Multi-modal robotic key actuator dynamic power migration method, system, medium, and apparatus

Abstract

The present application relates to the technical field of robot energy management, and provides a multi-modal robot key actuator dynamic power migration method, system, medium and equipment, comprising: for the current task stage, performing task risk assessment on each actuator, and determining key actuators and non-key actuators; according to the energy state, estimating the current available power, determining the minimum holding power and power budget of each actuator; when the actual output of the key actuator is lower than the target output, and the insufficient output affects the task holding index, judging whether the insufficient output is caused by multi-actuator energy competition, if yes, reducing the power budget of at least one non-key actuator, and configuring the energy supply margin to the key actuator; re-detecting the actual output of the key actuator and the task holding index, and maintaining, expanding, canceling the power migration or performing task degradation according to the detection result. Thus, the task reliability in the key stage is improved.

Description

Technical Field

[0001] This invention belongs to the field of robot energy management technology, and in particular relates to a method, system, medium and equipment for dynamic power transfer of key actuators in multimodal robots. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] As robot applications continue to become more complex and unstructured, traditional single-mode robots can usually only complete specific tasks. Wheeled robots have high mobility but struggle to overcome complex obstacles, legged robots have strong environmental adaptability but high mobility and control complexity, and flying robots have the ability to move quickly in space but are limited by endurance and energy capacity.

[0004] Multimodal robots integrate multiple motion capabilities such as crawling, jumping, flying, and attitude adjustment, enabling them to select different motion modes according to environmental changes and achieve autonomous movement and task execution in complex scenarios. However, the introduction of multiple motion modalities also makes robot energy management more complex: unlike single-actuator robots, multimodal robots typically contain multiple actuators with different power requirements. For example, crawling actuators usually need to operate stably for a long time and have high continuous power requirements; jumping actuators usually need to release a large amount of energy in a short time and have high transient power requirements; flying actuators need to continuously provide high-frequency, high-response power output; and attitude adjustment actuators need to respond quickly to changes in the robot's attitude.

[0005] In actual operation, the importance of actuators at different stages of motion is not fixed. During a search on flat ground, the crawling module undertakes the main motion task; when encountering obstacles, the jumping module becomes crucial for completing the task; and after takeoff, the flight module and attitude adjustment mechanism determine whether the robot can stably complete the crossing and landing. Therefore, the core issue of energy management for multimodal robots is not simply reducing energy consumption, but rather, under limited energy conditions, dynamically determining the most critical motion capabilities for the current task and rationally adjusting energy resource allocation.

[0006] Existing robot energy management methods have significant shortcomings: The power allocation method based on fixed priority pre-sets the importance of different actuators (e.g., the flight module has the highest priority and the auxiliary module has the lowest priority). However, the task state changes constantly during the operation of a multimodal robot, and the fixed priority cannot adapt to the switching of motion modes. For example, the flight module does not need high priority power supply during the crawling phase, but the flight module becomes a key actuator during the aerial adjustment phase. The fixed strategy is prone to energy waste or insufficient key capabilities. Battery-based energy protection methods typically detect battery voltage, current, and remaining charge to provide undervoltage protection or overall load reduction. However, this method focuses primarily on energy security rather than task completion capability. When energy is insufficient, traditional methods usually reduce the output of all actuators, which may cause the robot to lose its mobility, terminate the task prematurely, or fail to complete critical actions. While multi-actuator coordinated control methods optimize output synchronization, motion accuracy, and control stability among multiple motors or actuators, they lack a dynamic power scheduling mechanism to maintain critical tasks under energy competition conditions.

[0007] In summary, existing robot energy management methods typically employ fixed power allocation, preset priorities, or uniform load reduction strategies based on battery status. These methods do not fully consider the dynamic changes in critical execution capabilities during continuous motion tasks. When multiple actuators operate simultaneously and the system power supply is limited, problems such as insufficient output from critical actuators, non-critical loads consuming limited energy, bus voltage fluctuations, and task interruptions can easily occur. Summary of the Invention

[0008] To address the technical problems mentioned above, this invention provides a method, system, medium, and device for dynamic power transfer of key actuators in multimodal robots. This transforms the robot's energy management objective from simply reducing power consumption or protecting the power system to maintaining the ability to complete the current task under limited energy constraints. It directly links energy scheduling with the robot's task objectives, dynamically identifying key actuators based on the motion task and the risk of task failure. This avoids energy waste or insufficient protection of critical capabilities caused by fixed priorities. Furthermore, by identifying energy competition among multiple actuators and implementing power transfer, power resources released by non-critical actuators are prioritized for supplying key actuators, thereby improving the reliability of tasks during critical phases such as jumping, flight assistance, attitude adjustment, and landing recovery.

[0009] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of the present invention provides a dynamic power transfer method for a key actuator of a multimodal robot, comprising: Acquire the current task state, motion state, target output and actual output of each actuator, and energy status of the multimodal robot; Based on the current task status and motion status, determine the current task phase and the corresponding task maintenance indicators; For the current task phase, analyze the necessity of each actuator, the risk of task failure due to decreased output capability, the deviation between target output and actual output, and the substitutable capabilities, conduct a task risk assessment for each actuator, and identify critical and non-critical actuators. Estimate the current available power based on the energy status, determine the minimum holding power and power budget for each actuator; when the actual output of a critical actuator is lower than the target output and the insufficient output affects the mission holding performance, determine whether the insufficient output is caused by energy competition among multiple actuators; When it is determined that the insufficient output is caused by energy competition among multiple actuators, the power budget of at least one non-critical actuator is reduced, and the resulting power margin is allocated to the critical actuator. After the power migration is completed, the actual output and task retention indicators of the critical actuator are re-evaluated, and the power migration is maintained, expanded, or canceled, or the task is degraded, based on the evaluation results.

[0010] Furthermore, the task risk assessment of each actuator and the identification of critical and non-critical actuators include: Calculate the task risk assessment value R for the i-th actuator. i =a1N i +a2F i +a3D i +a4H i +a5B i , where N i F indicates the degree to which the executor is necessary for the current task. i D indicates the risk of primary task failure due to actuator failure. i H represents the normalized deviation between the actuator's target output and its actual output. i This indicates the impact of the actuator output decrease on the robot's safety status, B. i This indicates the ability of other actuators to substitute for its function; a1, a2, a3, a4, and a5 are weighting coefficients. The task risk assessment value R i With the critical threshold R H Non-critical threshold R L Comparison, where R H Greater than R L ; When R i Above the critical threshold R H Furthermore, if the duration exceeds a preset threshold, the corresponding actuator will be identified as a critical actuator; When R i Below the non-critical threshold R L Furthermore, if the duration exceeds a preset threshold, the corresponding actuator will be identified as a non-critical actuator; When R i In R H With R L If the duration exceeds a preset threshold, the corresponding actuator will be identified as a restricted actuator.

[0011] Furthermore, the step of estimating the current available power based on the state of energy and determining the minimum holding power and power budget for each actuator includes: The currently allocable power is determined based on the bus voltage, allowable output current, battery continuous output capacity, power conversion module limits, and safety reserve power: P ava =min(U bus I lim ,P bat ,P conv )-P safe Among them, P ava U represents the currently allocable power. bus I represents the bus voltage. lim P represents the allowable current of the busbar. bat P represents the permissible output power determined based on the battery's state of charge, temperature, and internal resistance. conv Indicates the allowable output power of the power conversion module, P safe This refers to the safety power reserved for control, necessary sensing, and safety functions. For the i-th actuator, determine the minimum holding power P. i,min Among them, the key actuator P i,min Based on the mission retention metric, the P of non-critical actuators is determined. i,min To maintain the power required for standby or safe operation, the auxiliary load that is allowed to be shut down has a P i,min It is zero; For the i-th actuator, the power budget Of which, the remaining power excluding the minimum holding power Remaining power allocation weight Maximum task risk assessment value , is the task risk assessment value, and n is the number of actuators participating in the remaining power allocation.

[0012] Furthermore, the target output O of the key actuator is detected. k,tar With actual output O k,act The deviation between them is calculated, and the normalized output deficiency D is calculated. k =max(0,O) k,tar -O k,act ) / O k,tar When D k When the output deviation exceeds the preset threshold and continues for a predetermined control period, it is determined whether the insufficient output affects the task maintenance index. If the task risk assessment value of the key actuator is higher than the trigger threshold, it is determined whether the insufficient output is caused by energy competition among multiple actuators.

[0013] Furthermore, the determination of whether insufficient output is caused by energy contention among multiple actuators is based on at least two of the following conditions: The bus voltage is lower than the normal operating range or drops when the critical actuator is overloaded; The total current is close to or has reached the permissible current limit; The critical actuator requested more power than it actually received. At least one non-critical actuator has an actual power higher than its minimum holding power and has releasable power; No stall, over-temperature, open circuit, or independent fault was detected in the critical actuator driver; After reducing the trial power of non-critical actuators, the output of critical actuators showed a recovery trend.

[0014] Furthermore, the current mission phase includes at least three consecutive phases: ground approach phase, jump energy storage phase, takeoff transition phase, air support phase, attitude adjustment or landing preparation phase, and landing recovery phase.

[0015] Furthermore, the power migration is achieved by at least one of the following methods: adjusting the power budget of the power supply branch, adjusting the branch current limiting threshold, adjusting the actuator drive current limit, adjusting the torque limit, adjusting the target speed, adjusting the operating frequency, adjusting the PWM duty cycle, adjusting the action amplitude, adjusting the operating duty cycle, and adjusting the startup sequence of multiple actuators.

[0016] A second aspect of the present invention provides a dynamic power transfer system for a key actuator of a multimodal robot, comprising: The sensing module is configured to acquire the current task state, motion state, target output and actual output of each actuator, and energy status of the multimodal robot. The task status analysis module is configured to determine the current task stage and the corresponding task maintenance index based on the current task status and motion status. The task risk assessment module is configured to: for the current task stage, analyze the necessity of each actuator, the risk of task failure due to decreased output capability, the deviation between target output and actual output, and the substitutable capability, conduct task risk assessment for each actuator, and identify critical and non-critical actuators. The energy status detection module is configured to: estimate the current available power based on the energy status, determine the minimum holding power and power budget of each actuator; and determine whether the insufficient output is caused by energy competition among multiple actuators when the actual output of a critical actuator is lower than the target output and the insufficient output affects the task holding index. The control module is configured to: when it is determined that the insufficient output is caused by energy competition among multiple actuators, reduce the power budget of at least one non-critical actuator and allocate the resulting power margin to the critical actuator; after the power migration is completed, re-detect the actual output and task hold indicators of the critical actuator, and maintain, expand or cancel the power migration, or perform task degradation based on the detection results.

[0017] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the dynamic power transfer method for key actuators of a multimodal robot as described above.

[0018] A fourth aspect of the present invention provides a computer device including a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, wherein the processor executes the program to implement the steps in the dynamic power transfer method for key actuators of a multimodal robot as described above.

[0019] Compared with the prior art, the beneficial effects of the present invention are: This invention transforms the goal of robot energy management from simply reducing power consumption or protecting the power system to maintaining the ability to complete the current task under limited energy constraints. It directly links energy scheduling with the robot's task objectives, dynamically identifies critical actuators based on the motion task and the risk of task failure, avoids energy waste or insufficient protection of critical capabilities caused by fixed priorities, and prioritizes the supply of power resources released by non-critical actuators to critical actuators by identifying energy competition among multiple actuators and implementing power migration, thereby improving the reliability of tasks in critical stages such as jumping, flight assistance, attitude adjustment, and landing recovery.

[0020] The key actuator identification, task risk assessment, and power transfer mechanisms of this invention do not depend on specific robot mechanical structures and can be extended to multimodal biomimetic robots that fly, climb, and jump, air-ground collaborative robots, legged robots, and other unmanned systems containing multiple heterogeneous actuators. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0022] Figure 1 This is a flowchart of the dynamic power transfer method for key actuators of a multimodal robot according to Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the multimodal robot composition according to Embodiment 1 of the present invention; Figure 3This is a schematic diagram of the functional architecture and information closed loop of the task-keeping dynamic power transfer system according to Embodiment 1 of the present invention; Figure 4 This is a schematic diagram illustrating the dynamic migration relationship between the continuous task phase and the key actuator in Embodiment 1 of the present invention. Figure 5 This is a schematic diagram of task risk assessment and actuator classification determination according to Embodiment 1 of the present invention; Figure 6 This is a schematic diagram of the energy competition determination and dynamic power migration closed loop of key actuators in Embodiment 1 of the present invention; Figure 7 This is a schematic diagram illustrating the power budget changes of each actuator at different stages of a continuous task according to Embodiment 1 of the present invention. Figure 8 This is a schematic diagram of graded protection and gradual recovery under abnormal energy conditions according to Embodiment 1 of the present invention; Figure 9 This is a schematic diagram of the structure of a computer device according to Embodiment 4 of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0024] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] Terminology Explanation: Critical actuators are those whose output capability, during the current task phase, falls below the task requirements, directly leading to the failure to achieve the main motion objective, a significant decrease in robot stability, or the inability to maintain a safe state. Non-critical actuators, on the other hand, are those whose output can be reduced, delayed, or paused within a predetermined range without immediately causing the failure of the current main task. The distinction between critical and non-critical actuators is not permanently fixed by actuator type but is dynamically determined by the current task objective, motion phase, output state, and task failure risk.

[0026] Power migration: It is not limited to the direct physical transfer of electrical energy between two actuators, but refers to the control module redetermining the power budget, current limit threshold, drive duty cycle, output frequency, torque limit, start-up sequence or working duty cycle of each power branch, so that non-critical actuators release adjustable power and the resulting power supply margin is allocated to critical actuators.

[0027] Task retention constraint: This refers to the requirement that, under conditions of energy constraints or energy competition among multiple actuators, at least one task performance indicator corresponding to the robot's current primary task must not fall below a preset retention threshold. The task performance indicator may include crawling speed, obstacle-crossing energy storage, available power for takeoff, flight actuator rotation speed or flapping frequency, attitude angle error, angular velocity, ground clearance, landing attitude, and other state variables that characterize the robot's task completion capability.

[0028] Example 1 To address the mismatch between energy scheduling and motion task requirements in existing multimodal robots during continuous task execution in complex environments, this embodiment provides a dynamic power transfer method for key actuators of multimodal robots.

[0029] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment aims to solve the following key technical problems: (1) How to dynamically identify key actuators that affect the completion of the task based on the current task objective and motion phase: Traditional methods usually assume that the importance of actuators is fixed; This embodiment dynamically determines the key actuators that must be guaranteed in the current phase by analyzing the current motion mode, task objective, actuator contribution, and task failure risk. (2) How to determine whether the output drop of the key actuator is caused by energy competition among multiple actuators, and how to solve the energy competition problem among multiple actuators: When multiple actuators are running at the same time, limited energy may not be able to meet all the needs; This embodiment determines whether there is energy competition by detecting the output deviation of the key actuator, the energy supply and demand relationship, and the power occupation of non-key actuators, and performs power migration. (3) How to prioritize key motion capabilities and maintain the robot's task completion capability under limited energy supply conditions by reducing the energy consumption of non-critical actuators and migrating power resources: This embodiment differs from traditional energy protection strategies. It does not aim to reduce overall power consumption, but rather prioritizes key motion capabilities that affect task completion while allowing the sacrifice of non-critical functions. (4) How to maintain control, perception and critical mission capabilities during energy anomalies and recovery processes, and avoid overall system failure or repeated entry into protection states.

[0030] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment is applicable to biomimetic robots with multiple motion modes, air-ground collaborative robots, complex environment exploration robots, small unmanned systems, and other autonomous robot systems containing multiple heterogeneous actuators.

[0031] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment dynamically evaluates the importance of different actuators based on the robot's current motion task requirements, execution capability status, and energy status, and ensures key motion capabilities through power resource reallocation.

[0032] The dynamic power migration method for key actuators of multimodal robots provided in this embodiment realizes intelligent energy scheduling for multimodal robots in complex environments by constructing a closed loop of task status perception, key capability assessment, energy competition judgment, and power migration control.

[0033] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment has the following specific objectives: (1) Establish a mapping relationship between motion tasks and actuator energy requirements to achieve dynamic identification of key actuators; (2) A power transfer mechanism based on task-keeping constraints is proposed to ensure key motion capabilities under limited energy conditions; (3) Improve the reliability of task completion of multimodal robots during continuous motion mode switching; (4) Reduce energy competition caused by multiple actuators operating simultaneously and improve system stability.

[0034] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment proposes a dynamic power transfer technology concept oriented towards task preservation. Its core is not to simply reduce the overall power consumption, but to maximize the robot's ability to complete the current task under limited energy conditions. To this end, a coupling relationship is established between task state, execution capability and energy state.

[0035] Among them, the task status is used to characterize the motion target and task stage that the robot needs to complete at present; the execution capability is used to characterize the contribution of different actuators to the current task and the risk of task failure caused by the performance degradation; the energy status is used to characterize the power that the system can currently provide, the bus voltage, the battery status, and the power occupancy of each actuator.

[0036] During robot operation, the importance of each actuator is dynamically assessed based on the current task objective and motion phase, identifying critical and non-critical actuators. When the actual output of a critical actuator is lower than the target output required to meet the current task, and it is determined that this insufficient output is related to energy competition within the system, instead of uniformly reducing the load on all actuators, the power consumption of non-critical actuators or auxiliary loads with lower current task contribution is prioritized for reduction, and the released power resources are transferred to critical actuators. Through this mechanism, energy resources are dynamically transferred from low-task-contribution actuators to high-task-contribution actuators, thereby maintaining the robot's critical motion capabilities and continuous task execution capabilities.

[0037] The dynamic power transfer method for key actuators in multimodal robots provided in this embodiment is applicable to, for example, Figure 2 The robot 100 is shown as a multi-modal robot capable of flying, crawling, and jumping. The robot 100 includes a power module 101, a power management module 102, an energy status detection module 103, an actuator status detection module 104, a control module 105, a crawling module 106, a jumping module 107, a flight module 108, an attitude adjustment mechanism 109, a sensing module 110, a communication module 111, an actuator drive module 112, an auxiliary load module 113, a task status analysis module 114, and a task risk assessment module 115.

[0038] The power module 101 may include a rechargeable battery, a supercapacitor, or a hybrid energy storage unit consisting of a battery and a supercapacitor, for providing power to the various functional modules of the robot. The power management module is used to perform power conversion, voltage regulation, distribution, and power supply protection. The power management module 102 is connected to the power module 101 and supplies power to each execution module through at least one common bus and multiple controllable power supply branches. Each power supply branch may be equipped with independent switching devices, current limiting devices, DC-DC converter circuits, or programmable drive circuits, so that the control module 105 can adjust the power limit of the corresponding module.

[0039] The energy status detection module 103 is used to acquire at least one of the following: battery terminal voltage, bus voltage, total current, current of each branch, current available power, remaining state of charge, temperature, and allowable output capacity of the power supply.

[0040] The actuator status detection module 104 is used to acquire the drive current, input power, output power, position, working status, speed, displacement, torque, working frequency, or output saturation status of the crawling motor, jump energy storage driver, flight actuator, and attitude adjustment actuator.

[0041] The control module 105 is connected to the energy status detection module 103, actuator status detection module 104, sensing module 110, task status analysis module 114, task risk assessment module 115, and actuator drive module 112, respectively, and is used to perform task analysis, critical actuator identification, energy competition judgment, power migration decision-making, and recovery control. The control module 105 can be implemented by a single embedded processor, or it can be implemented collaboratively by a main controller and a power management slave controller. The task status analysis module 114 and the task risk assessment module 115 can be set as independent processing units, or they can be set as software function modules executed by the control module 105.

[0042] The crawling module 106 is used for the robot's continuous movement on the ground; the jumping module 107 is used to complete the rapid obstacle crossing formed by energy storage, locking and releasing; the flight module 108 is used to provide aerodynamic force during the robot's independent flight, flight assistance after jumping, or aerial trajectory correction phase; the attitude adjustment mechanism 109 is used to generate attitude adjustment torque to adjust the robot's attitude, actuator angle, or motion output direction.

[0043] The sensing module 110 is used to acquire information about the robot's motion state, environmental state, and contact state.

[0044] The communication module 111 is used for data communication between the robot and external terminals, other robots, or task systems.

[0045] The actuator drive module 112 is used to drive each actuator according to the power migration and output constraint commands output by the control module.

[0046] The auxiliary load module 113 may include lighting, non-critical camera, environmental sampling, extended sensing, data logging, auxiliary computing, extended communication, and other loads not directly related to the current main motion task.

[0047] The task status analysis module 114 is used to analyze the current task type, task stage, task objective, and key motion capabilities required to complete the task.

[0048] The task risk assessment module 115 is used to assess the impact of performance degradation of each actuator on task completion and to identify critical and non-critical actuators accordingly.

[0049] During operation, the task status analysis module 114 determines the current task stage based on motion commands, robot status, and environmental information; the task risk assessment module 115 identifies critical actuators based on the impact of the output decrease of each actuator on the current main task; the control module 105 generates migration commands based on the available power of the system, the target power of critical actuators, and the releaseable power of non-critical actuators, and the branch power reconfiguration is jointly completed by the power management module 102 and the actuator drive module 112.

[0050] This embodiment establishes a closed loop for task analysis and risk assessment between the multimodal motion mechanism and the power system, so that power scheduling is no longer determined solely by battery voltage or fixed priority, thus forming a feedback path from the system structure to the task objective on energy allocation.

[0051] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment, such as... Figure 1 and Figure 3 As shown, it includes the following steps: Step 1: Task status analysis module 114 and dynamic identification of key actuators.

[0052] This embodiment illustrates one implementation of the task status analysis module 114 and the task risk assessment module 115.

[0053] Step 101: Obtain the current task state, robot motion state, actuator state (target output and actual output of each actuator), and energy state of the multimodal robot.

[0054] The current task status includes at least one of the following: task instructions, target motion mode, environmental constraints, and task objective.

[0055] Specifically, the task state analysis module 114 receives task instructions, environmental perception results, posture information, speed information, height information, foot contact information, and the working status of each motion module, and constructs the current task state vector: ; Where m represents the target motion mode, s represents the task stage, g represents the task objective, q represents the robot posture state, v represents the velocity state, h represents the height state, c represents the contact state, and e represents the environmental constraint state.

[0056] The motion state includes at least one of attitude angle, angular velocity, velocity, acceleration, height, and foot contact state.

[0057] The energy status includes bus voltage, allowable output current, battery continuous output capacity, power conversion module limits, and safety reserved power.

[0058] Step 102: Based on the current task status and motion status, determine the current task stage and at least one task maintenance indicator corresponding to the current task stage.

[0059] The task status analysis module 114 can use a state machine, rule table, decision tree, probability model or trained classification model to determine the current task stage.

[0060] For flying, crawling, and jumping robots, the mission phases can be divided into ground approach, jump energy storage, jump preparation, takeoff transition, aerial assistance, attitude adjustment, landing cushioning, and post-landing recovery. These phase divisions are used to characterize changes in power requirements and key execution capabilities, and do not need to be exactly the same as the specific switching conditions in the robot motion control patent.

[0061] The current mission phase includes at least three consecutive phases: ground approach phase, hop energy storage phase, takeoff transition phase, air support phase, attitude adjustment or landing preparation phase, and landing recovery phase; correspondingly, the key actuators dynamically migrate between the crawling module, hop module, flight module, and attitude adjustment mechanism as the mission phase progresses.

[0062] The mission retention metrics include at least one of the following: crawl speed, obstacle crossing energy storage, energy storage completion time, flight actuator rotation speed, flapping frequency, flight thrust, attitude angle error, angular velocity, expected crossing distance, and landing attitude.

[0063] Step 103: Calculate the task risk of each actuator and identify critical actuators, restricted actuators and non-critical actuators.

[0064] Specifically, for the current task phase, we analyze the necessity of each actuator, the risk of task failure due to the decline in the output capability of each actuator, the deviation between the target output and the actual output of each actuator, and the substitutability of each actuator. We conduct a task risk assessment for each actuator and determine the critical and non-critical actuators based on the results of the task risk assessment.

[0065] like Figure 5 As shown, the task risk assessment module 115 calculates the task risk assessment value R for the i-th actuator. i The evaluation value can be determined based on the necessity of the executor for the current task, the degree of task failure caused by the lack of executor output, the deviation of the current output, and the substitution capability.

[0066] In one implementation, the evaluation value can be expressed as: ; Where, N i F indicates the degree to which the executor is necessary for the current task. i D indicates the risk of primary task failure due to actuator failure. i H represents the normalized deviation between the actuator's target output and its actual output. i This indicates the impact of the actuator output decrease on the robot's safety status, B. i This indicates the ability of other actuators to substitute for its function. , , , , These are the weighting coefficients.

[0067] In other implementations, remaining mission time, estimated energy consumption, environmental hazard level, and energy required to return to a safe area can be incorporated into the mission risk assessment to coordinate between immediate mission sustainment and overall mission endurance.

[0068] In other embodiments, the task risk assessment module 115 can use a rule-based risk matrix, fuzzy inference, Bayesian networks, model prediction results, or a trained machine learning model to determine the actuator risk level. Regardless of the algorithm used, as long as its input includes at least the current task state and the actuator output state, and is used to determine the power protection object, the task-maintaining power migration of the present invention can be realized.

[0069] Task risk assessment module 115 will R i The task risk assessment value R is compared with at least one risk threshold. i With the critical threshold R H Non-critical threshold R L Comparison, where R H Greater than R L When R i Above the critical threshold R H And the duration exceeds the threshold T hold When R is selected, the corresponding actuator is identified as the critical actuator; when R is selected, the corresponding actuator is identified as the critical actuator. i Below the non-critical threshold R L And the duration exceeds the threshold T hold When the threshold is between two thresholds, the corresponding actuator is identified as a non-critical actuator; the actuator is identified as a restricted actuator, whose power is allowed to be reduced but must not be lower than the minimum holding power.

[0070] To avoid frequent switching of critical actuators near the threshold, this embodiment employs hysteresis judgment and minimum hold time. A lower entry threshold is used when an actuator switches from a non-critical state to a critical state, and a higher exit condition is used when exiting from a critical state, or the exit condition is required to be met for multiple consecutive control cycles. That is, the switching between critical, restricted, and non-critical actuator levels uses hysteresis judgment composed of the entry threshold, exit threshold, number of consecutive judgments, and / or minimum hold time to suppress frequent switching of actuator levels near the threshold.

[0071] like Figure 4 As shown, during the ground approach phase, the crawling module 106 has a higher task necessity and risk of failure, and is therefore identified as a critical actuator; the flight module 108 does not participate in the current main task and can be identified as a non-critical actuator. After the robot leaves the ground, the foot contact state disappears and the altitude increases. The task state analysis module 114 updates the task phase to the aerial assistance phase. At this time, the task risk assessment values ​​of the flight module 108 and the attitude adjustment mechanism 109 increase, and they become critical actuators, while the crawling module 106 becomes a non-critical actuator.

[0072] The aforementioned dynamic identification process enables the same actuator to change its power guarantee level according to the mission phase, thereby avoiding energy waste caused by permanently setting the flight module, jump module, or crawl module as the highest priority.

[0073] Step 2: Estimation of available power and generation of mission-sustaining power budget.

[0074] This embodiment illustrates how the current available power and the power budget for each actuator are determined.

[0075] Specifically, the current available power is estimated based on the energy state, and the minimum holding power of the critical actuators and the release power of the non-critical actuators are determined.

[0076] The control module 105 determines the system's currently allocable power (currently available power) based on the bus voltage, allowable output current, battery continuous output capacity, power conversion module limits, and safety reserve power. ; Among them, P ava U represents the currently allocable power. bus I represents the bus voltage. lim P represents the allowable current of the busbar. bat P represents the permissible output power determined based on the battery's state of charge, temperature, and internal resistance. conv Indicates the allowable output power of the power conversion module, P safe This refers to the safety power reserved for the control module 105, the sensing module 110, and the necessary communication functions (the reserved power corresponding to the control, necessary sensing, and safety functions).

[0077] When the system does not have independent battery power estimation or conversion module power limits set, U can be used. bus I lim Subtract P safe As a simplified estimate.

[0078] For the i-th actuator, determine the minimum holding power P. i,min Target power P i,tar and maximum allowable power P i,max P, the key actuator i,min Based on the current task retention threshold (task retention metric), the P of non-critical actuators is determined. i,min The P of auxiliary loads that can be turned off can be used to maintain the power required for standby or safe operation. i,min It can be zero.

[0079] The control module 105 first reserves power for control, sensing and safety functions, then allocates minimum holding power to key actuators, and then allocates incremental power from the remaining power based on the task risk assessment value and output deviation.

[0080] Power budget (reference power) for each actuator satisfy: ,and ; like Figure 7 The figure shows the power budget changes of each actuator at different stages of a continuous task.

[0081] When the available power is higher than the sum of the target power of all actuators, each actuator can operate at the target power. When the available power is lower than the sum of the target power of all actuators but higher than the sum of the safety reserve power and the minimum holding power of critical actuators, the control module 105 compresses the budget of non-critical actuators and restricted actuators to meet the holding power of critical actuators. When the available power further decreases and cannot meet the minimum holding power of all critical actuators at the same time, the task risk assessment module 115 re-compares the risk levels of multiple task objectives and enters the task degrade or safe exit state.

[0082] In one alternative implementation, the remaining power P, excluding the minimum holding power, is... rem Weighted allocation: ; ; Among them, P rem This represents the remaining allocable power after satisfying control, necessary sensing, safety functions, and the minimum holding power of each actuator. To avoid double-weighting factors such as task necessity and output deviation, the remaining power allocation weight can be determined based on the task risk assessment value. ; ; and satisfy , where n represents the number of actuators participating in the remaining power allocation.

[0083] It should be noted that P rem To satisfy the minimum holding power requirement, the remaining allocable power is compared with the release power ΔP of non-critical actuators. j,rel The meanings are different; the total releasable power of multiple selected non-critical actuators is .

[0084] When there are at least two critical actuators and the current available power is insufficient to meet the target power of all critical actuators, the following steps are taken in sequence: allocate safety reserve power for control and necessary sensing functions, allocate minimum holding power for critical actuators to avoid robot instability or loss of control, allocate minimum holding power for critical actuators corresponding to the current main task, and allocate the remaining power according to the task risk assessment value of each critical actuator.

[0085] The above formula is only one implementation method. In actual control, priority queues, linear programming, model predictive control, or lookup table allocation can also be used, but all of them should meet the constraints of prioritizing the minimum holding power of key actuators and ensuring that the total power does not exceed the available power.

[0086] Step 3: Energy competition determination and power migration triggering mechanism.

[0087] This embodiment illustrates how to distinguish between actuator malfunctions and insufficient output of critical actuators caused by energy competition, in order to avoid blindly increasing power for mechanical jamming, drive failures, or sensor anomalies. Figure 6 As shown.

[0088] Specifically, when the actual output of a critical actuator is lower than the corresponding target output and the insufficient output affects the mission maintenance index, the system determines whether the insufficient output is caused by energy competition among multiple actuators based on the system's power supply capacity, the power supply and demand relationship of the critical actuator, and the releaseable power of the non-critical actuator.

[0089] Step 301: Detect whether the output of the key actuator is lower than the target output, and determine whether insufficient output affects task persistence.

[0090] The control module 105 first detects the target output O of the key actuator. k,tar With actual output O k,act The deviation between them is calculated, and the normalized output deficiency is calculated: ; When D k When the output deviation exceeds a preset threshold and continues for a predetermined control period, the task risk assessment module 115 determines whether the insufficient output affects task maintenance. Specifically, at least one task maintenance indicator and its allowable range or maintenance threshold are set for different task stages; when the insufficient output of the key actuator causes the actual value or predicted value of at least one task maintenance indicator to exceed the corresponding allowable range and continues to reach the preset judgment time, it is determined that the insufficient output affects the current task maintenance.

[0091] For example, during the ground crawling phase, when the crawling speed v is lower than the minimum holding speed v min At that time, it can be determined that the ground movement mission is affected; during the jump energy storage phase, when the stored energy E s Below the required energy storage E for the missionreq If the energy storage mechanism fails to reach the target energy storage within the specified time, the jump mission can be deemed affected; during the air assist phase, if the flight actuator speed, flapping frequency, flight thrust, or expected traverse distance is lower than the corresponding minimum hold threshold, the air mission can be deemed affected; during the attitude adjustment or landing preparation phase, if the attitude angle error... Greater than the maximum value and / or angular velocity Greater than the maximum allowed value If this occurs, it can be determined that the attitude maintenance task is affected.

[0092] When all task maintenance indicators remain within the allowable range, maintain the current power allocation and continue monitoring; when it is determined that the insufficient output has affected task maintenance and the task risk assessment value of the critical actuator is higher than the trigger threshold, further determine whether the insufficient output is caused by energy competition among multiple actuators.

[0093] Step 302: Determine whether there is energy competition and whether non-critical actuators have releasable power.

[0094] Energy competition can be determined based on at least two of the following conditions: (1) The bus voltage is lower than the normal operating range or drops significantly when the critical actuator is overloaded; (2) The total current is close to or reaches the allowable current limit; (3) The critical actuator requests more power than it actually receives; (4) The actual power of at least one non-critical actuator is higher than its minimum holding power, and there is releasable power; (5) No stall, over-temperature, open circuit, or other independent fault was detected in the critical actuator drive; (6) After reducing the trial power of non-critical actuators, the output of critical actuators showed a recovery trend.

[0095] Before determining the energy competition among multiple actuators, check whether the critical actuators are stalled, overheated, open-circuited, mechanically jammed, sensor abnormal, or drive abnormal; when the above abnormalities are detected, stop increasing the power to the critical actuators and perform fault handling or safety degradation.

[0096] When critical actuator output is insufficient, task hold risk exceeds a threshold, energy contention exists in the system, and non-critical actuators have releasable power, the control module 105 generates a migration trigger signal. To avoid false triggering caused by transient disturbances, trigger delay, number of consecutive checks, and recovery hysteresis interval can be set.

[0097] If the output of a critical actuator is insufficient but no energy competition is detected, power migration is not blindly executed. Instead, the state is marked as actuator malfunction, abnormal mechanical load, or unreasonable control objective, and fault handling or safety degradation is performed. Therefore, the migration triggering mechanism in this embodiment includes not only insufficient power detection but also a joint determination of task impact and energy competition.

[0098] Step 4: Dynamic power migration and closed-loop correction of key actuators.

[0099] Specifically, when it is determined that the insufficient output is caused by energy competition among multiple actuators, the power budget and / or drive output of at least one of the non-critical actuators is reduced, and the resulting power margin is allocated to the critical actuator; after the power migration is completed, the actual output and task retention indicators of the critical actuator are re-detected, and the power migration is maintained, expanded or canceled based on the detection results, or the task is degraded.

[0100] Step 401: Reduce the power of non-critical actuators and transfer the released power to critical actuators.

[0101] Once the migration triggering condition in step 3 is met, the control module 105 determines the migration target power and the power release source.

[0102] For the critical actuator k, the requested compensation power can be calculated based on the normalized output insufficiency D. k The actuator power-output characteristics and the current allowable power limit are determined as follows: ; Among them, D k This indicates that the normalized output of the critical actuator k is insufficient; This represents the conversion factor between insufficient output of the critical actuator and the required compensation power. This represents the target power of the critical actuator k in the current task phase; This indicates the current actual power received by the critical actuator k; This indicates the maximum allowable input power for the critical actuator k; This represents the requested compensation power for the critical actuator k.

[0103] in, It can be determined based on the actuator's pre-calibrated power-output characteristic curve, lookup table relationships, or online identification results.

[0104] In other implementations, the mapping relationship between the key actuator output state and the input power required to achieve that output state can also be used. Determine the compensation power, that is: ; For a non-critical actuator j, its releaseable power is expressed as: ; in, This represents the power that the j-th non-critical actuator can release without compromising the current task safety state; This represents the current actual power consumption of the j-th non-critical actuator; This represents the minimum holding power required for the j-th non-critical actuator to maintain standby, safe, or minimal auxiliary functions; non-critical loads that can be completely shut down can be set as follows: .

[0105] The control module 105 selects power sources from low to high based on the release cost; that is, it selects non-critical actuators for power release from low to high power release cost. The release cost can be determined based on the impact of load shutdown on the current task, restart time, power surge during recovery, and safety impact. Priority is given to releasing auxiliary lighting, unnecessary calculations, and non-critical communication power; secondly, motion modules not involved in the current task are reduced; and thirdly, actuators that provide auxiliary support to the current task but are redundant are reduced.

[0106] Actual power increment migrated to critical actuators The minimum value among: the requested compensation power of critical actuators, the total releaseable power of non-critical actuators, the current allowable margin of the power system, and the total releaseable power is not exceeded. ; in, This represents the final actual increase in power to the critical actuator k. P represents the requested compensation power for the critical actuator k. margin J represents the unused power margin of the system before power migration, and J represents the set of non-critical actuators selected to release power. This represents the releasable power corresponding to the j-th non-critical actuator.

[0107] ; in, Indicates the currently available power. This indicates the current actual power consumption of each actuator.

[0108] Control module 105 will ΔP k Convert the current limit, torque limit, drive duty cycle, operating frequency, or voltage command to the critical actuator, and reduce the power of the selected non-critical actuator to the new budget value.

[0109] Power migration is performed in stages or using a ramp method. Non-critical actuators first release power at a preset ramp. After the control module 105 confirms that the bus condition has improved, the power of critical actuators is gradually increased to prevent secondary voltage fluctuations caused by simultaneous changes in multiple branches.

[0110] Power migration is achieved through at least one of the following methods: adjusting the power budget of the power supply branch, adjusting the branch current limiting threshold, adjusting the actuator drive current limit, adjusting the torque limit, adjusting the target speed, adjusting the operating frequency, adjusting the PWM duty cycle, adjusting the action amplitude, adjusting the operating duty cycle, and adjusting the startup sequence of multiple actuators; power release and power increase of key actuators are performed in stages according to a preset slope.

[0111] Step 402: Verify the task retention effect and perform maintain, expand, undo migration or gradual recovery.

[0112] After migration, the actuator status detection module 104 continues to acquire the actual output of the critical actuator, and the task risk assessment module 115 recalculates the task risk. When the output of the critical actuator recovers and the task risk decreases, the current migration result is maintained; when the output is still insufficient and there is still releasable power, the migration range is expanded; when there is no releasable power or the output of the critical actuator does not improve with the increase of power supply, the migration stops and the task is degraded or handled.

[0113] This embodiment forms a closed loop that includes output insufficiency detection, energy competition confirmation, power release, critical actuator compensation, and task effect verification, which differs from the open-loop scheme that redistributes power in a fixed proportion at once.

[0114] In this embodiment, power regulation can be achieved by adjusting the branch current limit value of the power management module 102, or by adjusting the motor torque command, speed limit, PWM duty cycle, operating frequency, operating amplitude, or working duty cycle through the actuator drive module 112. For loads that cannot be continuously adjusted, power can be released by intermittent operation, delayed start-up, or shutdown.

[0115] In this embodiment, the control module 105 can adopt a centralized control structure, or each actuator's local controller can report the required power and task risk, with the energy management controller arbitrating the data. Multiple controllers can exchange power budgets, actuator status, and task phase information via bus, serial communication, or wireless communication.

[0116] Explanation 1: Task-maintaining power transfer during the ground crawling phase.

[0117] Taking the robot's task of approaching a target and inspecting the environment on complex terrain as an example, the task status analysis module 114 determines that the robot is in the ground crawling stage based on the continuous contact of its feet, the basic stability of the robot's height, and the validity of the crawling command. The task risk assessment module 115 identifies the crawling module 106 as a critical actuator, and identifies the jumping module 107 and the flight module 108 as non-critical actuators or standby actuators.

[0118] Under normal power conditions, the crawling module 106 operates at the target speed, the sensing module 110 and the communication module 111 maintain normal sampling and communication frequencies, the flight module 108 is in a shut-off or low-power standby state, and the jumping module 107 only maintains locking and safety monitoring.

[0119] When the robot enters soft ground, slopes, or areas with dense obstacles, the load on the crawling motors increases, and the actual speed is lower than the target speed. If the energy status detection module 103 simultaneously detects that the total current is close to the limit and finds that the auxiliary load module 113, unnecessary environmental sampling, or other modules in standby heating state are consuming power, then the task risk assessment module 115 determines that the continuous mobility is affected by energy competition.

[0120] The control module 105 first reduces the brightness of the auxiliary lighting, reduces the frequency of non-critical image processing, or suspends the uploading of data unrelated to the current motion; if the released power is still insufficient, it reduces the pre-stored power of the jump module 107 or shuts down the standby drive circuit of the flight module 108; the released power is used to increase the upper limit of the allowable current or torque of the crawling motor, so that the robot's crawling speed is restored to above the task holding threshold.

[0121] In this embodiment, the crawling speed does not need to fully recover to the original target value; as long as it is higher than the minimum speed required to continuously traverse the current area, the task is considered successful. Therefore, the system avoids further insufficient crawling motor torque caused by averaging down the entire load when energy is limited.

[0122] When the ground load decreases or the power status is restored, the system does not immediately restore all auxiliary loads at the same time. Instead, it first confirms that the creep speed and current are within a stable range, and then gradually restores the power of data processing, lighting and standby actuators in a low-impact sequence.

[0123] Explanation 2: Key power guarantee during the jump energy storage and obstacle crossing phases.

[0124] Taking the robot's transition from crawling to jumping obstacle-crossing as an example, the task status analysis module 114 determines that the robot has entered the jumping energy storage phase based on the obstacle-crossing command, obstacle recognition results, and the preparation status of the jumping module 107. The task risk assessment module 115 identifies the jumping energy storage driver as the key actuator and sets the minimum holding power required to reach the target energy storage within a specified time.

[0125] The jumping module 107 can store energy by driving an elastic element with a motor. The control module 105 determines the energy storage progress based on the displacement of the energy storage mechanism, the driving current, the energy storage time, or an estimated elastic potential energy. When the energy storage speed is lower than the task requirements and the total system power is close to the limit, the control module 105 reduces the unnecessary motion output of the crawling module 106, causing the robot to enter a low-speed stable or stopped state; at the same time, it shuts down the unnecessary pre-start of the flight module 108, reduces the power of the auxiliary load module 113, and uses the released power for the jumping energy storage driver.

[0126] During the jump preparation phase, the system reserves safe power for the control, sensing, and jump locking mechanisms, and restricts the simultaneous activation of other high-power actuators. If subsequent missions require the immediate use of flight module 108 after takeoff, the system may reserve some air assist power as jump energy storage nears completion, but the jump energy storage must not fall below the minimum energy required for obstacle clearance.

[0127] At the moment of jump release, this invention does not limit the specific release identification and attitude stabilization algorithm, but only changes the power protection object according to the current mission stage. After the jump release is completed, the mission risk of the jump energy storage driver decreases rapidly, the control module 105 adjusts it from a critical actuator to a non-critical actuator, and releases the power of its drive branch to establish a power supply margin for the flight module 108 and attitude adjustment mechanism 109 in the air phase.

[0128] This embodiment demonstrates how critical actuators migrate with the successive task phases. Instead of permanently setting the skip module as a high priority in advance, it protects the skip module during the skip energy storage phase, and actively removes this protection after release, migrating the power budget to subsequent critical actuators.

[0129] Explanation 3: Power transfer during the in-flight assistance and attitude adjustment phase.

[0130] Taking the robot's entry into aerial motion after completing a jump as an example, the task status analysis module 114 determines that the robot has entered the aerial assistance phase based on the disappearance of foot contact, changes in altitude, and aerial mission commands. The task risk assessment module 115 identifies at least one of the flight module 108 and the attitude adjustment mechanism 109 as a key actuator based on the impact of the flight module 108 output on altitude, horizontal crossing distance, or descent speed, and the impact of the attitude adjustment mechanism 109 output on attitude angle error.

[0131] For flapping-wing flight modules, flapping frequency, drive current, wing flapping amplitude, or estimated aerodynamic force can be used as output states; for rotor flight modules, rotational speed, estimated thrust, or motor current can be used as output states. If the actual output is lower than the target output and the bus voltage drops and the total current approaches the limit, the system determines that there is energy competition.

[0132] Control module 105 shuts down the drive output of crawling module 106, keeping it in a mechanically free state or a low-power braking state; stops the jump energy storage driver from continuing to operate; reduces the power of non-critical communication, lighting, and image processing; and uses the released power to increase the permissible drive power of flight module 108.

[0133] When the attitude angle error exceeds the attitude holding threshold, the mission risk assessment value of the attitude adjustment mechanism 109 is increased. The control module 105 can further perform internal power transfer between the flight module 108 and the attitude adjustment mechanism 109. For example, when the altitude margin is sufficient but the attitude deviation is large, unnecessary thrust increments of the flight module are appropriately limited, and some power is allocated to the attitude adjustment mechanism; when the attitude has recovered but the altitude is decreasing too quickly, the holding power of the attitude adjustment mechanism is reduced, and the power of the flight module is increased.

[0134] The aforementioned power scheduling is based on current mission maintenance metrics, rather than a fixed rule that flight modules or attitude control mechanisms always have priority. The system compares the risks of the two types of critical capabilities using state variables such as altitude, attitude, angular velocity, and expected landing point, allocating limited energy to actuators with higher current risk.

[0135] When the robot is close to landing and the need for flight assistance decreases, the task status analysis module 114 updates the task phase to the landing buffer phase, gradually reduces the power of the flight module, and re-establishes the minimum holding power for the crawling module after landing, thereby completing the continuous migration of the critical power guarantee object.

[0136] Explanation 4: Mission-maintaining graded protection and recovery under abnormal energy conditions.

[0137] like Figure 8 As shown, this embodiment illustrates task retention protection when battery power decreases, bus voltage drops, or output capacity is limited by temperature. The control module 105 sets multiple energy state thresholds U1, U2, U3, and U4 that decrease sequentially, where U1 is higher than U2, U2 is higher than U3, and U3 is higher than U4. Each threshold can be set according to battery type, load characteristics, and safety requirements.

[0138] When the bus voltage is below U1, the system enters Level 1 protection, reducing auxiliary lighting, non-critical communication, and unnecessary computing loads without changing the current critical actuator target. When it is below U2, the system enters Level 2 protection, further limiting the output of non-critical motion actuators and restricted actuators, and increasing the power protection level of critical actuators. When it is below U3, the system enters Level 3 protection, selecting degraded tasks that can reduce the mission target but maintain robot safety, such as reducing crawling speed, shortening flight assistance time, or performing controlled landing ahead of schedule. When it is below U4, the system enters a safe state, reserving the power required for control, necessary sensing, and safe actions, and shutting down other loads.

[0139] Task degradation is not simply a year-on-year reduction in all loads, but rather a selection made by the task risk assessment module 115 based on the current environment and motion state. For example, when the robot is in the air, it should prioritize attitude recovery or controlled landing; when the robot is on the ground, it can choose to stop moving forward and maintain communication; when the robot is jumping to store energy but has not yet reached the safe release condition, it can stop energy storage and maintain locking.

[0140] When the energy state is restored, the control module 105 uses a recovery threshold higher than the protection threshold and requires the voltage, current, or temperature to continuously meet the recovery conditions for a predetermined time. The recovery process proceeds in the following order: control and necessary sensing, current critical actuators, restricted actuators, non-critical actuators, and auxiliary loads.

[0141] The power of each branch is gradually restored according to a preset slope. After each stage of restoration, the control module 105 re-detects the bus voltage and the output of key actuators. If restoring a certain load causes the bus voltage to drop again, subsequent restorations are paused and the system reverts to the previous stable power state. Through threshold hysteresis and gradual restoration, frequent switching between protection and normal states and secondary voltage drops caused by simultaneous restoration of multiple actuators can be prevented.

[0142] Note 5: Integrated application in continuous missions involving flight, crawling, and jumping.

[0143] This embodiment illustrates the application of the present invention in a complete continuous task. The robot first approaches the obstacle area on the ground using the crawling module 106. At this time, the task status analysis module 114 identifies it as the ground approach phase, with the crawling module 106 acting as the key actuator, while the jumping module 107 and the flight module 108 remain in standby mode. The control module 105 ensures that the crawling speed does not fall below the approach task holding threshold and limits unnecessary loads.

[0144] When environmental perception indicates that the obstacle ahead cannot be directly crossed by crawling, the mission objective is updated to obstacle crossing by jumping. The mission status analysis module 114 enters the jump energy storage phase, and the mission risk assessment module 115 increases the risk level of the jump module 107. The control module 105 reduces the crawling speed and releases auxiliary load power, transferring the power budget to the jump energy storage driver, enabling the energy storage mechanism to reach the energy required for obstacle crossing within the specified preparation time.

[0145] After the jump release is completed, the jump energy storage driver no longer undertakes the primary task. The mission status analysis module 114 updates the mission phase to the air-assisted phase based on the takeoff status, the control module 105 cancels the high power budget of the jump module, and allocates the power of the crawl module, jump module, and auxiliary load release to the flight module 108 and attitude adjustment mechanism 109.

[0146] If the aerial attitude deviation is small but the expected crossing distance is insufficient, the flight module 108 is identified as the main critical actuator, and the system prioritizes its output; if the robot is disturbed and generates a large attitude angle and angular velocity, the task risk of the attitude adjustment mechanism 109 increases, and the system will transfer some of the flight redundant power to the attitude adjustment mechanism until the attitude is restored to within the holding threshold.

[0147] As the robot approaches landing, the mission objective shifts from aerial traversal to a stable landing. The system gradually reduces the power of the flight module, reserving power for attitude adjustment and necessary sensing. After the feet re-contact the ground, the mission status analysis module 114 enters the landing recovery phase, and the crawling module 106 once again becomes the key actuator. The control module 105 resumes crawling drive in a ramp manner and, after confirming that the robot's attitude and busbar status are stable, resumes communication, image processing, and other auxiliary functions.

[0148] Throughout the continuous mission, the critical actuators sequentially migrate from the crawling module 106 to the jumping module 107, then to the flight module 108 or the attitude adjustment mechanism 109, and finally return to the crawling module 106. Each change in the critical object is accompanied by the release of power in non-critical branches, the compensation of power in critical branches, and the verification of mission hold-up effects.

[0149] This embodiment demonstrates that the present invention can embed energy scheduling into continuous motion tasks, so that limited energy is no longer allocated according to a fixed ratio, but dynamically migrates among multiple actuators as the risk of task failure changes. It should be noted that the above-described flight, crawling, and jumping are merely examples, and the present invention is also applicable to wheel-leg switching, air-ground switching, swimming and crawling switching, and other multi-actuator robots that include robotic arms, mobile chassis, and attitude stabilization mechanisms.

[0150] The dynamic power migration method for key actuators of multimodal robots provided in this embodiment changes the robot's energy management objective from simply reducing power consumption or protecting the power system to maintaining the ability to complete the current task under limited energy constraints, thus directly linking energy scheduling with the robot's task objectives.

[0151] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment can dynamically identify key actuators based on the motion task and the risk of task failure, avoiding energy waste or insufficient guarantee of key capabilities caused by using fixed priorities.

[0152] The dynamic power transfer method for critical actuators of multimodal robots provided in this embodiment identifies energy competition among multiple actuators and implements power transfer, so that the power resources released by non-critical actuators are preferentially supplied to critical actuators, thereby improving the reliability of tasks in critical stages such as jumping, flight assistance, attitude adjustment and landing recovery.

[0153] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment prioritizes maintaining control, perception, and key motion capabilities under abnormal energy conditions, while degrading non-critical functions. Compared with overall load reduction or direct shutdown, this method can improve task continuity and system safety in complex environments.

[0154] The dynamic power transfer method for key actuators of multimodal robots provided in this embodiment does not depend on specific robot mechanical structures for key actuator identification, task risk assessment, and power transfer mechanisms. It can be extended to multimodal biomimetic robots that fly, climb, and jump, air-ground collaborative robots, legged robots, and other unmanned systems that contain multiple heterogeneous actuators.

[0155] This embodiment provides a dynamic power transfer method for critical actuators in multimodal robots. It acquires the task state, motion state, target and actual output of the actuators, and energy state. Based on the task stage and the risk of task failure, it dynamically determines critical and non-critical actuators. When insufficient output from a critical actuator affects task retention metrics and is caused by energy competition among multiple actuators, it reduces the power budget or drive output of non-critical actuators, allocating the resulting energy margin to the critical actuators. Based on the transferred output and task retention results, it maintains, expands, or cancels the transfer, or performs task degradation. This method is applicable to multimodal robots capable of flight, crawling, jumping, and attitude adjustment, ensuring critical motion capabilities and continuous task execution under limited energy conditions.

[0156] Example 2 The multimodal robot key actuator dynamic power transfer system provided in this embodiment includes: a robot motion execution system, a sensing module, a task status analysis module, a task risk assessment module, an energy status detection module, a control module, and an actuator drive module.

[0157] The robot motion execution system includes at least two of the following: a crawling module, a jumping module, a flying module, and an attitude adjustment mechanism, to enable the robot to perform various motions under different environmental conditions.

[0158] The sensing module is used to acquire at least one of the following: robot attitude angle, angular velocity, acceleration, height, speed, foot contact state, actuator operating state, and environmental state, providing input for task analysis and performance evaluation.

[0159] The task state analysis module is used to determine the current task type, task stage, and key motion capabilities required to complete the task based on the robot's current task instructions, environmental state, motion stage, and target motion mode. Unlike conventional state recognition, which only determines the robot's motion state, the task state analysis system further determines the execution capabilities necessary to complete the current task.

[0160] The task risk assessment module evaluates the task risk level of each actuator based on its contribution to the current task, the deviation between the actuator's actual output and the target output, and the impact of actuator performance degradation on task completion. Actuators with a high task risk level and whose output capability directly affects the completion of the current task are identified as critical actuators; other actuators with a smaller impact on the current task or that can be temporarily downgraded are identified as non-critical actuators.

[0161] The energy status detection module is used to acquire at least one of the following: battery voltage, bus voltage, total current, current of each actuator, current available power, and energy safety margin, and to determine whether the system has insufficient power supply or energy competition among multiple actuators.

[0162] The control module is used to identify non-critical actuators from which power can be released when the output of a critical actuator is lower than the target output, insufficient output affects task completion, and energy competition exists in the system. It then reduces the output power, operating frequency, startup slope, duty cycle, or continuous operating time of the non-critical actuators and reallocates the released power budget to the critical actuators. After the migration is complete, the system re-detects the output of the critical actuators and the robot's task status, and maintains, increases, or cancels the power migration based on the task retention results.

[0163] Based on the power migration decision results, the actuator drive module implements current limiting, power limiting, PWM duty cycle adjustment, target speed adjustment, target frequency adjustment, startup timing adjustment, or power recovery control for each actuator.

[0164] Furthermore, when the energy status is abnormal but the current task still needs to be maintained, priority is given to maintaining the power supply to the control module, sensing module and current critical actuators, while reducing the power of auxiliary loads and non-critical actuators; when the energy status is restored, the output is gradually restored in the order of control and sensing module, current critical actuators, non-critical actuators and auxiliary loads, in order to avoid secondary voltage drops.

[0165] It should be noted that each module in this embodiment corresponds one-to-one with each step in Embodiment 1, and their specific implementation processes are the same, so they will not be repeated here.

[0166] Example 3 This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the dynamic power transfer method for key actuators of a multimodal robot as described in Embodiment 1 above.

[0167] Example 4 This embodiment provides a computer device, such as... Figure 9 As shown, the system includes a computer-readable storage medium 1003, a processor 1001, a communication interface 1002, and a computer program stored on the computer-readable storage medium 1003 and executable on the processor 1001. The processor 1001, communication interface 1002, and computer-readable storage medium 1003 can be connected via a bus or other means. The communication interface 1002 is used to receive and send data. When the processor 1001 executes the program, it implements the steps in the dynamic power transfer method for key actuators of a multimodal robot as described in Embodiment 1 above.

[0168] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A dynamic power transfer method for key actuators of a multimodal robot, characterized in that, include: Acquire the current task state, motion state, target output and actual output of each actuator, and energy status of the multimodal robot; Based on the current task status and motion status, determine the current task phase and the corresponding task maintenance indicators; For the current task phase, analyze the necessity of each actuator, the risk of task failure due to decreased output capability, the deviation between target output and actual output, and the substitutable capabilities, conduct a task risk assessment for each actuator, and identify critical and non-critical actuators. Estimate the current available power based on the energy status, determine the minimum holding power and power budget for each actuator; when the actual output of a critical actuator is lower than the target output and the insufficient output affects the mission holding performance, determine whether the insufficient output is caused by energy competition among multiple actuators; When it is determined that insufficient output is caused by energy competition among multiple actuators, the power budget of at least one non-critical actuator is reduced, and the resulting power margin is allocated to the critical actuator. After the power migration is completed, the actual output and task retention metrics of the critical actuators are re-tested, and the power migration is maintained, extended, or canceled, or the task is degraded, based on the test results.

2. The dynamic power transfer method for key actuators of a multimodal robot as described in claim 1, characterized in that, The process of conducting task risk assessments for each actuator and identifying critical and non-critical actuators includes: Calculate the task risk assessment value R for the i-th actuator. i =a1N i +a2F i +a3D i +a4H i +a5B i , where N i F indicates the degree to which the executor is necessary for the current task. i D indicates the risk of primary task failure due to actuator failure. i H represents the normalized deviation between the actuator's target output and its actual output. i This indicates the impact of the actuator output decrease on the robot's safety status, B. i This indicates the ability of other actuators to substitute for its function; a1, a2, a3, a4, and a5 are weighting coefficients. The task risk assessment value R i With the critical threshold R H Non-critical threshold R L Comparison, where R H Greater than R L ; When R i Above the critical threshold R H Furthermore, if the duration exceeds a preset threshold, the corresponding actuator will be identified as a critical actuator; When R i Below the non-critical threshold R L Furthermore, if the duration exceeds a preset threshold, the corresponding actuator will be identified as a non-critical actuator; When R i In R H With R L If the duration exceeds a preset threshold, the corresponding actuator will be identified as a restricted actuator.

3. The dynamic power transfer method for key actuators of a multimodal robot as described in claim 1, characterized in that, The process of estimating the current available power based on the energy state and determining the minimum holding power and power budget for each actuator includes: The currently allocable power is determined based on the bus voltage, allowable output current, battery continuous output capacity, power conversion module limits, and safety reserve power: P ava =min(U bus I lim ,P bat ,P conv )-P safe Among them, P ava U represents the currently allocable power. bus I represents the bus voltage. lim P represents the allowable current of the busbar. bat P represents the permissible output power determined based on the battery's state of charge, temperature, and internal resistance. conv Indicates the allowable output power of the power conversion module, P safe This refers to the safety power reserved for control, necessary sensing, and safety functions. For the i-th actuator, determine the minimum holding power P. i,min Among them, the key actuator P i,min Based on the mission retention metric, the P of non-critical actuators is determined. i,min To maintain the power required for standby or safe operation, the auxiliary load that is allowed to be shut down has a P i,min Zero; For the i-th actuator, the power budget Of which, the remaining power excluding the minimum holding power Remaining power allocation weight Maximum task risk assessment value , is the task risk assessment value, and n is the number of actuators participating in the remaining power allocation.

4. The dynamic power transfer method for key actuators of a multimodal robot as described in claim 1, characterized in that, Detect the target output O of the key actuator k,tar With actual output O k,act The deviation between them is calculated, and the normalized output deficiency D is calculated. k =max(0,O) k,tar -O k,act ) / O k,tar When D k When the output deviation exceeds the preset threshold and continues for a predetermined control period, it is determined whether the insufficient output affects the task maintenance index. If the task risk assessment value of the key actuator is higher than the trigger threshold, it is determined whether the insufficient output is caused by energy competition among multiple actuators.

5. The dynamic power transfer method for key actuators of a multimodal robot as described in claim 1, characterized in that, Whether the insufficient output is caused by energy contention among multiple actuators is determined based on at least two of the following conditions: The bus voltage is lower than the normal operating range or drops when the critical actuator is overloaded; The total current is close to or has reached the permissible current limit; The critical actuator requested more power than it actually received. At least one non-critical actuator has an actual power higher than its minimum holding power and has releasable power; No stall, over-temperature, open circuit, or independent fault was detected in the critical actuator driver; After reducing the trial power of non-critical actuators, the output of critical actuators showed a recovery trend.

6. The dynamic power transfer method for key actuators of a multimodal robot as described in claim 1, characterized in that, The current mission phase includes at least three consecutive phases: ground approach phase, jump energy storage phase, takeoff transition phase, air support phase, attitude adjustment or landing preparation phase, and landing recovery phase.

7. The dynamic power transfer method for key actuators of a multimodal robot as described in claim 1, characterized in that, The power migration is achieved by at least one of the following methods: adjusting the power budget of the power supply branch, adjusting the branch current limiting threshold, adjusting the actuator drive current limit, adjusting the torque limit, adjusting the target speed, adjusting the operating frequency, adjusting the PWM duty cycle, adjusting the action amplitude, adjusting the operating duty cycle, and adjusting the startup sequence of multiple actuators.

8. A dynamic power transfer system for key actuators of a multimodal robot, characterized in that, include: The sensing module is configured to acquire the current task state, motion state, target output and actual output of each actuator, and energy status of the multimodal robot. The task status analysis module is configured to determine the current task stage and the corresponding task maintenance index based on the current task status and motion status. The task risk assessment module is configured to: for the current task stage, analyze the necessity of each actuator, the risk of task failure due to decreased output capability, the deviation between target output and actual output, and the substitutable capability, conduct task risk assessment for each actuator, and identify critical and non-critical actuators. The energy status detection module is configured to: estimate the current available power based on the energy status, determine the minimum holding power and power budget of each actuator; and determine whether the insufficient output is caused by energy competition among multiple actuators when the actual output of a critical actuator is lower than the target output and the insufficient output affects the task holding index. The control module is configured to: when it is determined that insufficient output is caused by energy competition among multiple actuators, reduce the power budget of at least one non-critical actuator and allocate the resulting power margin to the critical actuator; After the power migration is completed, the actual output and task retention metrics of the critical actuators are re-tested, and the power migration is maintained, extended, or canceled, or the task is degraded, based on the test results.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the dynamic power transfer method for key actuators of a multimodal robot as described in any one of claims 1-7.

10. A computer device comprising a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the dynamic power transfer method for key actuators of a multimodal robot as described in any one of claims 1-7.

Images