Service robot system for modular function extension and dynamic task scheduling

Abstract

The invention discloses a modular function extension and dynamic task scheduling service robot system, which relates to the technical field of robots, and comprises the following specific steps: step 1, task feature extraction, step 2, machine learning prediction, step 3, dynamic priority calculation, step 4, heuristic task allocation, and step 5, load balancing scheduling. And step 6, task execution and feedback, in the method, the execution time and the resource demand of the task are predicted by using a machine learning model, and an optimal task scheduling scheme is calculated by using dynamic priority calculation, heuristic task allocation and load balancing scheduling. In the first step, tasks are fragmented, then pass through calculation and then are combined, the calculation complexity can be reduced, and the algorithm operation efficiency can be improved. In the fourth step, multiple iterations are carried out after randomization operation, and the scheduling scheme is optimized step by step, so that the global optimality of the scheduling scheme is improved, and falling into a local optimal solution is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of robots, and particularly to a service robot system with modular function expansion and dynamic task scheduling. Background Art

[0002] A robot is an intelligent machine that can work semi-autonomously or fully autonomously. It can execute various tasks, such as operations and movements, through programming and automatic control. A robot has basic characteristics such as perception, decision-making, and execution, and can assist or even replace humans to complete dangerous, heavy, and complex work, thereby improving work efficiency and quality, serving human life, and expanding or extending the scope of human activities and capabilities.

[0003] In current service robots, when dynamically scheduling tasks, only task priorities are considered, which brings certain limitations to the task scheduling of service robots and affects the practicability of service robots.

[0004] In summary, it is necessary to design a service robot system with modular function expansion and dynamic task scheduling. Summary of the Invention

[0005] In order to overcome the above deficiencies, the present invention provides a service robot system with modular function expansion and dynamic task scheduling.

[0006] The present invention achieves the above object through the following technical solutions:

[0007] A service robot system with modular function expansion and dynamic task scheduling includes the following specific steps:

[0008] Step 1, task feature extraction: Collect information such as the execution time, resource requirements, and priority of tasks. The task set T = {T1, T2,..., T n}, and each task T i includes a feature vector X i = (x1, x2,..., x m ), and the feature vector X i includes task type, resource requirements, and historical execution time;

[0009] Step 2, machine learning prediction: Use a machine learning model to predict the execution time and resource requirements of tasks. The formula is where is the predicted execution time of task T i , and f is a machine learning model (such as linear regression, random forest, neural network, etc.);

[0010] Step 3, dynamic priority calculation: Combine task priority, predicted execution time, and system load to dynamically adjust task priority;

[0011] Step 4, Heuristic Task Allocation: Optimize the task allocation scheme through a heuristic algorithm. The heuristic algorithm is as follows:

[0012]

[0013] where S opt is the optimal task allocation scheme, P i is the dynamic priority of task T i C i is the execution cost of task T i , Load_Balance(S) is the load balance degree of the scheduling scheme S, and λ is the load balance weight parameter;

[0014] Step 5, Load Balancing Scheduling: According to the node load situation, allocate tasks to the node with the lowest load to ensure the maximization of resource utilization;

[0015] Step 6, Task Execution and Feedback: Execute the task and collect the actual execution data for optimizing the machine learning model and the heuristic algorithm.

[0016] Preferably, the said Step 1 includes the following specific steps:

[0017] S11, Task Sharding: Divide the task set T = {T1, T2,..., T n} into multiple subsets T k , and allocate each subset to a different computing node. T k = {T i |i ∈ Node k}, where T k is the task subset allocated to node k, and Node k is the k-th computing node;

[0018] S12, Parallel Computing: Each node independently executes task scheduling, and uses a multi-core CPU or GPU to accelerate the calculation. The scheduling formula is where Schedule k is the local scheduling scheme of node k, P i is the dynamic priority of task T i C i is the execution cost of task T i ;

[0019] S13, Result Merging: Merge the local scheduling schemes of each node into a global scheduling scheme. The merging formula is where Schedule global is the global scheduling scheme, m is the total number of computing nodes, which reduces the calculation complexity and improves the algorithm operation efficiency.

[0020] Preferably, the formula for calculating the dynamic priority in step three is as follows:

[0021]

[0022] Where P i is the dynamic priority of task T i , is the predicted execution time of task T i , R i is the static priority (user-defined) of task T i , L j is the current load of node j, and α, β, γ are weight coefficients used to adjust the influence of various factors.

[0023] Preferably, a randomization operation is added in step four to increase the search space. The formula is as follows

[0024] S new = Mutate(S current , Random_Factor)

[0025] Where Mutate is a randomization operation function used to perturb the current scheduling scheme, S new is the new scheduling scheme, S current is the current scheduling scheme, and Random_Factor is a randomization factor. After the randomization operation, multiple iterations are performed to gradually optimize the scheduling scheme, thereby improving the global optimality of the scheduling scheme and avoiding falling into a local optimal solution.

[0026] Preferably, in steps three and four, a weight parameter tuning tool is added to improve the adaptability and performance of the scheduling scheme by simplifying the parameter tuning process. The specific steps are as follows:

[0027] S41. An automatic parameter tuning tool (using Bayesian optimization to search for the optimal weight parameters), the formula is as follows:

[0028]

[0029] Where α * , β * , γ * , λ * are the optimal weight parameters, and P(S) is the performance index of the scheduling scheme S (such as task completion time, resource utilization rate);

[0030] S42. Adaptive parameter adjustment, dynamically adjusting the weight parameters according to the system state and task requirements. The formula is as follows:

[0031] α t = α t-1 + ηΔα

[0032] where α t is the weight parameter for the t-th iteration, η is the learning rate, and Δα is the parameter adjustment amount;

[0033] S43. Feedback mechanism, optimizing the weight parameter through actual execution data

[0034]

[0035] where Δα is the parameter adjustment amount.

[0036] Preferably, the formula for step five is as follows:

[0037]

[0038] where Node next is the node for the next task assignment, and L j is the current load of node j.

[0039] Preferably, the formula for step six is as follows:

[0040]

[0041] where is the predicted execution time of task T i , Error i is the prediction error of task T i , and T i is the actual execution time of task T i . Thus, by collecting actual execution data, the machine learning model and heuristic algorithm are optimized to improve prediction accuracy and scheduling efficiency.

[0042] Preferably, it includes the following main modules:

[0043] Perception module: responsible for environmental perception and data collection;

[0044] Decision module: based on artificial intelligence algorithms for task planning and decision-making;

[0045] Execution module: controls the robot to execute specific tasks;

[0046] Communication module: realizes data interaction and collaborative work between the robot and external systems;

[0047] Energy module: manages the energy supply and consumption of the robot to ensure long-term stable operation.

[0048] The beneficial effects of the present invention are as follows: In the service robot system with modular function expansion and dynamic task scheduling:

[0049] 1. Predict the execution time and resource requirements of tasks using a machine learning model, and at the same time, calculate the optimal task scheduling scheme by using dynamic priority calculation, heuristic task allocation, and load balancing scheduling.

[0050] 2. In step one, by fragmenting the tasks and then performing calculation and merging, the computational complexity can be reduced and the algorithm running efficiency can be improved.

[0051] 3. After randomization operations in step four, perform multiple iterations to gradually optimize the scheduling scheme, thereby improving the global optimality of the scheduling scheme and avoiding falling into local optimal solutions.

[0052] 4. In steps three and four, add a weight parameter tuning tool to improve the adaptability and performance of the scheduling scheme by simplifying the parameter tuning process. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The present invention will be described by way of examples with reference to the accompanying drawings, where:

[0054] Figure 1 is the step diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] Now, the present invention will be further described in detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0056] As Figure 1 shown, a service robot system for modular function expansion and dynamic task scheduling includes the following specific steps:

[0057] Step 1. Task feature extraction: Collect information such as the execution time, resource requirements, and priority of tasks. The task set T = {T1, T2,..., T n}, and each task T i includes a feature vector X i = (x1, x2,..., x m ), and the feature vector X i includes task type, resource requirements, and historical execution time;

[0058] Step 2. Machine learning prediction: Use a machine learning model to predict the execution time and resource requirements of tasks. The formula is where is the predicted execution time of task T i , and f is a machine learning model (such as linear regression, random forest, neural network, etc.);

[0059] Step 3. Dynamic priority calculation: Combine task priority, predicted execution time, and system load to dynamically adjust task priority;

[0060] Step 4, Heuristic Task Allocation: Optimize the task allocation scheme through a heuristic algorithm. The heuristic algorithm is as follows:

[0061]

[0062] Among them, S opt is the optimal task allocation scheme, P i is the dynamic priority of task T i C i is the execution cost of task T i Load_Balance(S) is the load balance degree of the scheduling scheme S, and λ is the load balance weight parameter;

[0063] Step 5, Load Balancing Scheduling: According to the node load situation, allocate tasks to the node with the lowest load to ensure the maximization of resource utilization;

[0064] Step 6, Task Execution and Feedback: Execute tasks and collect actual execution data for optimizing the machine learning model and heuristic algorithm.

[0065] Specifically, the said Step 1 includes the following specific steps:

[0066] S11, Task Sharding: Divide the task set T = {T1, T2,..., T n} into multiple subsets T k , and allocate each subset to a different computing node. T k = {T i |i ∈ Node k}, where T k is the task subset allocated to node k, and Node k is the k-th computing node;

[0067] S12, Parallel Computing: Each node independently executes task scheduling, and uses multi-core CPU or GPU to accelerate the computing. The scheduling formula is Among them, Schedule k is the local scheduling scheme of node k, P i is the dynamic priority of task T i C i is the execution cost of task T i ;

[0068] S13, Result Merging: Merge the local scheduling schemes of each node into a global scheduling scheme. The merging formula is Among them, Schedule global is the global scheduling scheme, m is the total number of computing nodes, reducing the computing complexity and improving the algorithm operation efficiency.

[0069] Specifically, the formula for dynamic priority calculation in step 3 is as follows:

[0070]

[0071] where P i is the dynamic priority of task T i , is the predicted execution time of task T i , R i is the static priority (user-defined) of task T i , L j is the current load of node j, and α, β, γ are weight coefficients used to adjust the influence of various factors.

[0072] Specifically, a randomization operation is added in step 4 to increase the search space, and the formula is as follows

[0073] S new = Mutate(S current , Random_Factor)

[0074] where Mutate is a randomization operation function used to perturb the current scheduling scheme, S new is the new scheduling scheme, S current is the current scheduling scheme, and Random_Factor is a randomization factor. After the randomization operation, multiple iterations are performed to gradually optimize the scheduling scheme, thereby improving the global optimality of the scheduling scheme and avoiding falling into a local optimal solution.

[0075] Specifically, in steps 3 and 4, a weight parameter tuning tool is added to improve the adaptability and performance of the scheduling scheme by simplifying the parameter tuning process. The specific steps are as follows:

[0076] S41. An automatic parameter tuning tool (using Bayesian optimization to search for the optimal weight parameters), and the formula is as follows:

[0077]

[0078] where α * , β * , γ * , λ * are the optimal weight parameters, and P(S) is the performance index of the scheduling scheme S (such as task completion time, resource utilization rate);

[0079] S42. Adaptive parameter adjustment, dynamically adjusting the weight parameters according to the system state and task requirements, and the formula is as follows:

[0080] α t = α t-1+ηΔα

[0081] where α t is the weight parameter for the t-th iteration, η is the learning rate, and Δη is the parameter adjustment amount;

[0082] S43. Feedback mechanism, which optimizes the weight parameter through the actual execution data

[0083]

[0084] where Δη is the parameter adjustment amount.

[0085] Specifically, the formula for step five is as follows:

[0086]

[0087] where Node next is the node for the next task assignment, and L j is the current load of node j.

[0088] Specifically, the formula for step six is as follows:

[0089]

[0090] where is the predicted execution time of task T i Error i is the prediction error of task T i T i is the actual execution time of task T i Thus, by collecting the actual execution data, the machine learning model and the heuristic algorithm are optimized to improve the prediction accuracy and the scheduling efficiency.

[0091] Specifically, it includes the following main modules:

[0092] Perception module: responsible for environmental perception and data collection;

[0093] Decision module: performs task planning and decision-making based on artificial intelligence algorithms;

[0094] Execution module: controls the robot to execute specific tasks;

[0095] Communication module: realizes data interaction and collaborative work between the robot and the external system;

[0096] Energy module: manages the energy supply and consumption of the robot to ensure long-term stable operation.

[0097] Embodiment 1: Application of a service robot system with modular function expansion and dynamic task scheduling in logistics warehousing

[0098] Step 1: Task Feature Extraction

[0099] 1. Task Sharding: Divide logistics warehousing tasks (such as goods sorting, inventory management, transportation scheduling) into multiple subsets, and each subset is assigned to a different computing node. For example, assign the goods sorting task to node A and the inventory management task to node B.

[0100] 2. Parallel Computing: Each node independently executes task scheduling and utilizes multi-core CPUs or GPUs to accelerate the calculation. For example, node A uses a GPU to accelerate the scheduling calculation of the sorting task.

[0101] 3. Result Merging: Merge the local scheduling schemes of each node into a global scheduling scheme. For example, merge the scheduling schemes of node A and node B to form a global logistics scheduling scheme.

[0102] Step 2: Machine Learning Prediction

[0103] 1. Model Training: Use historical logistics data to train a machine learning model to predict the execution time and resource requirements of tasks. For example, use a random forest model to predict the execution time of the goods sorting task.

[0104] 2. Prediction Application: Apply the prediction results to task scheduling to optimize resource allocation. For example, adjust the priority of the sorting task according to the prediction results.

[0105] Step 3: Dynamic Priority Calculation

[0106] 1. Priority Adjustment: Dynamically adjust task priorities by combining task priorities, predicted execution times, and system loads. For example, when the system load is high, increase the priority of the inventory management task.

[0107] 2. Weight Coefficient Adjustment: Dynamically adjust weight coefficients according to the actual operating conditions to optimize priority calculation. For example, increase the weight of the system load and decrease the weight of the predicted execution time.

[0108] Step 4: Heuristic Task Allocation

[0109] 1. Application of Heuristic Algorithm: Optimize the task allocation scheme through a heuristic algorithm. For example, use a genetic algorithm to optimize the allocation of goods transportation tasks.

[0110] 2. Randomization Operation: Add randomization operations to increase the search space and avoid falling into local optimal solutions. For example, randomly adjust the allocation order of transportation tasks.

[0111] Step 5: Load Balancing Scheduling

[0112] 1. Load Monitoring: Real-time monitor the load conditions of each node. For example, monitor the CPU and memory usage of node A and node B.

[0113] 2. Task Assignment: Assign tasks to the node with the lowest load to ensure maximum resource utilization. For example, assign a new transportation task to node B with a lower load.

[0114] Step Six: Task Execution and Feedback

[0115] 1. Task Execution: Execute the task and collect actual execution data. For example, execute the goods sorting task and record the actual execution time.

[0116] 2. Model Optimization: Utilize the actual execution data to optimize machine learning models and heuristic algorithms. For example, adjust the parameters of the random forest model according to the actual execution time.

[0117] Example Two: Application of a Service Robot System with Modular Function Expansion and Dynamic Task Scheduling in Medical Care

[0118] Step One: Task Feature Extraction

[0119] 1. Task Sharding: Divide medical care tasks (such as patient monitoring, medicine delivery, surgical assistance) into multiple subsets, and assign each subset to different computing nodes. For example, assign the patient monitoring task to node C and the medicine delivery task to node D.

[0120] 2. Parallel Computing: Each node independently executes task scheduling and uses multi-core CPUs or GPUs to accelerate the calculation. For example, node C uses the CPU to accelerate the scheduling calculation of the monitoring task.

[0121] 3. Result Merging: Merge the local scheduling schemes of each node into a global scheduling scheme. For example, merge the scheduling schemes of node C and node D to form a global medical scheduling scheme.

[0122] Step Two: Machine Learning Prediction

[0123] 1. Model Training: Utilize historical medical data to train machine learning models to predict the execution time and resource requirements of tasks. For example, use a neural network model to predict the execution time of surgical assistance tasks.

[0124] 2. Prediction Application: Apply the prediction results to task scheduling to optimize resource allocation. For example, adjust the priority of the medicine delivery task according to the prediction results.

[0125] Step Three: Dynamic Priority Calculation

[0126] 1. Priority Adjustment: Dynamically adjust task priorities by combining task priorities, predicted execution times, and system loads. For example, when the system load is high, increase the priority of surgical assistance tasks.

[0127] 2. Weight coefficient adjustment: Dynamically adjust the weight coefficients according to the actual operation situation to optimize the priority calculation. For example, increase the weight of the predicted execution time and reduce the weight of the system load.

[0128] Step Four: Heuristic task allocation

[0129] 1. Application of heuristic algorithm: Optimize the task allocation scheme through heuristic algorithms. For example, use the ant colony algorithm to optimize the allocation of drug delivery tasks.

[0130] 2. Randomization operation: Add randomization operations to increase the search space and avoid falling into local optimal solutions. For example, randomly adjust the allocation order of delivery tasks.

[0131] Step Five: Load balancing scheduling

[0132] 1. Load monitoring: Monitor the load conditions of each node in real time. For example, monitor the CPU and memory usage of Node C and Node D.

[0133] 2. Task allocation: Allocate tasks to the node with the lowest load to ensure maximum utilization of resources. For example, allocate new guardianship tasks to the less-loaded Node C.

[0134] Step Six: Task execution and feedback

[0135] 1. Task execution: Execute tasks and collect actual execution data. For example, execute surgical assistance tasks and record the actual execution time.

[0136] 2. Model optimization: Utilize the actual execution data to optimize machine learning models and heuristic algorithms. For example, adjust the parameters of the neural network model according to the actual execution time.

[0137] Example Three: Application of a service robot system with modular function expansion and dynamic task scheduling in home services

[0138] Step One: Task feature extraction

[0139] 1. Task sharding: Divide home service tasks (such as cleaning, cooking, and security) into multiple subsets, and allocate each subset to different computing nodes. For example, allocate cleaning tasks to Node E and cooking tasks to Node F.

[0140] 2. Parallel computing: Each node independently executes task scheduling and uses multi-core CPUs or GPUs to accelerate the calculation. For example, Node E uses a GPU to accelerate the scheduling calculation of cleaning tasks.

[0141] 3. Result merging: Merge the local scheduling schemes of each node into a global scheduling scheme. For example, merge the scheduling schemes of Node E and Node F to form a global home service scheduling scheme.

[0142] Step 2: Machine Learning Prediction

[0143] 1. Model Training: Use historical home service data to train a machine learning model to predict the execution time and resource requirements of tasks. For example, use a linear regression model to predict the execution time of cooking tasks.

[0144] 2. Prediction Application: Apply the prediction results to task scheduling to optimize resource allocation. For example, adjust the priority of cleaning tasks according to the prediction results.

[0145] Step 3: Dynamic Priority Calculation

[0146] 1. Priority Adjustment: Combine task priority, predicted execution time, and system load to dynamically adjust task priority. For example, when the system load is high, increase the priority of security tasks.

[0147] 2. Weight Coefficient Adjustment: Dynamically adjust the weight coefficients according to the actual running situation to optimize priority calculation. For example, increase the weight of system load and decrease the weight of predicted execution time.

[0148] Step 4: Heuristic Task Allocation

[0149] 1. Application of Heuristic Algorithm: Optimize the task allocation scheme through a heuristic algorithm. For example, use the simulated annealing algorithm to optimize the allocation of cooking tasks.

[0150] 2. Randomization Operation: Add a randomization operation to increase the search space and avoid falling into a local optimal solution. For example, randomly adjust the allocation order of cooking tasks.

[0151] Step 5: Load Balancing Scheduling

[0152] 1. Load Monitoring: Monitor the load conditions of each node in real time. For example, monitor the CPU and memory usage of Node E and Node F.

[0153] 2. Task Allocation: Allocate tasks to the node with the lowest load to ensure maximum resource utilization. For example, allocate a new security task to Node E with a lower load.

[0154] Step 6: Task Execution and Feedback

[0155] 1. Task Execution: Execute the task and collect actual execution data. For example, execute a cooking task and record the actual execution time.

[0156] 2. Model Optimization: Use the actual execution data to optimize the machine learning model and heuristic algorithm. For example, adjust the parameters of the linear regression model according to the actual execution time.

[0157] Through the above three specific embodiments, the specific implementation steps of the service robot system with modular function expansion and dynamic task scheduling in different application scenarios are demonstrated, and the efficient operation of the system and the accuracy of task scheduling can be achieved.

[0158] Based on the inspiration of the present invention, through the above description, relevant staff can completely make various changes and modifications without departing from the technical idea of the present invention. The technical scope of the present invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A service robot system with modular function expansion and dynamic task scheduling, characterized in that: It includes the following specific steps: Step 1. Task feature extraction: Collect the execution time, resource requirements, and priority information of the tasks. The task set T = {T1, T2, …, T n}, and each task T i includes a feature vector X i =(x1, x2, …, x m ), and the feature vector X i includes the task type, resource requirements, and historical execution time; Step 2, Machine learning prediction: Use a machine learning model to predict the execution time and resource requirements of a task. The formula is where is the predicted execution time of task T i , f is the machine learning model; Step 3, Dynamic Priority Calculation: Combine task priority, predicted execution time, and system load to dynamically adjust task priority; Step 4, Heuristic Task Allocation: Optimize the task allocation scheme through a heuristic algorithm. The heuristic algorithm is: Among them, S opt is the optimal task allocation scheme, P i is the dynamic priority of task T i , C i is the execution cost of task T i , Load_Balance(S) is the load balancing degree of the scheduling scheme S, and λ is the load balancing weight parameter; Step 5, Load Balancing Scheduling: According to the node load situation, allocate tasks to the node with the lowest load to ensure maximum resource utilization; Step 6, Task Execution and Feedback: Execute tasks and collect actual execution data for optimizing machine learning models and heuristic algorithms.

2. The service robot system for modular function expansion and dynamic task scheduling according to claim 1, characterized in that: The said Step 1 includes the following specific steps: S11. Task sharding: Divide the task set T = {T1, T2, …, T n} into multiple subsets T k , and assign each subset to a different computing node. T k = {T i | i ∈ Node k}, where T k is the task subset assigned to node k, and Node k is the k-th computing node; S12. Parallel computing, where each node independently executes task scheduling and utilizes multi-core CPUs or GPUs to accelerate the computing. The scheduling formula is where Schedule k is the local scheduling scheme for node k, P i is the dynamic priority of task T i , and C i is the execution cost of task T i . S13. Result merging: Merge the local scheduling schemes of each node into a global scheduling scheme, and the merging formula is where Schedule global is the global scheduling scheme, and m is the total number of computing nodes.

3. The service robot system for modular function expansion and dynamic task scheduling according to claim 1, wherein: The formula for dynamic priority calculation in Step 3 is as follows: Among them, P i is the dynamic priority of task T i , is the predicted execution time of task T i , R i is the static priority of task T i , L j is the current load of node j, and α, β, γ are weight coefficients used to adjust the influence of each factor.

4. The service robot system for modular function expansion and dynamic task scheduling according to claim 1, characterized in that: In Step 4, a randomization operation is added to increase the search space. The formula is as follows S new = Mutate(S current , Random_Factor) Among them, Mutate is a randomization operation function used to perturb the current scheduling plan. S new is the new scheduling plan, S current is the current scheduling plan, and Random_Factor is the randomization factor. After the randomization operation, multiple iterations are carried out to gradually optimize the scheduling plan.

5. The service robot system for modular function expansion and dynamic task scheduling according to claim 3, wherein: In Step 3 and Step 4, a weight parameter tuning tool is added to improve the adaptability and performance of the scheduling scheme by simplifying the parameter tuning process. The specific steps are as follows: S41, Automated Parameter Tuning Tool. The formula is as follows: Among them, α * , β * , γ * , λ * are the optimal weight parameters, and P(S) is the performance index of the scheduling scheme S; S42, Adaptive Parameter Adjustment. Dynamically adjust the weight parameters according to the system state and task requirements. The formula is as follows: α t = α t-1 + ηΔα where α t is the weight parameter for the t-th iteration, η is the learning rate, and Δα is the parameter adjustment amount; S43, Feedback Mechanism. Optimize the weight parameters through actual execution data where Δα is the parameter adjustment amount.

6. The service robot system for modular function expansion and dynamic task scheduling according to claim 1, characterized in that: The formula for Step 5 is as follows: Among them, Node next is the node assigned to the next task, and L j is the current load of node j.

7. The service robot system for modular function expansion and dynamic task scheduling according to claim 1, characterized in that: The formula for Step 6 is as follows: Wherein, is the predicted execution time of task T i , Error i is the prediction error of task T i , T i is the actual execution time of task T i .

8. The service robot system for modular function expansion and dynamic task scheduling according to claim 1, characterized in that: It includes the following main modules: Perception Module: Responsible for environmental perception and data collection; Decision Module: Conduct task planning and decision-making based on artificial intelligence algorithms; Execution Module: Control the robot to execute specific tasks; Communication Module: Realize data interaction and collaborative work between the robot and external systems; Energy Module: Manage the energy supply and consumption of the robot to ensure long-term stable operation.

Images