A method and system for collaborative scheduling of components of a multi-vehicle-adaptive emergency rescue robot

Abstract

This invention relates to the field of rescue robot technology, addressing the problem of existing technologies where emergency medical equipment is idle during critical moments due to a surge in sudden emergency rescue missions, thus affecting rescue efficiency. Specifically, it relates to a collaborative scheduling method and system for multi-vehicle-adaptive emergency rescue robot components. Responding to an emergency rescue trigger signal, the method collects environmental disturbance parameters of the target scene, extracts the electromechanical state parameters of each physical node, performs preprocessing, constructs a heterogeneous resource matrix based on the electromechanical state parameters, and generates a time-series task set based on the medical classification indicators in the emergency rescue trigger signal. This invention achieves coordinated scheduling of multiple mobile vehicles and different emergency rescue components, avoiding the problem of emergency medical equipment being idle during critical moments due to a surge in sudden emergency rescue missions, eliminating the negative impact of traditional static allocation architectures on adaptability to changing rescue needs, reducing the waste of control resources in complex disaster scenarios, and improving the overall rescue response efficiency of multi-vehicle clusters.

Description

Technical Field

[0001] This invention relates to the field of rescue robot technology, and in particular to a method and system for collaborative scheduling of multi-vehicle-adaptive emergency rescue robot components. Background Technology

[0002] Currently, in disaster situations, multi-robot systems are used for emergency response in hazardous environments to replace human rescue efforts and reduce the risk of injury or death to search and rescue personnel. With the evolution of mechatronics technology, the physical form of emergency robots is developing from single-function vehicles to highly modular and heterogeneous designs. However, in practical disaster response engineering implementation, the adaptation of multiple vehicles and the coordinated scheduling of components still face significant technical bottlenecks.

[0003] Existing emergency rescue robot dispatching methods mostly adopt a one-machine-one-controller mode. In the ever-changing disaster scene, this mode is difficult to dynamically coordinate and dispatch multiple mobile vehicles and different modular components according to sudden emergency task needs, which reduces the adaptability to changing rescue needs and easily causes emergency rescue equipment to be idle at critical moments, thus seriously affecting the overall rescue efficiency. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a collaborative scheduling method and system for multi-vehicle-adaptive emergency rescue robot components, which solves the problem of emergency rescue equipment being idle at critical moments due to a surge in sudden emergency rescue tasks, thus affecting rescue efficiency.

[0005] To address the aforementioned technical problems, the present invention provides the following technical solution: a method for collaborative scheduling of multi-vehicle adapted emergency rescue robot components, the method comprising the following steps: S1. In response to the emergency rescue trigger signal, collect the environmental disturbance parameters of the target scene, extract the electromechanical status parameters of each physical node, and perform preprocessing. S2. Construct a heterogeneous resource matrix that characterizes the vehicle's movement constraints and component operation capabilities based on electromechanical state parameters, and generate a time-series task set with spatiotemporal constraints based on medical grading indicators in emergency response trigger signals. S3. Based on the heterogeneous resource matrix and environmental disturbance parameters, construct an anti-disturbance operation mapping model for a physical intelligent agent composed of dynamic coupling of vehicles and components, and initialize the expected completion probability and operation risk prediction for each task in the time-series task set accordingly. S4. Based on the estimated operation risk and guided by minimizing the global operation risk, under the constraint of satisfying the preset probability confidence threshold, the global scheduling node performs task optimization calculation to generate a global allocation instruction, and then sends the global allocation instruction to the corresponding physical agent. S5. Each physical intelligent agent responds to dynamic spatial interference events, triggers negotiation between adjacent nodes in the dynamic communication subnet as needed to resolve spatiotemporal conflicts, runs a local autonomous decision-making model to optimize electromechanical operating parameters, and generates drive signals to control vehicles and components to carry out emergency rescue operations. S6. Collect real-time operational deviation data of the physical agent in emergency rescue operations, correct the anti-disturbance operation mapping model in reverse based on the real-time operational deviation data, respond to physical triggering events such as environmental topology changes or operational risk exceeding limits, interrupt the current control command and re-trigger the task optimization calculation in step S4.

[0006] Furthermore, step S2 specifically includes the following steps: Extract vehicle kinematic parameters and component physical mounting parameters from electromechanical state parameters, and construct a heterogeneous resource matrix characterizing vehicle movement constraints and component operation capabilities through feature splicing; Based on the emergency trigger signal, three-dimensional spatial coordinates and medical classification indicators are extracted, and the medical classification indicators are mapped to time window constraints with rigid cutoff boundaries according to the nonlinear time-series decay rule. The three-dimensional spatial coordinates, time window constraints, and component capability type constraints required to perform the emergency rescue operation are jointly encapsulated to generate a time-series task set with spatiotemporal constraints.

[0007] Furthermore, step S3 specifically includes the following steps: The physical damping coefficient vector is extracted from the environmental disturbance parameters, and an environmental attenuation tensor matching the dimension of the heterogeneous resource matrix is ​​constructed. Through tensor Hadamard product operations, the heterogeneous resource matrix is ​​transformed into an anti-disturbance operation mapping matrix representing a physical agent composed of dynamically coupled vehicles and components. The calculation formula is as follows: In the formula, Represents the anti-disturbance operation mapping matrix; Represents a heterogeneous resource matrix; Represents the vector of physical damping coefficients; This represents the Hadama product operator; The nonlinear mapping function represents the physical space damping attenuation characteristics. Based on the temporal task set, the Euclidean space coordinate deviation and remaining absolute time window of each temporal task are extracted as spatiotemporal feature terms. Using the effective movement speed scalar in the anti-disturbance task mapping matrix as a benchmark, the expected completion probability of the i-th physical agent performing the j-th temporal task is independently quantized using the logistic function. The calculation formula is as follows: In the formula, This represents the expected probability of the i-th physical agent executing the j-th temporal task; This represents the collaborative confidence hyperparameter; Represents a Boolean variable characterizing the physical mounting capability matching constraint of the component; This represents the Euclidean space coordinate deviation of each time-series task; Represents the remaining absolute time window for each time-series task; Represents the effective movement speed scalar; The medical grading indicators in the time-series task set are coupled with the expected completion probability to calculate the operational risk prediction for each time-series task. The calculation formula is as follows: In the formula, This indicates the estimated risk level of the operation. Indicators representing medical grading; Indicates the priority penalty amplification factor; This is a characterization of the inherent damage risk of environmental background noise.

[0008] Furthermore, step S4 specifically includes the following steps: Extract the identifier set and temporal task set of physical agents, define a Boolean matrix, and construct an optimization objective function and probability confidence constraint equation guided by minimizing global operation risk based on the operation risk prediction value. The evolutionary optimization engine is activated by the global scheduling node, which maps the Boolean matrix into a positional encoding of the high-dimensional solution space. The optimization objective function is used as the fitness evaluation criterion. Under the rigid condition of satisfying the probability confidence constraint equation, iterative optimization is carried out to generate the optimal allocation matrix that achieves the minimum fitness value. Based on the index of the element with a value of 1 extracted from the optimal Boolean matrix, the corresponding spatiotemporal feature terms are recalled and reconstructed into a global allocation instruction containing waypoint coordinates and execution time scale. The V2X edge communication protocol stack is invoked to send the global allocation instruction point-to-point to the corresponding physical agent to trigger an electromechanical response.

[0009] Furthermore, the expressions for the optimization objective function and the probability confidence constraint equation are as follows: In the formula, This represents the optimization objective function aimed at minimizing the overall operational risk. This represents the total number of physical intelligent agents in the set; This represents the total number of tasks in the time-series task set; Represents the elements in a Boolean matrix, when The temporal representation assigns the j-th temporal task to the i-th physical agent; Indicates the weight of the electromechanical energy consumption penalty; This represents the predicted electromechanical loss of the corresponding physical intelligent agent performing a task across space; This represents the expected probability of the i-th physical agent executing the j-th temporal task; This represents the preset probability confidence threshold for ensuring system scheduling; This represents the estimated risk of the task, i.e., the local risk of the i-th physical agent performing the j-th task.

[0010] Furthermore, step S5 specifically includes the following steps: Each physical agent relies on the vehicle-mounted ad hoc network to broadcast pose sequences in real time to build a dynamic communication subnet. When the real-time spatial Euclidean distance between the continuously monitored i-th physical agent and the adjacent k-th physical agent is less than the preset safe distance threshold, a trajectory warning event is triggered. In response to a trajectory warning event, neighboring nodes negotiate and obtain the electromechanical avoidance coefficient of the i-th physical agent by exchanging the operational risk estimates of the currently assigned tasks between the two parties through the dynamic communication subnet. The electromechanical avoidance coefficient is input into the local autonomous decision-making model to optimize the electromechanical operating parameters, including motor torque limiting, and generate motor torque limiting command values. The output drive signal controls the vehicle and components to perform emergency rescue operations, thereby establishing absolute right-of-way priority for high-risk tasks at the physical execution level.

[0011] Furthermore, step S6 specifically includes the following steps: In emergency rescue operations, the electromechanical response data of the physical intelligent agent is collected in real time and compared with the execution benchmark corresponding to the globally assigned instructions in time and space to extract the electromechanical execution deviation that characterizes the physical damping interference. Electromechanical execution deviations are converted into error compensation factors, and feature mapping weights in the anti-disturbance operation mapping model are adaptively adjusted to correct the expected completion probability in reverse. In response to physical triggering events such as sudden changes in on-site topology or exceeding the limits of the estimated operational risk, the current drive signal is interrupted at the chassis hardware control layer, and the global scheduling node is woken up across layers to re-perform the task optimization calculation.

[0012] A collaborative scheduling system for emergency medical robot components with multi-vehicle adaptation includes a processor and a memory. The memory stores a computer program, which, when executed by the processor, implements a collaborative scheduling method for emergency medical robot components.

[0013] By employing the above technical solution, the present invention provides a method and system for collaborative scheduling of multi-vehicle adapted emergency rescue robot components, which has at least the following beneficial effects: 1. This invention achieves coordinated scheduling of multiple mobile vehicles and different emergency rescue components by constructing a heterogeneous resource matrix and a spatiotemporal constraint generation mechanism. This avoids the problem of emergency rescue equipment being idle at critical moments due to a surge in sudden emergency rescue tasks, eliminates the negative impact of the traditional static allocation architecture on the adaptability to changing rescue needs, reduces the waste of control resources in complex disaster scenarios, and improves the overall rescue response efficiency of multi-vehicle clusters.

[0014] 2. By constructing an anti-disturbance operation mapping model, this invention avoids the distortion problems caused by asynchronous sampling frequency difference of sensors at the disaster site and physical damping to the control system, reduces the impact of communication limitations and environmental interference on collaborative decision-making, reduces the uncertainty of system scheduling in unknown dynamic environments, and ensures high confidence operation of the emergency rescue robot under physical interference.

[0015] 3. This invention constructs a Boolean matrix evolution optimization model that integrates electromechanical energy consumption penalty and probability confidence constraint equations. This model transforms the hierarchical priority of clinical medicine into a global operational risk orientation for robots. It prevents the misallocation of emergency resources caused by conventional algorithms ignoring the timeliness of medical care, avoids low-priority tasks occupying high-risk medical resources and affecting the survival rate of critically injured patients, reduces the risk of slow rescue in multi-disaster scenarios, and improves the life-saving protection of high-risk patients.

[0016] 4. This invention solves the problem of path congestion caused by network communication delays or environmental topology changes by using trajectory warning events based on comprehensive safety distance thresholds and risk-weighted electromechanical avoidance coefficients. It not only gives high-risk emergency rescue missions absolute physical right-of-way priority and reduces the negative impact of traditional obstacle avoidance algorithms on the timeliness of emergency rescue, but also reduces the probability of disaster collapse caused by force majeure on site, thus ensuring the survivability of emergency rescue robot clusters performing high-risk missions. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of the collaborative scheduling method of the present invention. Detailed Implementation

[0018] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This will allow for a full understanding of how the present application uses technical means to solve technical problems and achieve technical effects, and to facilitate its implementation.

[0019] Current methods for scheduling emergency medical robots struggle to adapt to the ever-changing demands of disaster scenes, leading to chaotic coordination among multiple vehicles and modular components. This results in emergency equipment being idle at critical moments, impacting overall rescue response efficiency. To avoid the problem of idle emergency equipment due to a surge in sudden emergency tasks, reduce the waste of control resources in complex disaster scenarios, and improve the overall rescue response efficiency of multi-vehicle clusters, this invention proposes a collaborative scheduling method for multi-vehicle adapted emergency medical robot components. This method relies on an intelligent system comprising a global scheduling node, a heterogeneous sensing network, and a cluster of physical agents.

[0020] The global scheduling node is deployed in an emergency command vehicle on the outskirts of the disaster area. It is equipped with a GPU processor with high-concurrency tensor computing capabilities and is responsible for receiving emergency response trigger signals from the central command and performing high-dimensional space evolution calculations.

[0021] The heterogeneous sensing network includes a swarm of drones with downward-looking lidar deployed at high altitudes, RGB-D depth cameras, and IoT environmental detectors scattered throughout the site, which are responsible for providing macroscopic terrain and environmental disturbance parameters.

[0022] The physical intelligent agent cluster is dynamically coupled by multiple mobile vehicles with different electromechanical characteristics (such as steering off-road vehicles and tracked rescue chassis) equipped with differentiated emergency rescue functional components (such as AED defibrillator mounting modules, modular demolition robotic arms, and constant temperature blood transport chambers); each physical intelligent agent directly controls the motors and component actuators through the built-in MCU control board and is equipped with a V2V short-range radio frequency communication module.

[0023] The emergency medical robot component collaborative scheduling method in this embodiment is as follows: Figure 1 As shown, it includes the following steps: S1. In response to the emergency rescue trigger signal, the spatiotemporal metadata of the signal is parsed to define the three-dimensional physical boundary of the target work space. A heterogeneous sensing network mapped within the three-dimensional physical boundary is then activated. This network probes the target work space, collecting environmental disturbance parameters characterizing external physical interference. Simultaneously, electromechanical state parameters of each internal physical node are extracted via the underlying hardware bus. Based on spatiotemporal stamps, the acquired environmental disturbance parameters and electromechanical state parameters are filtered and denoised to remove outliers from the physical samples. The preprocessed environmental disturbance parameters and electromechanical state parameters are then output. Aligning the environmental disturbance parameters and electromechanical state parameters to the same spatiotemporal stamp plane avoids spatiotemporal misalignment of the data source, improving the scheduling accuracy of subsequent emergency rescue robots.

[0024] S2. To avoid the inability to recognize the time-sensitive nature of emergency rescue missions, leading to resource misallocation where high-value equipment is used for low-priority tasks and impacting the survival rate of high-risk patients, the kinematic parameters of the vehicle and the physical mounting parameters of the components are extracted from the preprocessed electromechanical state parameters. A heterogeneous resource matrix representing vehicle movement constraints and component operational capabilities is constructed through vector feature concatenation. Based on the emergency rescue trigger signal, three-dimensional spatial coordinates and medical classification indicators are extracted. According to a nonlinear temporal decay rule, the medical classification indicators are mapped to time window constraints with rigid cutoff boundaries. The three-dimensional spatial coordinates, time window constraints, and component capability type constraints required to perform the emergency rescue operation are jointly encapsulated to generate a temporal task set with spatiotemporal constraints. This reduces the probability of misallocation of emergency rescue resources and avoids the problem of emergency equipment being idle at critical moments due to a surge in sudden emergency rescue missions.

[0025] S3. Based on the heterogeneous resource matrix and preprocessed environmental disturbance parameters, a disturbance-resistant operation mapping model of a physical intelligent agent consisting of dynamically coupled vehicles and components is constructed. Based on this model, the expected completion probability and operation risk prediction are initialized for each task in the time-series task set. This avoids distortion problems caused by asynchronous sensor sampling frequencies and physical damping at the disaster site, and reduces the impact of communication limitations and environmental interference on collaborative decision-making.

[0026] In this embodiment, as a preferred implementation method, the implementation steps are as follows: First, the physical damping coefficient vector is extracted from the environmental disturbance parameters. An environmental attenuation tensor matching the dimension of the heterogeneous resource matrix is ​​constructed. Through tensor Hadamard product operation, the heterogeneous resource matrix is ​​transformed into an anti-disturbance operation mapping matrix representing a physical agent composed of dynamic coupling of vehicles and components. The calculation formula is as follows: In the formula, Represents the anti-disturbance operation mapping matrix; Represents a heterogeneous resource matrix; Represents the vector of physical damping coefficients; This represents the Hadama product operator; This represents a nonlinear mapping function that calibrates the damping attenuation characteristics of the physical space.

[0027] Secondly, based on the time-series task set, the Euclidean space coordinate deviation and remaining absolute time window of each time-series task are extracted as spatiotemporal feature terms. Taking the effective movement speed scalar in the anti-disturbance task mapping matrix as the benchmark, the expected completion probability of the i-th physical agent performing the j-th time-series task is calculated independently using the logistic function. The calculation formula is as follows: In the formula, This represents the expected probability of the i-th physical agent executing the j-th temporal task; This represents the collaborative confidence hyperparameter; Represents a Boolean variable characterizing the physical mounting capability matching constraint of the component; This represents the Euclidean space coordinate deviation of each time-series task; Represents the remaining absolute time window for each time-series task; This represents the effective movement speed scalar.

[0028] Finally, the medical grading indicators in the time-series task set are coupled with the expected completion probability to calculate the operational risk prediction for each time-series task. The calculation formula is as follows: In the formula, This indicates the estimated risk level of the operation. Indicators representing medical grading; Indicates the priority penalty amplification factor; This is a characterization of the inherent damage risk of environmental background noise.

[0029] S4. Based on the estimated operational risk and guided by minimizing global operational risk, under the constraint of a preset probability confidence threshold, the global scheduling node performs task optimization calculation to generate global allocation instructions and sends the global allocation instructions to the corresponding physical agents. This fundamentally reduces the possibility of secondary disasters caused by the emergency rescue robot stopping en route and eliminates the waste of computing power and transportation capacity caused by task overload allocation.

[0030] In this embodiment, as a preferred implementation method, the implementation steps are as follows: Extract the identifier set and temporal task set of the physical agent, define a Boolean matrix, and construct an optimization objective function and probability confidence constraint equation guided by minimizing the global task risk based on the task risk prediction value. The expressions are as follows: In the formula, This represents the optimization objective function aimed at minimizing the overall operational risk. This represents the total number of physical intelligent agents in the set; This represents the total number of tasks in the time-series task set; Represents the elements in a Boolean matrix, when The temporal representation assigns the j-th temporal task to the i-th physical agent; Indicates the weight of the electromechanical energy consumption penalty; This represents the predicted electromechanical loss of the corresponding physical intelligent agent performing a task across space; This represents the expected probability of the i-th physical agent executing the j-th temporal task; This represents the preset probability confidence threshold for ensuring system scheduling; This represents the estimated risk of the task, i.e., the local risk of the i-th physical agent performing the j-th task.

[0031] The evolutionary optimization engine is activated by the global scheduling node, which maps the Boolean matrix into a positional encoding of the high-dimensional solution space. Using the optimization objective function as the fitness evaluation criterion, iterative optimization is performed under the rigid condition of satisfying the probability confidence constraint equation to generate the optimal allocation matrix that achieves the minimum fitness value.

[0032] Based on the index of the element with a value of 1 extracted from the optimal Boolean matrix, the corresponding spatiotemporal feature terms are recalled and reconstructed into a global allocation instruction containing waypoint coordinates and execution time scale. The V2X edge communication protocol stack is invoked to send the global allocation instruction point-to-point to the corresponding physical agent to trigger an electromechanical response.

[0033] S5. Each physical agent responds to dynamic spatial interference events, triggers negotiation between adjacent nodes in the dynamic communication subnet as needed to resolve spatiotemporal conflicts, and runs a local autonomous decision-making model to optimize electromechanical operating parameters, generating drive signals to control vehicles and components to perform emergency rescue operations; thereby avoiding high-priority task blockage caused by conventional obstacle avoidance, reducing the probability of collisions caused by spatial interference, and achieving hard real-time, high-priority absolute right-of-way guarantee in weak network environments.

[0034] In this embodiment, as a preferred implementation method, the implementation steps are as follows: Each physical agent relies on the vehicle-mounted ad hoc network to broadcast pose sequences in real time to build a dynamic communication subnet. When the real-time spatial Euclidean distance between the continuously monitored i-th physical agent and the adjacent k-th physical agent is less than the preset safe distance threshold, a trajectory warning event is triggered.

[0035] In response to a trajectory warning event, neighboring nodes negotiate and obtain the electromechanical avoidance coefficient of the i-th physical agent by exchanging the operational risk estimates of the currently assigned tasks between the two parties through the dynamic communication subnet.

[0036] The electromechanical avoidance coefficient is input into the local autonomous decision-making model to optimize the electromechanical operating parameters, including motor torque limiting, and generate motor torque limiting command values. The output drive signal controls the vehicle and components to perform emergency rescue operations, thereby establishing absolute right-of-way priority for high-risk tasks at the physical execution level.

[0037] S6. In order to reduce the fatal impact of environmental changes on the overall rescue, real-time operational deviation data of physical agents in emergency rescue operations are collected. Based on the real-time operational deviation data, the anti-disturbance operation mapping model is corrected in reverse. In response to physical triggering events such as environmental topology changes or operational risk exceeding limits, the current control command is interrupted and the task optimization calculation in step S4 is re-triggered. This enables the emergency rescue robot cluster to take over in response to environmental changes and reduces the probability of emergency rescue robots getting stuck due to environmental changes.

[0038] In this embodiment, as a preferred implementation method, the implementation steps are as follows: In emergency rescue operations, the electromechanical response data of the physical intelligent agent is collected in real time and compared with the execution benchmark corresponding to the globally assigned instructions in time and space to extract the electromechanical execution deviation that characterizes the physical damping interference. Electromechanical execution deviations are converted into error compensation factors, and feature mapping weights in the anti-disturbance operation mapping model are adaptively adjusted to correct the expected completion probability in reverse. In response to physical triggering events such as sudden changes in on-site topology or exceeding the limit of the estimated operation risk, the current drive signal is interrupted at the chassis hardware control layer, and the global scheduling node is woken up across layers to re-perform the task optimization solution.

[0039] This method first responds to emergency response signals by preprocessing collected environmental disturbance parameters and electromechanical state parameters to construct a heterogeneous resource matrix and mapping medical grading indicators to generate a time-series task set via rigid deadline windows. Subsequently, it integrates environmental decay tensors to construct an anti-disturbance operation mapping model for physical agents, independently quantifying the expected completion probability and the estimated local operation risk. Based on this, the global scheduling node, under the rigid condition of satisfying the preset probability confidence constraint equation, performs high-dimensional evolutionary optimization on the Boolean matrix (incorporating electromechanical energy consumption penalties) and issues global allocation instructions, guided by minimizing global operation risk. Each physical agent responds to trajectory warning events triggering risk-weighted negotiation, directly constraining motor torque limit command values ​​through local decisions, establishing absolute priority for high-risk tasks at the physical execution level. Finally, it extracts electromechanical execution deviations to close-loop correct the anti-disturbance model, and in the event of physical events such as sudden topological changes, it directly interrupts signals through the hardware control layer and wakes up global rescheduling across layers.

[0040] The present invention also provides a multi-vehicle-adaptive emergency medical robot component collaborative scheduling system, including a processor and a memory, wherein the memory is used to store a computer program, and the computer program is executed by the processor to implement an emergency medical robot component collaborative scheduling method.

[0041] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0042] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Since the above embodiments are substantially similar to the method embodiments, their descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0043] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for collaborative scheduling of components of a multi-vehicle-adaptive emergency rescue robot, characterized in that, The method includes the following steps: S1. In response to the emergency rescue trigger signal, collect the environmental disturbance parameters of the target scene, extract the electromechanical status parameters of each physical node, and perform preprocessing. S2. Construct a heterogeneous resource matrix based on electromechanical state parameters, and generate a time-series task set based on medical classification indicators in the emergency response trigger signal; S3. Based on the heterogeneous resource matrix and environmental disturbance parameters, construct the disturbance-resistant operation mapping model of the physical agent, and initialize the expected completion probability and operation risk prediction value for each task in the time-series task set accordingly. S4. Based on the estimated operation risk and guided by minimizing the global operation risk, under the constraint of satisfying the preset probability confidence threshold, the global scheduling node performs task optimization calculation to generate a global allocation instruction and sends it to the corresponding physical agent. S5. Each physical intelligent agent responds to dynamic spatial interference events, triggers negotiation between adjacent nodes in the dynamic communication subnet as needed to resolve spatiotemporal conflicts, runs a local autonomous decision-making model to optimize electromechanical operating parameters, and generates drive signals to control vehicles and components to carry out emergency rescue operations. S6. Collect real-time operational deviation data of the physical agent, correct the anti-disturbance operation mapping model in reverse according to it, respond to physical triggering events such as environmental topology changes or operation risk exceeding limits, interrupt the current control command and re-trigger the task optimization calculation.

2. The cooperative scheduling method according to claim 1, characterized in that, Step S2 specifically includes the following steps: Extract vehicle kinematic parameters and component physical mounting parameters from electromechanical state parameters, and construct a heterogeneous resource matrix characterizing vehicle movement constraints and component operation capabilities through feature splicing; Based on the emergency trigger signal, three-dimensional spatial coordinates and medical classification indicators are extracted, and the medical classification indicators are mapped to time window constraints with rigid cutoff boundaries according to the nonlinear time-series decay rule. The three-dimensional spatial coordinates, time window constraints, and component capability type constraints required to perform the emergency rescue operation are jointly encapsulated to generate a time-series task set with spatiotemporal constraints.

3. The cooperative scheduling method according to claim 1, characterized in that, Step S3 specifically includes the following steps: The physical damping coefficient vector is extracted from the environmental disturbance parameters, and an environmental attenuation tensor matching the dimension of the heterogeneous resource matrix is ​​constructed. Through tensor Hadamard product operations, the heterogeneous resource matrix is ​​transformed into an anti-disturbance operation mapping matrix for the physical agent. The calculation formula is as follows: In the formula, Represents the anti-disturbance operation mapping matrix; Represents a heterogeneous resource matrix; Represents the vector of physical damping coefficients; This represents the Hadama product operator; The nonlinear mapping function represents the physical space damping attenuation characteristics. Based on the temporal task set, the Euclidean space coordinate deviation and remaining absolute time window of each temporal task are extracted as spatiotemporal feature terms. Using the effective movement speed scalar in the anti-disturbance task mapping matrix as a benchmark, the expected completion probability of the i-th physical agent performing the j-th temporal task is independently quantized using the logistic function. The calculation formula is as follows: In the formula, This represents the expected probability of the i-th physical agent completing the j-th temporal task; This represents the collaborative confidence hyperparameter; A Boolean variable representing the physical mounting capability matching constraint of the component; This represents the Euclidean space coordinate deviation of each time-series task; Represents the remaining absolute time window for each time-series task; Represents the effective movement speed scalar; The medical grading indicators in the time-series task set are coupled with the expected completion probability to calculate the operational risk prediction for each time-series task. The calculation formula is as follows: In the formula, This indicates the estimated risk level of the operation. Indicators representing medical grading; Indicates the priority penalty amplification factor; This is a characterization of the inherent damage risk of environmental background noise.

4. The cooperative scheduling method according to claim 1, characterized in that, Step S4 specifically includes the following steps: Extract the identifier set and temporal task set of physical agents, define a Boolean matrix, and construct an optimization objective function and probability confidence constraint equation guided by minimizing global operation risk based on the operation risk prediction value. The evolutionary optimization engine is activated by the global scheduling node, which maps the Boolean matrix into a positional encoding of the high-dimensional solution space. The optimization objective function is used as the fitness evaluation criterion. Under the rigid condition of satisfying the probability confidence constraint equation, iterative optimization is carried out to generate the optimal allocation matrix that achieves the minimum fitness value. Based on the index of the element with a value of 1 extracted from the optimal Boolean matrix, the corresponding spatiotemporal feature terms are recalled and reconstructed into a global allocation instruction containing waypoint coordinates and execution time scale. The V2X edge communication protocol stack is then invoked to send the global allocation instruction point-to-point to the corresponding physical agent.

5. The cooperative scheduling method according to claim 4, characterized in that, The expressions for the optimization objective function and the probability confidence constraint equation are as follows: In the formula, This represents the optimization objective function aimed at minimizing the overall operational risk. This represents the total number of physical intelligent agents in the set; This represents the total number of tasks in the time-series task set; Represents the elements in a Boolean matrix; Indicates the weight of the electromechanical energy consumption penalty; This represents the predicted electromechanical loss of the corresponding physical intelligent agent performing a task across space; This represents the expected probability of the i-th physical agent completing the j-th temporal task; This indicates a preset probability confidence threshold. This indicates the estimated risk level of the operation.

6. The cooperative scheduling method according to claim 1, characterized in that, Step S5 specifically includes the following steps: Each physical agent relies on the vehicle-mounted ad hoc network to broadcast pose sequences in real time to build a dynamic communication subnet. When the real-time spatial Euclidean distance between the continuously monitored i-th physical agent and the adjacent k-th physical agent is less than the preset safe distance threshold, a trajectory warning event is triggered. In response to a trajectory warning event, neighboring nodes negotiate and obtain the electromechanical avoidance coefficient of the i-th physical agent by exchanging the operational risk estimates of the currently assigned tasks between the two parties through the dynamic communication subnet. The electromechanical avoidance coefficient is input into the local autonomous decision-making model to optimize the electromechanical operating parameters, including the motor torque limit, and generate the motor torque limit command value to output the drive signal to control the vehicle and components to perform emergency rescue operations.

7. The cooperative scheduling method according to claim 1, characterized in that, Step S6 specifically includes the following steps: In emergency rescue operations, the electromechanical response data of the physical intelligent agent is collected in real time and compared with the execution benchmark corresponding to the globally assigned instructions in time and space to extract the electromechanical execution deviation that characterizes the physical damping interference. Electromechanical execution deviations are converted into error compensation factors, and feature mapping weights in the anti-disturbance operation mapping model are adaptively adjusted to correct the expected completion probability in reverse. In response to physical triggering events such as sudden changes in on-site topology or exceeding the limits of the estimated operational risk, the current drive signal is interrupted at the chassis hardware control layer, and the global scheduling node is woken up across layers to re-perform the task optimization calculation.

8. A collaborative scheduling system for components of an emergency medical robot as described in any one of claims 1-7, characterized in that, It includes a processor and a memory, the memory being used to store a computer program, which, when executed by the processor, implements the emergency robot component collaborative scheduling method as described in any one of claims 1-7.

Images