Robot path planning method based on multi-strategy improved Harris hawk algorithm

Abstract

The application discloses a robot path planning method based on a multi-strategy improved Harris hawk algorithm, and the method comprises the following steps: setting initial parameters; using a chaos initialization strategy, introducing a Tent chaos mapping to initialize a population; using a fuzzy energy strategy to calculate prey escape energy; selecting an exploration stage or a development stage according to the escape energy; when |E| is greater than or equal to 1, an individual enters the exploration stage, and when |E| is less than 1, the individual enters the development stage; using a positive sine disturbance method to try to jump out of a local optimal solution; judging whether a maximum iteration number is reached, if yes, outputting an optimal path, and if not, returning to the prey escape energy calculation step; and constructing a path trajectory graph and a fitness curve graph of the robot in a position space. The application aims to solve problems that the robot faces when performing path planning, such as path non-smoothness, slow convergence speed and easy falling into a dead zone.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot path planning, and relates to a robot path planning method based on a multi-strategy improved Harris Hawk algorithm. Background Art

[0002] The artificial potential field method is a commonly used robot path planning algorithm. Its basic concept is to construct a virtual potential field and use its forces to guide the robot to avoid obstacles and reach its target. The algorithm treats the target point as a source of gravity, exerting an attractive force on the robot; it treats obstacles as sources of repulsion, exerting a repulsive force on the robot. The robot moves under the influence of this combined force. Starting from a starting point, the robot calculates the total force (the combined force of gravity and repulsion) at its current position and moves in the direction of that force until it reaches the target point or encounters an unavoidable obstacle.

[0003] A major drawback of the artificial potential field method is its tendency to become trapped in local optima. For example, a robot may become trapped in a "power equilibrium point" or "dead zone," preventing it from continuing toward its target. This situation often occurs in environments with numerous obstacles or complex paths near the target. Furthermore, because the robot only moves in the direction of its current total power, the path planning process is relatively localized, potentially resulting in an uneven overall path and affecting the robot's efficiency and stability. Summary of the Invention

[0004] The technical solution adopted by the present invention is: a robot path planning method based on a multi-strategy improved Harris Eagle algorithm, which aims to solve the problems faced by robots during path planning, such as uneven paths, slow convergence speed, and easy falling into dead zones.

[0005] The robot path planning method based on the multi-strategy improved Harris Hawk algorithm includes the following specific steps:

[0006] Step 1: Set initial parameters;

[0007] Based on the range of the robot's motion space and the distribution of obstacles, the initial parameters of the method used are determined, including the size of the position space, the robot's starting and ending points, the population size, and the maximum number of iterations. Different individuals represent different paths in the position space, and prey represents the current optimal path. The population size represents the number of individuals existing at the same time, and the maximum number of iterations is the maximum number of times an individual path can be iterated. The quality of the path is calculated and evaluated by the objective function.

[0008] Step 2: Use the chaos initialization strategy and introduce the Tent chaos map to initialize the population;

[0009] To overcome the problem of insufficient initial population diversity in existing techniques, this paper introduces the Tent Chaos Map method to initialize the algorithm population. This method generates an initial population with high uniformity, ensuring a uniform distribution across the entire position space, thereby improving the algorithm's global search capability and initial diversity. This strategy helps prevent the algorithm from falling into local optimal solutions during the search process, improving the overall effectiveness of path planning.

[0010] Step 3: Calculate the prey escape energy using fuzzy energy strategy;

[0011] In the HHO algorithm (Harris Hawks Optimization, a swarm-based, nature-inspired optimization algorithm), the size of the escape energy E determines whether the Harris Hawk's search behavior enters the exploration phase or the exploitation phase, which has a huge impact on the algorithm's convergence speed. The escape energy in the original algorithm uses the random number E0 as its initial value and decreases linearly, which cannot meet the algorithm's requirements of more exploration behavior in the early stage and more exploitation in the later stage. Therefore, a nonlinear strategy is introduced to improve the HHO algorithm.

[0012] Step 4: Select the exploration phase or the development phase based on the escape energy;

[0013] When E≥1, the individual enters the exploration stage, and when E<1, the individual enters the development stage;

[0014] Step 6: Determine whether the maximum number of iterations has been reached. If so, output the optimal path. Otherwise, return to step 3.

[0015] Step 7: Construct the robot's path trajectory diagram and fitness curve diagram in the position space

[0016] According to the actual starting point, end point and spatial obstacle distribution of the robot's movement, a simulation graph is drawn, and the optimal path planning is performed according to the method proposed in this invention. The search results of the optimal path in 50 iterations are plotted as a fitness curve.

[0017] Furthermore, the Tent chaos mapping formula in step 2 above is:

[0018]

[0019] Where: i is the size of the initial population, λ is a constant, T i D is a random number in the interval [0,1], is an i×D-dimensional chaotic sequence; the population can be initialized as:

[0020]

[0021] Where: X is the position of the individual, t is the current iteration number, LB and UB are the lower and upper limits of the individual position space respectively.

[0022] Furthermore, the prey escape energy calculation in step 3 above introduces a nonlinear strategy to improve the HHO algorithm. The improved prey escape energy model is as follows:

[0023]

[0024] Where: E is the escape energy of the prey, E0 is a random number in the interval (-1, 1), representing the initial energy of the prey, and T is the maximum number of iterations;

[0025] When the prey's escape energy gradually decreases, the HHO algorithm will choose to enter the development phase, which makes it more likely to fall into the local optimal solution. Inspired by the simulated annealing algorithm, the exploration probability p is introduced, so that after the algorithm finds that it has entered the local optimal solution, it has a probability to ignore the prey's escape energy and enter the exploration phase to escape the local optimal solution. The probability formula is as follows:

[0026]

[0027] Where: C is the temperature coefficient, which gradually decreases with the number of iterations, C = αC0, C0 is the initial temperature, α is the decay factor, F(x) is the fitness function, and ΔX(t) is the distance between the prey position and the current individual position.

[0028] Furthermore, the method for selecting the exploration phase or development phase in step 4 above is:

[0029] When E≥1, the individual enters the exploration stage, and when E<1, the individual enters the development stage:

[0030] 1) Exploration phase

[0031] In this stage, the individual will update its position according to the following formula. q ≥0.5, the individual chooses to fly randomly. q When the value is less than 0.5, individuals move closer to other individuals and perform convergent flight:

[0032]

[0033] λ(t)=X rand (t)-2r3X(t) (6)

[0034] K=LB+r4(UB-LB) (7)

[0035] Where: X is the position of the individual, t is the current iteration number, X rand(t) is the position of an individual randomly selected from the population during iteration, λ(t) represents the movement trend of the current individual toward the random individual, r1~r4 and q are all random numbers in the interval (0, 1), K is the random position in the position space, LB and UB are the lower and upper limits of the individual position space, respectively, and X rabbit is the location of the prey, that is, the optimal solution so far, X m (t) is the average position of all individuals at the tth generation;

[0036] 2) Development phase

[0037] After entering the exploitation phase, the HHO algorithm selects one of the four exploitation strategies (1)-(4) according to the prey's escape energy E and the probability of escape r, where r is a random number in the interval (0, 1);

[0038] (1) When r ≥ 0.5 and |E| ≥ 0.5, the HHO algorithm selects the soft encirclement strategy, and the position update method is as follows:

[0039] X(t+1)=ΔX(t)-E|JX rabbit (t)-X(t)| (8)

[0040] ΔX(t)=X rabbit (t)-X(t) (9)

[0041] J=2(1-r5) (10)

[0042] Where: ΔX(t) is the distance between the prey's position and the current individual's position, J is the random jump intensity of the prey during the escape process, and r5 is a random number in the interval (0, 1);

[0043] (2) When r ≥ 0.5 and |E| < 0.5, the HHO algorithm selects the hard encirclement strategy, and the position update method is as follows:

[0044] X(t+1)=X rabbit (t)-E|X rabbit (t)-X(t)| (11)

[0045] (3) When r < 0.5 and E ≥ 0.5, the HHO algorithm selects the diving soft encirclement strategy and uses formula (12) to update the individual position. If the fitness is not improved, the Lévy flight is performed according to formula (13). If it is still not improved, the original position is used, as shown in formula (16):

[0046] Y1=X rabbit (t)-E|JX rabbit (t)-X(t)| (12)

[0047] Z1=Y1+S×L(D) (13)

[0048]

[0049] Where: D is the dimension of the problem to be solved, S is a D-dimensional random vector, L(x) is the levy flight function, u and v are random numbers in the interval (0, 1), β is a constant, which is 1.5, F(x) is the fitness function, and Γ(x) is the Gamma function;

[0050] (4) When r < 0.5 and |E| < 0.5, the HHO algorithm selects the dive hard encirclement strategy and uses formula (17) to update the individual position. If the fitness is not improved, the Lévy flight is performed according to formula (18). If it is still not improved, the original position is used, as shown in formula (19):

[0051] Y2=X rabbit (t)-EJX rabbit (t)-X m (t) (17)

[0052] Z2=Y2+S×L(D) (18)

[0053]

[0054] Furthermore, the sine-cosine perturbation method in step 5 above adopts an adaptive sine-cosine algorithm, and its specific position update formula is as follows:

[0055]

[0056] φ=|r8·X rabbit (t)-X(t) (21)

[0057] Where: w(t) and λ(t) are adaptive weight coefficients, which are updated according to the following formula respectively; r6, r7 and r8 are random numbers in the intervals [0, 2π], [0, 1] and [-2, 2] respectively;

[0058]

[0059] Furthermore, the objective function is determined as follows:

[0060] 1) Establish evaluation indicators for robot path planning, including path length, path smoothness, convergence speed, obstacle avoidance ability, and global optimality;

[0061] 2) Based on the evaluation indicators in step 1), the objective function for evaluating the robot path planning effect is determined as follows:

[0062]

[0063] Where: p is the path, n is the total number of nodes, C is the penalty factor, if the path passes through an obstacle and a collision occurs, C takes a larger value, if there is no collision, it takes a smaller value, i is the i-th node in the entire path. i and y i are the horizontal and vertical coordinates of the i-th node respectively.

[0064] Beneficial effects of the present invention: Compared with the prior art, the effects of the present invention are as follows:

[0065] 1) Improved initial population diversity: This paper introduces the Tent chaos mapping method to initialize the algorithm's initial population, making it evenly distributed throughout the entire position space. This measure effectively improves the diversity of the initial population, enhances the algorithm's global search capability, and reduces the possibility of the algorithm converging to a local optimal solution early on, thereby ensuring the comprehensiveness and accuracy of path planning.

[0066] 2) Accelerated algorithm convergence: This invention uses an adaptive sine-cosine algorithm to dynamically adjust the search behavior during path planning. This precise search strategy accelerates the algorithm's convergence, particularly in the later stages of path planning. Compared to traditional methods, this invention can find the optimal path more quickly, significantly improving path planning efficiency and making it particularly suitable for robotic applications requiring real-time response.

[0067] 3) Avoiding Local Optimal Solutions: To address the problem of traditional algorithms easily falling into local optimal solutions, this paper introduces a nonlinear fuzzy energy strategy to improve the search behavior of the Harris Hawk algorithm. By adjusting the nonlinear decrease of the escape energy, sufficient exploration is ensured in the early stages. In addition, the introduction of the exploration probability mechanism allows the algorithm to jump out of the local optimal solution with probability, further avoiding the problem of local optimal solutions and improving the global optimization capability of path planning.

[0068] 4) Enhanced smoothness and safety of path planning: By optimizing the algorithm structure, the present invention reduces the number of turning points in the path, significantly improving the smoothness of the planned path, thereby improving the stability and efficiency of the robot's movement; at the same time, because the algorithm can effectively avoid obstacles, the safety of path planning is significantly improved, reducing the risk of collisions of the robot in complex environments.

[0069] 5) Enhanced algorithm robustness: By combining multiple strategies, the algorithm exhibits high robustness under different environmental conditions and parameter settings. The algorithm can stably generate high-quality path planning results, is highly adaptable, and is suitable for a variety of complex application scenarios.

[0070] In summary, the present invention effectively solves the problems of many inflection points, slow convergence speed, and easy falling into local optimal solutions in the existing technology through the comprehensive application of multiple strategies, significantly improves the path planning effect of the robot in a multi-obstacle environment, and has significant technological advancement and practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 This is a flow chart of a robot path planning method based on a multi-strategy improved Harris Eagle algorithm;

[0072] Figure 2 This is the escape energy model diagram of the original HHO algorithm;

[0073] Figure 3 This is a diagram of an improved escape energy model based on the multi-strategy improved Harris Hawk algorithm;

[0074] Figure 4 It is a robot path planning method based on the multi-strategy improved Harris Eagle algorithm to generate a path trajectory diagram of the robot in a 100×100 simulation map;

[0075] Figure 5 This is a fitness curve diagram of the robot after 50 iterations in the simulation map using the robot path planning method based on the multi-strategy improved Harris Eagle algorithm. DETAILED DESCRIPTION

[0076] The present invention will be further described below with reference to specific embodiments.

[0077] Example 1: Figure 1-5 As shown in FIG, the robot path planning method based on the multi-strategy improved Harris Hawk algorithm includes the following specific steps:

[0078] Step 1: Set initial parameters;

[0079] Based on the range of the robot's motion space and the distribution of obstacles, the initial parameters of the method used are determined, including the size of the position space, the robot's starting and ending points, the population size, and the maximum number of iterations. Different individuals represent different paths in the position space, and prey represents the current optimal path. The population size represents the number of individuals existing at the same time, and the maximum number of iterations is the maximum number of times an individual path can be iterated. The quality of the path is calculated and evaluated by the objective function.

[0080] Step 2: Use the chaos initialization strategy and introduce the Tent chaos map to initialize the population;

[0081] To overcome the problem of insufficient initial population diversity in existing technologies, this paper introduces the Tent Chaos Map method to initialize the algorithm population. Tent Chaos Map can generate an initial population with high distribution uniformity, ensuring a uniform distribution of the population across the entire position space, thereby improving the algorithm's global search capability and initial diversity. This strategy helps prevent the algorithm from falling into local optimal solutions during the search process, improving the overall effectiveness of path planning.

[0082] The Tent chaos mapping formula in step 2 is:

[0083]

[0084] Where: i is the size of the initial population, λ is a constant, T i D is a random number in the interval [0,1], is an i×D-dimensional chaotic sequence; the population can be initialized as:

[0085]

[0086] Where: X is the position of the individual, t is the current iteration number, LB and UB are the lower and upper limits of the individual position space respectively.

[0087] Step 3: Calculate the prey escape energy using fuzzy energy strategy;

[0088] In the HHO algorithm (Harris Hawks Optimization, a swarm-based, nature-inspired optimization algorithm), the escape energy E determines whether the Harris Hawk's search behavior enters the exploration phase or the exploitation phase, which has a significant impact on the algorithm's convergence speed. The original algorithm uses the random number E0 as its initial value and decreases linearly, which cannot meet the algorithm's requirements of more exploration in the early stages and more exploitation in the later stages. Therefore, a nonlinear strategy is introduced to improve the HHO algorithm. The improved escape energy model is shown below:

[0089]

[0090] Where: E is the escape energy of the prey, E0 is a random number in the interval (-1, 1), representing the initial energy of the prey, and T is the maximum number of iterations;

[0091] When the prey's escape energy gradually decreases, the HHO algorithm will choose to enter the development phase, which makes it more likely to fall into the local optimal solution. Inspired by the simulated annealing algorithm, the exploration probability p is introduced, so that after the algorithm finds that it has entered the local optimal solution, it has a probability to ignore the prey's escape energy and enter the exploration phase to escape the local optimal solution. The probability formula is as follows:

[0092]

[0093] Where: C is the temperature coefficient, which gradually decreases with the number of iterations, C = αC0, C0 is the initial temperature, α is the decay factor, F(x) is the fitness function, ΔX(t) is the distance between the prey position and the current individual position;

[0094] Step 4: Select the exploration phase or the development phase based on the escape energy;

[0095] When E≥1, the individual enters the exploration stage, and when E<1, the individual enters the development stage:

[0096] The method for selecting the exploration phase or development phase in step 4 is:

[0097] 1) Exploration phase

[0098] In this stage, individuals will update their positions according to the following formula. When q≥0.5, individuals will choose to fly randomly. When q<0.5, individuals will move closer to other individuals and perform convergent flight:

[0099]

[0100] λ(t)=X rand (t)-2r3X(t) (6)

[0101] K=LB+r4(UB-LB) (7)

[0102] Where: X is the position of the individual, t is the current iteration number, X rand (t) is the position of an individual randomly selected from the population during iteration, λ(t) represents the movement trend of the current individual toward the random individual, r1~r4 and q are all random numbers in the interval (0, 1), K is the random position in the position space, LB and UB are the lower and upper limits of the individual position space, respectively, and X rabbit is the location of the prey, that is, the optimal solution so far, X m (t) is the average position of all individuals at the tth generation;

[0103] 2) Development phase

[0104] After entering the exploitation phase, the HHO algorithm selects one of the four exploitation strategies (1)-(4) according to the prey's escape energy E and the probability of escape r, where r is a random number in the interval (0, 1);

[0105] (1) When r ≥ 0.5 and |E| ≥ 0.5, the HHO algorithm selects the soft encirclement strategy, and the position update method is as follows:

[0106] X(t+1)=ΔX(t)-E|JXrabbit (t)-X(t)| (8)

[0107] ΔX(t)=X rabbit (t)-X(t) (9)

[0108] J=2(1-r5) (10)

[0109] Where: ΔX(t) is the distance between the prey's position and the current individual's position, J is the random jump intensity of the prey during the escape process, and r5 is a random number in the interval (0, 1);

[0110] (2) When r ≥ 0.5 and |E| < 0.5, the HHO algorithm selects the hard encirclement strategy, and the position update method is as follows:

[0111] X(t+1)=X rabbit (t)-E|X rabbit (t)-X(t)| (11)

[0112] (3) When r < 0.5 and |E| ≥ 0.5, the HHO algorithm selects the diving soft encirclement strategy and uses formula (12) to update the individual position. If the fitness is not improved, the Lévy flight is performed according to the formula. If it is still not improved, the original position is used, as shown in formula (16):

[0113] Y1=X rabbit (t)-E|JX rabbit (t)-X(t)| (12)

[0114] Z1=Y1+S×L(D) (13)

[0115]

[0116]

[0117] Where: D is the dimension of the problem to be solved, S is a D-dimensional random vector, L(x) is the levy flight function, u and v are random numbers in the interval (0, 1), β is a constant, which is 1.5, F(x) is the fitness function, and Γ(x) is the Gamma function;

[0118] (4) When r < 0.5 and |E| < 0.5, the HHO algorithm selects the dive hard encirclement strategy and uses formula (17) to update the individual position. If the fitness is not improved, the Lévy flight is performed according to formula (18). If it is still not improved, the original position is used, as shown in formula (19):

[0119] Y2=X rabbit (t)-EJX rabbit (t)-X m(t) (17)

[0120] Z2=Y2+S×L(D) (18)

[0121]

[0122] Step 5: Use the sine and cosine perturbation method to try to escape the local optimal solution;

[0123] Since the HHO algorithm is prone to falling into a local optimal solution during the iteration process, in order to solve this problem, the adaptive sine-cosine algorithm is introduced in the final stage of the algorithm. The specific position update formula is as follows:

[0124]

[0125] φ=|r8·X rabbit (t)-X(t) (21)

[0126] Where: w(t) and λ(t) are adaptive weight coefficients, which are updated according to the following formula respectively; r6, r7 and r8 are random numbers in the intervals [0, 2π], [0, 1] and [-2, 2] respectively;

[0127]

[0128]

[0129] Step 6: Determine whether the maximum number of iterations has been reached. If so, output the optimal path; otherwise, return to step 3.

[0130] Step 7: Construct the robot's path trajectory diagram and fitness curve diagram in the position space

[0131] According to the actual robot movement starting point, end point and position space obstacle distribution, a simulation graph is drawn, and the optimal path planning is performed according to the method proposed in this invention. The planning results are as follows: Figure 4 As shown, the search results of the optimal path in 50 iterations are plotted as a fitness curve, as shown in Figure 5 shown.

[0132] In summary, the present invention introduces a chaos initialization strategy, a fuzzy energy strategy, and a sine-cosine perturbation strategy. The present invention aims to improve the global search capability and convergence speed of the algorithm, and effectively avoid falling into a local optimal solution during the path planning process, thereby optimizing the overall path planning effect and operating efficiency of the robot, and effectively solving the problems faced by the robot during path planning, such as uneven path, slow convergence speed, and easy falling into dead zones.

Claims

1. A robot path planning method based on a multi-strategy improved Harris Hawk algorithm, characterized by: The method comprises the following steps: Step 1: Set initial parameters; The initial parameters include the size of the position space, the starting and ending points of the robot, the population size, and the maximum number of iterations. Different individuals represent different paths in the position space, and the prey represents the current optimal path. The population size represents the number of individuals existing at the same time, and the maximum number of iterations is the maximum number of times an individual path can be iterated. The quality of the path is calculated and evaluated by the objective function. Step 2: Use the chaos initialization strategy and introduce the Tent chaos map to initialize the population; Step 3: Calculate the prey escape energy using fuzzy energy strategy; Step 4: Select the exploration phase or the development phase based on the escape energy; When |E|≥1, the individual enters the exploration stage, and when |E|<1, the individual enters the development stage; Step 5: Use the sine and cosine perturbation method to try to escape the local optimal solution; Step 6: Determine whether the maximum number of iterations has been reached. If so, output the optimal path. Otherwise, return to step 3. Step 7: Construct the path trajectory diagram and fitness curve diagram of the robot in the position space; According to the actual robot movement starting point, end point and position space obstacle distribution, a simulation graph is drawn, and the optimal path planning is performed according to steps 1-7. The search situation of the optimal path in 50 iterations is plotted as a fitness curve.

2. The robot path planning method based on the multi-strategy improved Harris Hawk algorithm according to claim 1, characterized in that: The Tent chaos mapping formula in step 2 is: Where: i is the size of the initial population, λ is a constant, T i D is a random number in the interval [0,1], Is an i×D-dimensional chaotic sequence, from which the population is initialized as: Where: X is the position of the individual, t is the current iteration number, LB and UB are the lower and upper limits of the individual position space respectively.

3. The robot path planning method based on the multi-strategy improved Harris Hawk algorithm according to claim 1, characterized in that: In step 3, the prey escape energy calculation introduces a nonlinear strategy to improve the HHO algorithm. The improved prey escape energy model is as follows: Where: E is the escape energy of the prey, E0 is a random number in the interval (-1, 1), representing the initial energy of the prey, and T is the maximum number of iterations; The introduction of exploration probability p allows the algorithm to ignore the prey's escape energy and enter the exploration phase after finding that it has entered the local optimal solution, thereby escaping the local optimal solution. The probability formula is as follows: Where: C is the temperature coefficient, which gradually decreases with the number of iterations, C = αC0, C0 is the initial temperature, α is the decay factor, F(x) is the fitness function, and ΔX(t) is the distance between the prey position and the current individual position.

4. The robot path planning method based on the multi-strategy improved Harris Hawk algorithm according to claim 3, characterized in that: The method for selecting the exploration phase or development phase in step 4 is: 1) Exploration phase In this stage, the individual will update its position according to the following formula. q ≥0.5, the individual chooses to fly randomly. q When the value is less than 0.5, individuals move closer to other individuals and perform convergent flight: λ(t)=|X rand (t)-2r3X(t)| (6) K=LB+r4(UB-LB) (7) Where: X is the position of the individual, t is the current iteration number, X rand (t) is the position of an individual randomly selected from the population during iteration, λ(t) represents the movement trend of the current individual toward the random individual, r1~r4 and q are all random numbers in the interval (0, 1), K is the random position in the position space, LB and UB are the lower and upper limits of the individual position space, respectively, and X rabbit is the location of the prey, that is, the optimal solution so far, X m (t) is the average position of all individuals at the tth generation; 2) Development phase After entering the exploitation phase, the HHO algorithm selects one of the four exploitation strategies (1)-(4) according to the prey's escape energy E and the probability of escape r, where r is a random number in the interval (0, 1); (1) When r ≥ 0.5 and |E| ≥ 0.5, the HHO algorithm selects the soft encirclement strategy, and the position update method is as follows: X(t+1)=ΔX(t)-E∣JX rabbit (t)-X(t)| (8) ΔX(t)=X rabbit (t)-X(t) (9) J＝2(1-r5) (10) Where: ΔX(t) is the distance between the prey's position and the current individual's position, J is the random jump intensity of the prey during the escape process, and r5 is a random number in the interval (0, 1); (2) When r ≥ 0.5 and |E| < 0.5, the HHO algorithm selects the hard encirclement strategy, and the position update method is as follows: X(t+1)=X rabbit (t)-E|X rabbit (t)-X(t)| (11) (3) When r < 0.5 and |E| ≥ 0.5, the HHO algorithm selects the diving soft encirclement strategy and uses formula (12) to update the individual position. If the fitness is not improved, the Lévy flight is performed according to formula (13). If it is still not improved, the original position is used, as shown in formula (16): Y1=X rabbit (t)-E|JX rabbit (t)-X(t)| (12) Z1=Y1+S×L(D) (13) Where: D is the dimension of the problem to be solved, S is a D-dimensional random vector, L(x) is the levy flight function, u and v are random numbers in the interval (0, 1), β is a constant, F(x) is the fitness function, and Γ(x) is the Gamma function; (4) When r < 0.5 and |E| < 0.5, the HHO algorithm selects the dive hard encirclement strategy and uses formula (17) to update the individual position. If the fitness is not improved, the Lévy flight is performed according to formula (18). If it is still not improved, the original position is used, as shown in formula (19): Y2=X rabbit (t)-E|JX rabbit (t)-X m (t)| (17) Z2=Y2+S×L(D) (18) 5. The robot path planning method based on the multi-strategy improved Harris Hawk algorithm according to claim 4, characterized in that: The sine-cosine perturbation method in step 5 adopts the adaptive sine-cosine algorithm, and its specific position update formula is as follows: φ=|r8·X rabbit (t)-X(t)| (21) Where: w(t) and λ(t) are adaptive weight coefficients, which are updated according to the following formula respectively; r6, r7 and r8 are random numbers in the intervals [0, 2π], [0, 1] and [-2, 2] respectively; 6. The robot path planning method based on the multi-strategy improved Harris Hawk algorithm according to claim 1, characterized in that: The objective function is determined as follows: 1) Establish evaluation indicators for robot path planning, including path length, path smoothness, convergence speed, obstacle avoidance ability, and global optimality; 2) Based on the evaluation indicators in step 1), the objective function for evaluating the robot path planning effect is determined as follows: Where: p is the path, n is the total number of nodes, C is the penalty factor. If the path passes through an obstacle and a collision occurs, C takes a larger value, and if there is no collision, it takes a smaller value. i is the i-th node in the entire path; x i and y i are the horizontal and vertical coordinates of the i-th node respectively.

Images