the velocity, and Δt represents the time step; w j represents the process noise, which is subject to a Gaussian distribution w j ~ N(0, Q), represents the process noise covariance, θ j represents the direction angle at time j; the observation equation is established to describe how to calculate the state from the observation point:

[0291]

[0292] where, represents the state observation value at time j, and H represents an observation matrix: H = [0 0 1], which represents the angle information of the observation path point; l j represents the observation noise, which is subject to a Gaussian distribution l j ~ N(0, γ), and γ is the observation noise covariance. The direction angle at the next time is predicted according to the state transition equation:

[0293]

[0294] where, represents the predicted direction angle, θ j-1 represents the direction angle at time j-1; the prediction error covariance is:

[0295]

[0296] where, P j-1 represents the error covariance at time j-1; the Kalman gain K j is calculated, and the direction angle is updated by using the Kalman filter:

[0297]

[0298] where, θ j represents the direction angle at time j; the error covariance is updated:

[0299] P j = (I-K j H)P j|j-1 ,

[0300] where, P j represents the updated error covariance, I represents a unit matrix, and P j|j-1 represents the predicted error covariance; the above process is repeated to calculate the path steering change amount Δθ local and Δθ global :

[0301]

[0302] wherein, denotes the local orientation angle at time j, denotes the local orientation angle at time j-1, denotes the global orientation angle at time k, denotes the global orientation angle at time j-1. The fusion weights σ, ψ are dynamically adjusted according to the environmental characteristics; when the local path changes drastically, i.e., the steering change amount Δθ loacal of the local path is large, the weight ψ of the point projection distance δ projection is increased; when the local path is relatively smooth, i.e., the steering change amount Δθ global of the local path is small, the weight σ of the Euclidean distance δ euclidean is increased. The weights σ, ψ are dynamically adjusted:

[0303]

[0304] ψ = 1-σ,

[0305] The deviation data δ fused-linear or δ fused-nonliner changes dynamically with the weights σ, ψ:

[0306] δ fused-(non)linear = f(σ, ψ, δ projection , δ euclidean ),

[0307] wherein, f is a function of variables σ, ψ, δ projection , δ euclidean .

[0308] If δ fused-(non)linear > δ max : The multi-dimensional feasibility dynamic evaluation module under the MDC-AFI framework triggers the global path planner to re-optimize the global path, fits the feedback data of the local path planner to obtain a more reliable global path, and then takes it as the input of the local path planner for continuous iterative optimization.

[0309] The path switching priority is set, when the local path is seriously affected by dynamic obstacles, the global path point is taken as the priority reference level point, the global optimality and safety of the path planning are ensured, and the robot is prevented from being trapped in local optimization.

[0310] S5. Repeat S1-S4, through the path feedback mechanism enhanced by multi-dimensional cooperation and the dynamic optimization strategy of adaptive fusion, the real-time updated local path and global path are iteratively generated until the autonomous mobile robot reaches the target shelf, and the goods picking and transportation task is completed.

[0311] As Figure 3As shown, direct integration methods may introduce accumulated errors in numerical computation, especially during long integration processes. Quaternions, due to their mathematical properties (the normality of unit quaternions), maintain high numerical stability and are less prone to drift problems, such as... Figure 3 The three graphs on the left, from top to bottom, show the changes in roll, pitch, and yaw angles over time for direct integration (solid lines) and quaternion solutions (dashed lines) in a typical scenario. Direct integration, if Euler angles are used to represent rotation, can easily encounter gimbal lock, where, in certain attitudes, the two rotation axes coincide, resulting in the loss of a degree of freedom and preventing accurate rotation representation. Quaternions avoid this problem because they do not use Euler angles but represent rotation using four-dimensional vectors, thus always representing any three-dimensional rotation, such as... Figure 3 The three graphs on the right side of the middle section, from top to bottom, show the changes in roll, pitch, and yaw angles over time for direct integration (solid line) and quaternion solution (dashed line) in the Gimbal Lock scenario.

[0312] like Figure 4 As shown, when the autonomous mobile robot performs goods picking and transportation tasks between different warehouses, it evaluates and adjusts the local path planning through a multi-dimensional constraint matrix potential field optimization algorithm to effectively avoid dynamic obstacles. The four sub-graphs, arranged clockwise, show the global path planning and optimization of the transportation task, the dynamic avoidance of dynamic obstacles on the left side of the warehouse during local path tracking, the dynamic avoidance of dynamic obstacles on the right side of the warehouse during local path tracking, and the overall effect of global path optimization and dynamic obstacle avoidance during local path tracking.

[0313] like Figure 5 As shown, the Kalman filter algorithm can more accurately fit the angle data transmitted by the autonomous mobile robot when calculating the path turning change. The error level is lower than that of the other two algorithms, that is, the error lines in the figure are all below the other two algorithms.

[0314] like Figure 6 As shown, compared with the smoothness analysis method based on the second derivative integral, using the curvature of discrete points to evaluate the path smoothness makes the evaluation level fluctuate around the mean of smoothness, with a smaller error range and the evaluation level is closer to the theoretical value.

[0315] Example 2

[0316] In this embodiment, an MDC-AFI adaptive feedback fusion intelligent path planning system based on multi-dimensional collaborative enhancement includes: an initial information acquisition module, a global path optimization module, a local path optimization module, and a feedback fusion module.

[0317] The initial information acquisition module is configured to acquire initial state information of the intelligent robot, and the initial state information includes initial coordinates, attitude information, obstacle distribution layer information and target point coordinates.

[0318] The global path optimization module is configured to call a dynamic self-adaptive global path planner based on the initial state information to iteratively calculate and optimize a global path, obtain a global path node and generate a globally optimized path.

[0319] The local path optimization module is configured to decompose the globally optimized path to obtain a plurality of local target points, take the plurality of local target points as initial reference inputs of a local path planner, adaptively adjust parameters of the local path planner based on distribution information of dynamic obstacles in a local environment through a multi-dimensional feasibility dynamic evaluator under an MDC-AFI framework, iteratively generate a dynamically optimized local path according to dynamic constraint conditions of the intelligent robot and real-time environmental information.

[0320] The feedback fusion module is configured to, when a degree of deviation of the local path from the globally optimized path exceeds a preset tolerance range, judge whether the globally optimized path needs to be re-optimized by the multi-dimensional feasibility dynamic evaluator, and feed back the latest global path data to the local path planner for feedback fusion until the intelligent robot reaches the target point and completes a target navigation task.

[0321] The above-described embodiments are merely descriptions of the preferred modes of the present application and are not intended to limit the scope of the present application. Various modifications and improvements to the technical solutions of the present application made by those of ordinary skill in the art without departing from the design spirit of the present application shall fall within the protection scope of the present application as defined by the claims.

Claims

1. A multi-dimensional cooperative enhancement-based MDC-AFI adaptive feedback fusion intelligent path planning method, characterized in that, The method comprises the following steps: S1. Obtain initial state information of the intelligent robot, the initial state information comprising: initial coordinates, attitude information, obstacle distribution layer information, and target point coordinates; S2. Based on the initial state information, call a dynamic self-adaptive global path planner to search for a global path, iteratively calculate a global path node, and generate a globally optimized path; S3. Decompose the globally optimized path to obtain a plurality of local target points and use the local target points as initial reference inputs of a local path planner, adaptively adjust parameters of the local path planner based on distribution information of dynamic obstacles in the local environment through a multi-dimensional feasibility dynamic evaluator under the MDC-AFI framework, iteratively generate a dynamically optimized local path according to dynamic constraint conditions of the intelligent robot and real-time environmental information; S4. When the local path deviates from the globally optimized path by more than a preset tolerance range, the multi-dimensional feasibility dynamic evaluator determines whether the globally optimized path needs to be re-optimized, and feeds back the latest global path data to the local path planner for feedback fusion; S5. Repeat S1-S4 until the intelligent robot reaches the target point and completes the target navigation task.

2. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional cooperative enhancement according to claim 1, characterized in that, In S1, the method for obtaining the attitude information comprises: A formula for calculating a quaternion using accelerometer and magnetometer data in the IMU is as follows: ， where θ represents a rotation angle around the unit vector w represents a real part of the quaternion, x, y, z represent imaginary parts of the quaternion, and respectively represent components of rotation around the X, Y, and Z axes. Substitute the initial pose into the formula of value quaternion, get the initial quaternion ; An angular increment Δq is calculated by using the angular velocity of the gyroscope, the quaternion is updated, and an updated quaternion is obtained: ， wherein denotes the initial time instant, and At denotes the update time step. using the updated quaternion solving a pose angle to obtain the pose information ， ， ， Where Roll represents a roll angle of rotation around the X axis, Pitch represents a pitch angle of rotation around the Y axis, and Yaw represents a yaw angle of rotation around the Z axis; By quaternions The rotation matrix is computed : ， A rotation matrix is calculated based on the attitude information of each sensor ; Mapping of local coordinate systems of each sensor to a global coordinate system provides a uniform reference for subsequent algorithms, ensuring correct transfer and processing of data between different coordinate systems: ​ ， Where ε represents the index of each sensor in the sensor group; A reversible transformation, i.e., conversion of global coordinates to local coordinates, is as follows: ， wherein, represents the local coordinates of the sensor with index i. 3.The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional synergistic enhancement according to claim 1, characterized in that, The method for optimizing the global path comprises: Introducing dynamic adaptive cost heuristic function , combined with the obstacle distribution density, dynamic obstacle speed direction and the Euclidean distance of the target point, the path search efficiency is optimized; A path smoothing algorithm is used to optimize the objective function to process the initial global path, reduce the number of path break lines, and improve the smoothness and controllability of the path; A hierarchical grid modeling is used to set a higher weight for a high-complexity region and a lower weight for a low-complexity region, thereby improving the globality and efficiency of path searching.

4. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional cooperative enhancement according to claim 3, characterized in that, The method for optimizing the path searching efficiency comprises: Computational time cost i.e. the shortest time estimate from node n to goal point G: ， wherein, denotes the Euclidean distance from the node n to the goal point G, v denotes the linear velocity of the intelligent robot; Computing spatial distance cost i.e. the Euclidean distance of node n to goal point G: ， wherein, denotes the coordinates of the node n, denotes the coordinates of the target point G; Computing dynamic environment cost i.e. the influence of dynamic obstacles on path planning: ， where k represents the number of dynamic obstacles, d i represents the distance from the node n to the i-th dynamic obstacle, θ i represents the angle between the direction from the node n to the goal point G and the direction of the velocity of the dynamic obstacle; based on the temporal cost , the spatial distance cost and the dynamic environment cost to construct a dynamic adaptive cost heuristic function : ， wherein a, β, γ represent weight parameters of , , , and a + β + γ = 1. Dynamic adaptive cost heuristics function is expressed as the product of a weight vector and a cost vector: ， wherein, denotes a weight vector, denotes a cost vector; defining a dynamic environment parameter vector, and adapting said weight vector in dependence on said environment parameter vector dynamically self-adapting said weight vector : ， ， ， ， ， ， in, Represents the initial weight vector. Represents the gain matrix, ρ obs d represents the dynamic obstacle density. obs c represents the nearest distance between a node and a dynamic obstacle. env Indicates the openness of the environment. Represents the dynamic density function, Indicates environmental area area, Represents the two-dimensional coordinates of a dynamic obstacle. Let P represent a two-dimensional real number space, where i represents the index of the set of dynamic obstacles, and P represents the index of the set of obstacles. node v represents the coordinates of the node to be evaluated. i P represents the linear velocity vector of the dynamic obstacle with index i at time t. i (t) represents the coordinates of the dynamic obstacle with index i at time t. Represents a set of obstacles. Represents the Euclidean norm. This indicates the maximum obstacle density that can be reached within a unit area.

5. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional synergistic enhancement according to claim 3, characterized in that, The method for generating the dynamically optimized local path comprises: Improved dynamic adaptive cost heuristic function Optimized global paths are referenced to break the path into local waypoints; A local dynamic obstacle avoidance planner under a complex environment is called to obtain the local path according to the distribution environment of dynamic obstacles and the motion state of the intelligent robot, a multi-dimensional constraint matrix potential field optimization algorithm MDC-MPFO is used to evaluate the influence of local path planning, and dynamic obstacle avoidance adjustment is made in a timely manner according to the evaluation result; In the dynamic obstacle avoidance, the dynamic characteristics of the intelligent robot are combined, the path planning is constrained, and the curvature of the discrete points is calculated The smoothness of the path is evaluated to ensure the smoothness and dynamic feasibility of the path.

6. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional synergistic enhancement according to claim 5, characterized in that, The method for obtaining the local path comprises: A matrixed path point coordinate set P and an obstacle coordinate set O are obtained as follows: ， ， Where p represents the number of local path points to be evaluated, and m represents the number of obstacles; Computing waypoints Gravitational field strength to a target point : ， where d goal is the gravitational range of the target point, k att is the gravitational field gain coefficient; based on the gravitational field strength establishing a gravitational field matrix : ， wherein, represents the x-axis component of the gravitational field strength at each path point to be evaluated, represents the y-axis component of the gravitational field strength at each path point to be evaluated; Computing waypoints Repulsive field strength to nearest obstacle :​ ， ， where d safe represents the safety distance, k rep represents the repulsive field gain coefficient, d obs,i represents the path point to the nearest distance to the obstacle ​ based on the gravitational field strength establishing a repulsive field matrix : ， wherein, represents the x-axis component of the repulsive field strength of the closest obstacle to each path point to be evaluated, represents the y-axis component of the repulsive field strength of the closest obstacle to each path point to be evaluated; According to the superposition theorem, based on the aforementioned gravitational field matrix and the repulsive field matrix Constructing a synergistic field : ， The position of the path point of the local path planning is updated based on the resultant force field, and the dynamically optimized local path is generated: ； wherein, represents the updated coordinate of the local path point, represents the coordinate of the local path point before updating, represents the path updating step size, represents the force field intensity at each local path point.

7. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional synergistic enhancement according to claim 6, characterized in that, The method for evaluating the influence of local path planning by using the multi-dimensional constraint matrix potential field optimization algorithm MDC-MPFO comprises: path feasibility evaluation and path generation value evaluation; The method for path feasibility evaluation comprises: If , the local path planning is unreliable, and needs to be re-planned; If , the local path is evaluated by the feasibility, and the intelligent robot motion module is driven to the local target point; The method for path generation value evaluation comprises: the sum of the lengths of the force fields of the path points :​ ， wherein, represents the module length of the resultant force field p represents the number of local path points to be evaluated; Calculate the first-order smoothness: ， wherein, represents a first order smoothness value, P i represents the i-th local path point, P i-1 represents the i-1-th local path point; Computing second-order smoothness: ， wherein, represents a second order smoothness value, P i+1 represents the i+1th local path point; Weighted fusion, computing path planning quality assessment cost : ， wherein λ1 represents a first-order smoothness penalty coefficient, and λ2 represents a second-order smoothness penalty coefficient.

8. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional synergistic enhancement according to claim 7, characterized in that, The method for constraint path planning in combination with the dynamics characteristics of the intelligent robot comprises: According to the dynamics model of the intelligent robot, the local path needs to satisfy three constraints, including: Velocity-angualr velocity joint vector form constraint: ， ， wherein, denotes the velocity angular velocity joint vector, X lim denotes the velocity angular velocity joint constraint, v denotes the linear velocity of the intelligent robot, ω denotes the angular velocity, v min denotes the minimum linear velocity, v max denotes the maximum linear velocity, ω min denotes the minimum angular velocity, ω max denotes the maximum angular velocity; Linear-angular acceleration vector form constraint: ， ， wherein, denotes the linear angular acceleration joint vector, V lim denotes the linear angular acceleration joint constraint, a denotes the linear acceleration, β denotes the angular acceleration, a min denotes the minimum linear acceleration, a max denotes the maximum linear acceleration, β min denotes the minimum angular acceleration, β max denotes the maximum angular acceleration; Turning radius vector form constraint: ， where R represents a non-linear function of the turning radius, R min represents the minimum turning radius of the robot model.

9. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional cooperative enhancement according to claim 8, characterized in that, Using discrete point curvature An assessment of path smoothness is made, where the discrete point curvature The method of calculating the discrete point curvature comprises: selecting three consecutive local path points , and ; Calculating the area of a curvature triangle by determinant method : ； Computing the point pair distance between three points: ， ， ； where d i-1,i represents the distance between and d i,i+1 represents the distance between and d i-1,i+1 represents the distance between and ; Taking the maximum point pair distance between three points as the total chord length L: ； Substitute to solve discrete point curvature : ； When local path points are dense, the discrete point curvature is reduced to: 。 10. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional synergistic enhancement according to claim 3, characterized in that, The method for feedback fusion comprises: based on the dynamic adaptive cost heuristic function local goal points of the optimized global path decomposition, updating the initial position of the intelligent robot after each local goal point is completed When the local path deviates from the global path by more than a set threshold, the multi-dimensional feasibility dynamic evaluation module under the MDC-AFI framework judges whether to trigger the global path planner to re-optimize the global path, and dynamically updates the data fed back from the global path planner to the local path planning module; Setting path switching priority, when the local path is seriously affected by dynamic obstacles, taking the global path point as the priority reference level point, ensuring the global optimality and safety of path planning, and preventing the robot from falling into local optimality.

11. The MDC-AFI adaptive feedback fusion intelligent path planning method based on multi-dimensional synergistic enhancement according to claim 10, characterized in that, The method for judging whether to trigger the global path planner to re-optimize the global path comprises: Computing the Euclidean distance and point projection distance between the local path point and the global path point; Euclidean distance is: ， wherein, is a point coordinate of the local path, is a point coordinate of the global path; Point projection distance is: ， wherein and denotes the two end points of a certain segment of the global path. According to the complexity of the planning scene combined with the Euclidean distance and point projection distance , weighted fusion of comprehensive deviation data is obtained according to the weight parameter , ​ ， ， wherein, , denotes linear fusion bias data, denotes non-linear fusion bias data, denotes a fusion weight of Euclidean distance, denotes a fusion weight of point projection distance; When detecting fusion bias data And Greater than bias threshold When, the multi-dimensional feasibility dynamic evaluator under the MDC-AFI framework triggers the global path planning to re-optimize the global path, fits the feedback data of the local path planner to obtain a more reliable global path, and takes the obtained global path as the input of the local path planner for continuous iteration optimization.

12. A multi-dimensional synergistically enhanced MDC-AFI adaptive feedback fusion intelligent path planning system, the system applying the method of any one of claims 1-11, characterized in that, Comprise: An initial information acquisition module, a global path optimization module, a local path optimization module and a feedback fusion module; The initial information acquisition module is used to acquire initial state information of the intelligent robot, and the initial state information comprises initial coordinates, attitude information, obstacle distribution layer information and target point coordinates; The global path optimization module calls a dynamic adaptive global path planner to search for a global path based on the initial state information, iteratively calculates global path nodes, and generates a globally optimized path; The local path optimization module is used to decompose the globally optimized path to obtain a plurality of local target points and take them as initial reference inputs of a local path planner, adaptively adjust parameters of the local path planner through a multi-dimensional feasibility dynamic evaluator under the MDC-AFI framework and distribution information of dynamic obstacles in a local environment, iteratively generate a dynamically optimized local path according to dynamics constraint conditions of the intelligent robot and real-time environmental information; The feedback fusion module judges whether to re-optimize the global path when the degree of deviation of the local path from the globally optimized path exceeds a preset tolerance range, and feeds back the latest global path data to the local path planner for feedback fusion until the intelligent robot reaches the target point and completes the target navigation task.

Images