Omega-shaped cavity soft gripper optimization design method

Abstract

The invention relates to the technical field of soft gripper design, and provides an omega-shaped cavity soft gripper optimization design method which comprises the steps that key design parameters of an omega-shaped cavity soft gripper and the range of the key design parameters are determined, and a target function is established; based on the whale algorithm, generating a first-generation population of the key design parameters; on the basis of an NSGA-II algorithm, randomly generating Nnsga-k auxiliary solutions from the first generation population; combining the auxiliary solution with the first-generation population to obtain a second-generation population; performing crossover operation and mutation operation on the second-generation population, and processing based on an environment selection mechanism to obtain a third-generation population; the maximum iteration number T or solution set convergence is achieved, and finally a Pareto optimal solution set containing the grabbing force and the bending angle is output. The advantages of the whale algorithm and the NSGA-II are fused, the comprehensive performance of the Omega-shaped cavity soft gripper is improved, and the gripping force and the bending angle are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of soft gripper design, and in particular to an optimization design method for an Ω-shaped cavity soft gripper. Background Art

[0002] In the field of intelligent manufacturing and flexible robotics, Ω-shaped cavity soft grippers are widely used in scenarios such as precision assembly, minimally invasive medical surgery, and underwater operations due to their bionic properties and environmental adaptability.

[0003] As shown in Figure 1(a)-(b), the Ω-shaped cavity soft gripper has five Ω-shaped cavities 1 arranged in parallel. The five Ω-shaped cavities 1 are connected to each other, and an air inlet 2 is left on the outer side of one of the Ω-shaped cavities. Gas enters the five connected Ω-shaped cavities from the air inlet, generating positive pressure to cause the Ω-shaped cavities to expand and bend. Its main body is designed as an Ω-shaped cavity and is made of silicone material. The gripper can adapt to objects of different shapes and sizes. During the grasping process, it can deform according to the contour of the object to achieve a tight fit. In the design of the Ω-shaped cavity soft gripper, the structural parameters of the Ω-shaped cavity soft gripper include the total length L, cavity depth h, cavity width w, cavity radius r, cavity wall thickness d, and cavity spacing l. Among them, the cavity height h, cavity width w, cavity radius r, cavity wall thickness d, cavity spacing l and silicone material hardness Shore A need to be determined to maximize the grasping force F and the bending angle θ. Traditional soft grippers are complex in structure and manufacturing, resulting in high costs. They rely heavily on the designer's experience for fabrication and parameter adjustment, making adjustments and production difficult and inefficient. They also have limited understanding and application of biological principles and may not fully replicate biological grasping abilities. Furthermore, the design process typically focuses on a single objective, failing to consider multiple objectives, making it difficult to achieve a multi-objective balance between gripping force and bending angle. Summary of the Invention

[0004] The present invention mainly solves the technical problems that traditional soft grippers mainly rely on the designer's experience for production and parameter adjustment, which are difficult to adjust and produce, have low efficiency, and it is difficult to achieve a multi-objective balance between grasping force and bending angle. An optimization design method for Ω-shaped cavity soft grippers is proposed, which combines the advantages of whale algorithm and NSGA-II to improve the comprehensive performance of Ω-shaped cavity soft grippers and greatly improve the grasping force and bending angle.

[0005] The present invention provides an optimization design method for an Ω-shaped cavity soft gripper, comprising the following steps:

[0006] Step 1: Clarify the key design parameters and their ranges of the Ω-shaped cavity soft gripper and establish the objective function;

[0007] Step 2: Generate the first generation population of key design parameters based on the whale algorithm;

[0008] Step 3: Based on the NSGA-II algorithm, N is randomly generated from the first generation population. nsga -k auxiliary solutions; then merge the auxiliary solutions with the first generation population to obtain the second generation population;

[0009] Step 4: Perform crossover and mutation operations on the second generation population, and then process it based on the environmental selection mechanism to obtain the third generation population;

[0010] Step 5: When the maximum number of iterations T is reached or the solution set converges, the Pareto optimal solution set including the grasping force and bending angle is output.

[0011] Furthermore, the key design parameters include: cavity depth h, cavity width w, cavity radius r, cavity wall thickness d, cavity spacing l, and silicone material hardness Shore A.

[0012] Furthermore, the ranges of the key design parameters include: cavity depth h ranges from 10-30 mm, cavity width w ranges from 5-20 mm, cavity radius r ranges from 10-25 mm, cavity wall thickness d ranges from 1-3 mm, cavity spacing l ranges from 20-55 mm, and silicone material hardness Shore A ranges from 20-40.

[0013] Furthermore, the objective function includes: maximizing the gripping force F as a first objective function f1, and maximizing the bending angle θ as a second objective function f2;

[0014]

[0015] Where μ is the friction coefficient; k F is the contact area correction factor; β is the high-order nonlinear coefficient; C1 is the first hyperelastic material constant, C2 is the second hyperelastic material constant; p is the air pressure; E eq is the equivalent elastic modulus;

[0016] The calculation formula of the bending angle θ is:

[0017]

[0018] Among them, α is the nonlinear correction coefficient; k θ is the multi-cavity geometry correction factor.

[0019] Furthermore, step 2 includes the following steps 201 to 205:

[0020] Step 201: using the whale algorithm to randomly generate M solutions for key design parameters;

[0021] Step 202: Determine the current set of better solutions among the M randomly generated solutions. For each solution X in the current set of better solutions i , calculate the corresponding grasping force F i and bending angle θ i , and the calculated grasping force F i and bending angle θ i As the fitness value of the whale individual in the multi-objective space;

[0022] Step 203: Use the NSGA-II algorithm to perform non-dominated sorting on the M solutions in the current high-quality solution set, and divide the M solutions into different non-dominated levels;

[0023] Step 204: In each iteration, the individual whale generates a random number P and uses the coefficient vector Determine the predation behavior pattern; then update the position of the individual whale solution based on the determined predation behavior pattern;

[0024] The predation behavior patterns include: surrounding prey, spiral update and random search. If P < 0.5 and For the prey encirclement mode, if P < 0.5 and It is a random search mode. If P≥0.5, it executes the spiral update mode.

[0025] The position update formula of the whale algorithm is:

[0026]

[0027] in, is the updated individual position; is the shrinking radius, which decreases linearly with the number of iterations, from 2 to 0, controlling the shrinkage of the search range; l is a random number;

[0028] Step 205: After each iteration from step 202 to step 204, recalculate the fitness values ​​of all whale individuals, update the current set of high-quality solutions, and update the positions of whale individuals again according to the new positions until the maximum number of iterations T is reached. woa , the updated current high-quality solution set is used as the first generation population.

[0029] Furthermore, in step 201, the parameters of the whale algorithm pre-search generation population are set: population size W woa , number of iterations T woa , randomly initialize individual N woa ;

[0030] Using the whale algorithm, M solutions are randomly generated in the search space. Each solution is a whale individual, and each solution is a vector X containing key design parameters. i =[h i ,wi ,r i ,d i ,l i ,ShoreA i ], where i = 1, 2, ...;

[0031] The method of randomly generating solutions:

[0032] For the i-th cavity depth h i , the solution is generated by the formula:

[0033] h i =h min +rand()*(h max -h min )(3)

[0034] For the width w of the i-th cavity i , the solution is generated by the formula:

[0035] w i =w min +rand()*(w max -w min )(4)

[0036] For the radius r of the i-th cavity i , the solution is generated by the formula:

[0037] r i =r min +rand()*(r max -r min )(5)

[0038] For the wall thickness d of the i-th cavity i , the solution is generated by the formula:

[0039] d i =d min +rand()*(d max -d min )(6)

[0040] For the i-th cavity spacing l i , the solution is generated by the formula:

[0041] l i =l min +rand()*(l max -l min )(7)

[0042] For the i-th silicone material hardness ShoreA i , the solution is generated by the formula:

[0043] ShoreAi =ShoreA min +rand()*(ShoreA max -ShoreA min )(8)

[0044] Generate, where rand() is a random number in the interval [0,1].

[0045] Furthermore, step 4 includes the following steps 401 to 404:

[0046] Step 401, using simulated binary crossover to perform a crossover operation on the second generation population;

[0047] Step 402, using polynomial mutation to perform mutation operation on the second generation population;

[0048] Step 403: update the position of each individual whale and check whether the individual whale exceeds the range of the key design parameter; if the position of the individual whale exceeds the range of the key design parameter, adjust the position of the individual whale to the boundary value of the range of the key design parameter;

[0049] Step 404: Processing is performed based on the environment selection mechanism to obtain the third generation population.

[0050] Furthermore, in step 401, the crossover operation adopts simulated binary crossover, including:

[0051] Randomly select two parent individuals parent1 and parent2, and generate the parameters of the offspring individual child for each design parameter dimension;

[0052] The cavity depth parameter h of the offspring individual is generated by the following formula:

[0053]

[0054] The cavity width parameter w of the offspring individual is generated by the following formula:

[0055]

[0056] The cavity radius parameter r of the offspring individual is generated by the following formula:

[0057]

[0058] The cavity wall thickness parameter d of the offspring individual parameter is generated by the following formula:

[0059]

[0060] The cavity spacing parameter l of the offspring individual is generated by the following formula:

[0061]

[0062] The silicone material hardness parameter Shore A of the offspring individual is generated by the following formula:

[0063]

[0064] Among them, β is dynamically adjusted according to the random number.

[0065] Furthermore, in step 402, the mutation operation is performed on the individual to be mutated, if P<0.5 and The individual moves towards the target prey position Close to, such as:

[0066]

[0067] like Then randomly select an individual from the population for search update; if P ≥ 0.5, perform spiral update:

[0068]

[0069] Among them, l is a random number and b is a constant.

[0070] Furthermore, step 404 includes the following steps 4041 to 4044:

[0071] Step 4041, merging the second generation population with the population subjected to the crossover and mutation operations on the second generation population to obtain a merged population;

[0072] Step 4042, using the NSGA-II algorithm to perform non-dominated sorting on the merged population again;

[0073] Step 4043, calculate the congestion degree;

[0074] For the first objective function f1 and the second objective function f2, let n be the population size and the crowding degree of whale individual i distance(i) The calculation formula is:

[0075]

[0076] in, and are the maximum and minimum values ​​of the jth objective function in the current level respectively;

[0077] Step 4044, select the top N nsga The whale individuals are regarded as the third generation population, and the non-dominated sorting and crowding degree calculation are performed on the final third generation population.

[0078] The present invention provides an optimization design method for an Ω-shaped cavity soft gripper, which has the following advantages over the prior art:

[0079] 1. Enhanced local search capability: The whale algorithm's position update strategy (such as spiral search and adaptive shrinkage factor) improves the directionality of mutation operations, enhances local search capabilities, and avoids blind perturbations;

[0080] 2. Optimize the quality of the initial population: The first generation of population with higher quality generated by pre-search based on the whale algorithm provides a better starting point for the NSGA-II algorithm and accelerates convergence;

[0081] 3. Balancing global and local search: Using the Whale Algorithm (WOA), dynamically switching between encircling prey and random search modes, achieving a balance between exploration and exploitation;

[0082] 4. Improve the distribution of solution sets: Combined with the congestion maintenance mechanism of the NSGA-II algorithm, ensure the uniformity of the Pareto front solution set.

[0083] 5. The present invention takes maximizing the gripping force as the first objective function and maximizing the bending angle as the second objective function, and can achieve a multi-objective balance between the gripping force and the bending angle; it effectively improves the comprehensive performance of the Ω-shaped cavity soft gripper, and greatly improves the gripping force and the bending angle. The Ω-shaped cavity soft gripper designed by the present invention takes into account the coupling of multiple physical fields (air pressure drive, nonlinear deformation of materials, contact mechanics), while taking into account multiple conflicting objectives such as gripping force and bending angle. The method of the present invention can be used to conveniently adjust and manufacture the Ω-shaped cavity soft gripper, thereby improving efficiency. By integrating the advantages of the whale algorithm and NSGA-II, the present invention significantly improves the efficiency of solving multi-objective optimization problems and the quality of solutions, and is suitable for complex engineering optimization scenarios.

[0084] 6. The present invention can be extended to other continuous parameter optimization problems, including but not limited to deep learning model hyperparameter optimization, robot path planning, and engineering structure design. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1(a) is a schematic diagram of the structure of the Ω-shaped cavity soft gripper;

[0086] Figure 1(b) is a cross-sectional schematic diagram of the Ω-shaped cavity soft gripper;

[0087] Figure 2 This is a flow chart for implementing the optimization design method of the Ω-shaped cavity soft gripper provided by the present invention;

[0088] Figure 3 This is a detailed flow chart of the optimization design method of the Ω-shaped cavity soft gripper provided by the present invention;

[0089] Figure 4It is a detailed flow chart of the first generation population for generating key design parameters in step 2 of the present invention;

[0090] Figure 5 This is a detailed flow chart of the crossover operation and mutation operation performed on the second generation population in step 4 of the present invention. DETAILED DESCRIPTION

[0091] To make the technical problems solved, the technical solutions adopted, and the technical effects achieved by the present invention more clearly apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present invention, rather than all of the contents.

[0092] like Figure 2-3 As shown, an embodiment of the present invention provides an optimization design method for an Ω-shaped cavity soft gripper, which includes the following steps:

[0093] Step 1: Clarify the key design parameters and the range of key design parameters of the Ω-shaped cavity soft gripper, and establish the objective function.

[0094] The key design parameters include: cavity depth h, cavity width w, cavity radius r, cavity wall thickness d, cavity spacing l, and silicone material hardness Shore A.

[0095] The ranges of the key design parameters include: cavity depth h ranges from 10-30 mm, cavity width w ranges from 5-20 mm, cavity radius r ranges from 10-25 mm, cavity wall thickness d ranges from 1-3 mm, cavity spacing l ranges from 20-55 mm, and silicone material hardness Shore A ranges from 20-40.

[0096] The objective function includes: maximizing the grasping force F as a first objective function f1, and maximizing the bending angle θ as a second objective function f2.

[0097] The calculation formula of the gripping force F is:

[0098]

[0099] Where μ is the friction coefficient (usually μ = 0.4 to 0.7); k F is the contact area correction factor (k F =0.5~0.8); β is the high-order nonlinear coefficient (β≈0.05~0.1); C1 is the first hyperelastic material constant, C2 is the second hyperelastic material constant; p is the air pressure; E eq is the equivalent elastic modulus, E eq ≈6C1.

[0100] The calculation formula of the bending angle θ is:

[0101]

[0102] Where α is the nonlinear correction coefficient (usually α≈0.1~0.3); k θ is the multi-cavity geometry correction factor (usually k θ ≈0.6~0.9).

[0103] Step 2: Generate the first generation population of key design parameters based on the whale algorithm. Specifically, Figure 4 As shown, step 2 includes the following steps 201 to 205:

[0104] Step 201: Use the whale algorithm to randomly generate M solutions for key design parameters.

[0105] Specifically, set the parameters of the whale algorithm pre-search to generate the population: population size W woa , number of iterations T woa , randomly initialize individual N woa .

[0106] Using the whale algorithm, M solutions are randomly generated in the search space. Each solution is a whale individual, and each solution is a vector X containing key design parameters. i =[h i ,w i ,r i ,d i ,l i ,ShoreA i ], where (i=1,2,……).

[0107] Here is a way to randomly generate solutions:

[0108] For the i-th cavity depth h i , the solution is generated by the formula:

[0109] h i =h min +rand()*(h max -H min ) (3)

[0110] For the width w of the i-th cavity i , the solution is generated by the formula:

[0111] w i =w min +rand()*(w max -w min ) (4)

[0112] For the radius r of the i-th cavity i , the solution is generated by the formula:

[0113] r i =r min +rand()*(r max -r min ) (5)

[0114] For the wall thickness d of the i-th cavity i , the solution is generated by the formula:

[0115] d i =d min +rand()*(d max -d min ) (6)

[0116] For the i-th cavity spacing l i , the solution is generated by the formula:

[0117] l i =l min +rand()*(l max -l min ) (7)

[0118] For the i-th silicone material hardness ShoreA i , the solution is generated by the formula:

[0119] ShoreA i =ShoreA min +rand()*(ShoreA max -ShoreA min ) (8)

[0120] Generate, where rand() is a random number in the interval [0,1].

[0121] Step 202: Determine the current set of better solutions among the M randomly generated solutions. For each solution X in the current set of better solutions i , calculate the corresponding grasping force F i and bending angle θ i , and the calculated grasping force F i and bending angle θ i As the fitness value of the whale individual in the multi-objective space.

[0122] Step 203 : Using the NSGA-II algorithm (Non-dominated Sorting Genetic Algorithm II), the M solutions in the current high-quality solution set are non-dominated sorted, and the M solutions are divided into different non-dominated levels.

[0123] The non-dominated sorting rule is: if a solution X a Better than or equal to another solution X in both the first and second objective functions b , and is better than X in at least one objective function b , then X a DominateX b Solutions that are not dominated by any other solution belong to the first non-dominated level F1. Next, remove the individuals in the first non-dominated level from the population and search for individuals that are not dominated by other individuals among the remaining individuals, classifying them as the second non-dominated level F2. Repeat these steps until all individuals have been classified into a non-dominated level.

[0124] Step 204: In each iteration, the individual whale generates a random number P and uses the coefficient vector Determine the predation behavior pattern; then update the position of the individual whale solution based on the determined predation behavior pattern;

[0125] Among them, the random number P is randomly generated by the algorithm.

[0126] The predation behavior patterns include: surrounding prey, spiral update and random search. If P < 0.5 and For the prey encirclement mode, if P < 0.5 and It is a random search mode. If P≥0.5, it executes the spiral update mode.

[0127] The position update formula of the whale algorithm is:

[0128]

[0129] in, is the updated individual position; is the shrinking radius, which decreases linearly with the number of iterations, from 2 to 0, controlling the shrinkage of the search range; l is a random number.

[0130] Step 205: After each iteration from step 202 to step 204, recalculate the fitness values ​​of all whale individuals, update the current high-quality solution set (target prey position), and update the position of the whale individual again according to the new position until the maximum number of iterations T is reached. woa , the updated current high-quality solution set is used as the first generation population.

[0131] Step 3, based on the NSGA-II algorithm (Non-dominated Sorting Genetic Algorithm II), randomly generate N from the first generation population. nsga -k auxiliary solutions; then merge the auxiliary solutions with the first generation population to obtain the second generation population.

[0132] Step 4: Perform crossover and mutation operations on the second-generation population, and then process it based on the environmental selection mechanism to obtain the third-generation population.

[0133] like Figure 5 As shown, step 4 includes the following steps 401 to 404:

[0134] Step 401 : Perform a crossover operation on the second generation population using simulated binary crossover.

[0135] The crossover operation uses simulated binary crossover (SBX) including: randomly selecting two parent individuals parent1 and parent2, and generating the parameters of the offspring individual child for each design parameter dimension.

[0136] The following is how to generate offspring individuals:

[0137] The cavity depth parameter h of the offspring individual is generated by the following formula:

[0138]

[0139] The cavity width parameter w of the offspring individual is generated by the following formula:

[0140]

[0141] The cavity radius parameter r of the offspring individual is generated by the following formula:

[0142]

[0143] The cavity wall thickness parameter d of the offspring individual parameter is generated by the following formula:

[0144]

[0145] The cavity spacing parameter l of the offspring individual is generated by the following formula:

[0146]

[0147] The silicone material hardness parameter Shore A of the offspring individual is generated by the following formula:

[0148]

[0149] Among them, β is dynamically adjusted according to the random number to control the degree of crossover.

[0150] Step 402: Use polynomial mutation to perform mutation operation on the second generation population.

[0151] The mutation operation is used for individuals that need to be mutated. If P < 0.5 and The individual moves towards the target prey position Close to, such as:

[0152]

[0153] like Then randomly select an individual from the population for search update; if P ≥ 0.5, perform spiral update:

[0154]

[0155] Among them, l is a random number and b is a constant.

[0156] Step 403: Update the position of each whale (solution) and check whether the whale exceeds the range of the key design parameter. If the whale's position exceeds the range of the key design parameter, adjust the whale's position to the boundary value of the range of the key design parameter (the range of the parameter in the solution set).

[0157] For example, if h i <h min , then let h i =h min ; if h i >h max , then let h i >h max , and finally complete the mutation.

[0158] Step 404: Processing is performed based on the environmental selection mechanism to obtain the third generation population;

[0159] Step 404 (environment selection mechanism) includes the following steps 4041 to 4044:

[0160] Step 4041, merging the second generation population with the population subjected to the crossover and mutation operations on the second generation population to obtain a merged population;

[0161] Step 4042, using the NSGA-II algorithm to perform non-dominated sorting on the merged population again;

[0162] The non-dominated sorting method of NSGA-II is used to perform non-dominated sorting on the merged population; the merged population is divided into different non-dominated levels F1, F2...

[0163] Step 4043, calculate the congestion degree.

[0164] In each level, the crowding of individuals is calculated. For the first objective function f1 and the second objective function f2, let n be the population size, and the crowding of whale individual i is distance(i) The calculation formula is:

[0165]

[0166] in, and are the maximum and minimum values ​​of the j-th objective function in the current level, respectively.

[0167] Step 4044, select the top N nsga The whale individuals are regarded as the third generation population, and the non-dominated sorting and crowding degree calculation are performed on the final third generation population.

[0168] Step 5: When the maximum number of iterations T (T is the number of iterations of the non-dominated sorting genetic algorithm II) is reached or the solution set converges (such as the optimal solution does not change significantly over multiple generations), the Pareto optimal solution set containing the grasping force and bending angle is output.

[0169] The present invention uses the whale algorithm to pre-search and generate a high-quality initial population (first-generation population), and selects the optimal solution (parameter combination) through non-dominated sorting and crowding. In the iterative optimization stage, the current optimal solution is determined as the target prey position of the whale algorithm through non-dominated sorting and crowding calculation, the parameter a is dynamically adjusted (linearly decreasing from 2 to 0), and different predation modes are selected according to the probability P: when P<0.5, encirclement of prey or random search is performed, and the absolute value of the coefficient A is used to judge whether to shrink to the optimal solution or randomly explore; when P≥0.5, spiral update is performed to simulate the spiral motion trajectory of the whale around the prey. After the mutation operation, the offspring population is generated by combining simulated binary crossover, and the parent and offspring populations (second-generation population) are merged for non-dominated sorting. High-level individuals are given priority and the diversity of the solution set is retained according to the crowding (third-generation population). Individual performance is verified by ABAQUS static simulation, target values ​​such as grasping force and bending angle are extracted, and contact stability and stress safety constraints are embedded. The algorithm continues to iterate until the termination condition (such as the maximum number of iterations) is met, and finally outputs the Pareto optimal solution set including the grasping force and bending angle.

[0170] This method effectively balances global search and local exploitation, achieving a 33% faster convergence rate and an 8% higher solution diversity than traditional methods, enabling comprehensive optimization of the multi-objective performance of soft grippers. The method is applicable to the field of grid-like soft actuator grippers and provides an efficient solution for parameter optimization of soft robots.

[0171] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications to the technical solutions described in the above embodiments, or equivalent replacement of some or all of the technical features therein, do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the design of an Ω-shaped cavity soft gripper, characterized in that: The following processes are included: Step 1: Clarify the key design parameters and their ranges of the Ω-shaped cavity soft gripper and establish the objective function; Step 2: Generate the first generation population of key design parameters based on the whale algorithm; Step 3: Based on the NSGA-II algorithm, N is randomly generated from the first generation population. nsga -k auxiliary solutions; then merge the auxiliary solutions with the first generation population to obtain the second generation population; Step 4: Perform crossover and mutation operations on the second generation population, and then process it based on the environmental selection mechanism to obtain the third generation population; Step 5: When the maximum number of iterations T is reached or the solution set converges, the Pareto optimal solution set including the grasping force and bending angle is output.

2. The optimization design method of the Ω-shaped cavity soft gripper according to claim 1 is characterized in that: The key design parameters include: cavity depth h, cavity width w, cavity radius r, cavity wall thickness d, cavity spacing l, and silicone material hardness Shore A.

3. The optimization design method of the Ω-shaped cavity soft gripper according to claim 2 is characterized in that: The ranges of the key design parameters include: cavity depth h ranges from 10-30 mm, cavity width w ranges from 5-20 mm, cavity radius r ranges from 10-25 mm, cavity wall thickness d ranges from 1-3 mm, cavity spacing l ranges from 20-55 mm, and silicone material hardness Shore A ranges from 20-40.

4. The optimization design method of the Ω-shaped cavity soft gripper according to claim 3 is characterized in that: The objective function includes: maximizing the grasping force F as a first objective function f1, and maximizing the bending angle θ as a second objective function f2; Where μ is the friction coefficient; k F is the contact area correction factor; β is the high-order nonlinear coefficient; C1 is the first hyperelastic material constant, C2 is the second hyperelastic material constant; p is the air pressure; E eq is the equivalent elastic modulus; The calculation formula of the bending angle θ is: Among them, α is the nonlinear correction coefficient; k θ is the multi-cavity geometry correction factor.

5. The optimization design method of the Ω-shaped cavity soft gripper according to claim 1 is characterized in that: Step 2 includes the following steps 201 to 205: Step 201: using the whale algorithm to randomly generate M solutions for key design parameters; Step 202: Determine the current set of better solutions among the M randomly generated solutions. For each solution X in the current set of better solutions i , calculate the corresponding grasping force F i and bending angle θ i , and the calculated grasping force F i and bending angle θ i As the fitness value of the whale individual in the multi-objective space; Step 203: Use the NSGA-II algorithm to perform non-dominated sorting on the M solutions in the current high-quality solution set, and divide the M solutions into different non-dominated levels; Step 204: In each iteration, the individual whale generates a random number P and uses the coefficient vector Identify patterns of predatory behavior; Then update the position of the individual whale solution according to the determined predation behavior pattern; The predation behavior patterns include: surrounding prey, spiral update and random search. If P < 0.5 and For the prey encirclement mode, if P < 0.5 and It is a random search mode. If P≥0.5, it executes the spiral update mode. The position update formula of the whale algorithm is: in, is the updated individual position; is the shrinking radius, which decreases linearly with the number of iterations, from 2 to 0, controlling the shrinkage of the search range; l is a random number; Step 205: After each iteration from step 202 to step 204, recalculate the fitness values ​​of all whale individuals, update the current set of high-quality solutions, and update the positions of whale individuals again according to the new positions until the maximum number of iterations T is reached. woa , the updated current high-quality solution set is used as the first generation population.

6. The optimization design method of the Ω-shaped cavity soft gripper according to claim 5, characterized in that: In step 201, set the parameters of the whale algorithm pre-search to generate the population: population size W woa , number of iterations T woa , randomly initialize individual N woa ; Using the whale algorithm, M solutions are randomly generated in the search space. Each solution is a whale individual, and each solution is a vector X containing key design parameters. i =[h i ,w i ,r i ,d i ,l i ,ShoreA i ], where i = 1, 2, ...; The method of randomly generating solutions: For the i-th cavity depth h i , the solution is generated by the formula: h i =h min +rand()*(h max -h min )(3) For the width w of the i-th cavity i , the solution is generated by the formula: w i =w min +rand()*(w max -w min )(4) For the radius r of the i-th cavity i , the solution is generated by the formula: r i =r min +rand()*(r max -r min )(5) For the wall thickness d of the i-th cavity i , the solution is generated by the formula: d i =d min +rand()*(d max -d min )(6) For the i-th cavity spacing l i , the solution is generated by the formula: L i =l min +rand()*(l max -L min )(7) For the i-th silicone material hardness ShoreA i , the solution is generated by the formula: ShoreA i =ShoreA min +rand()*(ShoreA max -ShoreA min )(8) Generate, where rand() is a random number in the interval [0,1].

7. The optimization design method of the Ω-shaped cavity soft gripper according to claim 1 is characterized in that: Step 4 includes the following steps 401 to 404: Step 401, using simulated binary crossover to perform a crossover operation on the second generation population; Step 402, using polynomial mutation to perform mutation operation on the second generation population; Step 403: update the position of each individual whale and check whether the individual whale exceeds the range of the key design parameter; if the position of the individual whale exceeds the range of the key design parameter, adjust the position of the individual whale to the boundary value of the range of the key design parameter; Step 404: Processing is performed based on the environment selection mechanism to obtain the third generation population.

8. The optimization design method of the Ω-shaped cavity soft gripper according to claim 7, characterized in that: In step 401, the crossover operation adopts simulated binary crossover, including: Randomly select two parent individuals parent1 and parent2, and generate the parameters of the offspring individual child for each design parameter dimension; The cavity depth parameter h of the offspring individual is generated by the following formula: The cavity width parameter w of the offspring individual is generated by the following formula: The cavity radius parameter r of the offspring individual is generated by the following formula: The cavity wall thickness parameter d of the offspring individual parameter is generated by the following formula: The cavity spacing parameter l of the offspring individual is generated by the following formula: The silicone material hardness parameter Shore A of the offspring individual is generated by the following formula: Among them, β is dynamically adjusted according to the random number.

9. The optimization design method of the Ω-shaped cavity soft gripper according to claim 8, characterized in that: In step 402, the mutation operation is performed on the individual to be mutated, if P < 0.5 and The individual moves towards the target prey position Close to, such as: like Then randomly select an individual from the population for search update; if P ≥ 0.5, perform spiral update: Among them, l is a random number and b is a constant.

10. The optimization design method of the Ω-shaped cavity soft gripper according to claim 9, characterized in that: Step 404 includes the following steps 4041 to 4044: Step 4041, merging the second generation population with the population subjected to the crossover and mutation operations on the second generation population to obtain a merged population; Step 4042, using the NSGA-II algorithm to perform non-dominated sorting on the merged population again; Step 4043, calculate the congestion degree; For the first objective function f1 and the second objective function f2, let n be the population size and the crowding degree of whale individual i distance(i) The calculation formula is: in, and are the maximum and minimum values ​​of the jth objective function in the current level respectively; Step 4044, select the top N nsga The whale individuals are regarded as the third generation population, and the non-dominated sorting and crowding degree calculation are performed on the final third generation population.

Images