A robot dynamic energy efficiency optimization system and method

Abstract

The application provides a robot dynamic energy efficiency optimization system and method, and belongs to the field of mobile robots. The system comprises: a high-precision energy consumption modeling module, which is used for collecting multi-dimensional data of the robot in real time, calculating real-time total power consumption and remaining endurance of the robot, and generating an energy consumption cost map; a dynamic energy efficiency path planning module, which is used for receiving a task target and a real-time map containing obstacles, and generating an energy efficiency optimal path by using the energy consumption cost map; and an adaptive charging decision module, which is used for making hierarchical charging decisions according to real-time power, remaining endurance and a task queue of the robot. The system can realize real-time high-precision energy consumption perception, deeply integrate energy consumption into path decision, and actively perform energy scheduling based on prediction, so that the endurance and overall energy efficiency of the mobile robot are fundamentally improved.

Description

Technical Field

[0001] This invention belongs to the field of mobile robots, specifically relating to a robot dynamic energy efficiency optimization system and method. Background Technology

[0002] Mobile robots, as core transportation and operation platforms in fields such as intelligent manufacturing, smart logistics, and unmanned inspection, directly impact the operational efficiency and reliability of the entire system through their continuous operating capability. Battery life, i.e., the amount of work or distance a robot can complete on a single charge, is a key performance indicator. However, limited by the energy density of current battery technology, simply increasing battery capacity to improve battery life leads to increased robot weight, higher costs, and greater space requirements for payload. Therefore, maximizing energy efficiency and extending effective operating time through advanced energy management strategies based on existing hardware has become a critical technological challenge that urgently needs to be addressed in the field of mobile robots.

[0003] Currently, most mainstream technologies for improving robot battery life have limitations such as being isolated, static, and passive, failing to form a systematic energy efficiency optimization loop. Specifically, these limitations manifest in the following aspects:

[0004] 1. The energy consumption estimation model is coarse and cannot reflect dynamic operating conditions. Traditional robot energy management typically relies on simple estimations based on average power consumption or empirical models based on historical data. For example, a system might use a fixed "power consumption per kilometer" constant, combined with remaining battery power, to roughly estimate remaining range. This method severely ignores the dynamic variables in actual robot operation. A robot's instantaneous power consumption is significantly affected by real-time load, acceleration / deceleration states, road surface characteristics (e.g., the significant difference in friction coefficients between smooth and rough surfaces), slope, and additional equipment used to adapt to the environment (e.g., air conditioning, special-purpose tools). During uphill climbing, heavy acceleration, or adverse road conditions, actual power consumption can be several times higher than the average estimate, leading to severely optimistic range predictions based on fixed models. This can easily cause the robot to stall due to battery depletion mid-task, resulting in work interruptions, scheduling chaos, and even safety risks. Current technology lacks a physical energy consumption model capable of fusing multi-source sensor data (e.g., IMU acceleration, pressure sensor load, and visually recognized ground type) for real-time, high-precision calculations.

[0005] 2. Path planning focuses on the geometric shortest path, neglecting energy consumption. Existing robot navigation and path planning algorithms (such as Dijkstra's algorithm, A* algorithm, and their variants) generally prioritize minimizing the geometric length of the path or the estimated time as a single or primary optimization objective. Although some advanced algorithms consider indirectly influencing energy consumption through speed optimization, they do not embed "energy consumption" itself as a direct, quantifiable cost dimension into the path search decision function. This leads to the planned "shortest path" potentially being the least energy-efficient path. For example, a seemingly shorter path might include steep inclines, sharp bends, or high-friction surfaces, resulting in significantly higher energy consumption than a slightly longer but flatter, smoother path. Because existing systems lack the ability to quantify the energy cost of terrain (i.e., an "energy cost map"), their planners cannot "know" the differences in energy consumption between different paths and therefore cannot make the most energy-efficient decision.

[0006] 3. The charging strategy is reactive and lacks proactive scheduling. In terms of charging management, most current robotic systems employ a passive response strategy based on fixed thresholds. Typically, a low battery threshold is set (e.g., SOC below 20%). When the battery level drops to this threshold, the robot interrupts its current task and searches for a charging station. This strategy has significant drawbacks: First, the "critical response" mode is high-risk, as the robot may run out of power on its way back to a charging station or at a location far from one. Second, it leads to low operational efficiency; frequent charging interruptions disrupt the continuity of the task queue, and charging may occur during peak task periods, failing to coordinate with the work rhythm. Finally, it cannot utilize fragmented time; even if returning to a charging station to replenish a small amount of power during brief intervals between tasks would be a net benefit, the system will not proactively utilize it. Current technology lacks an intelligent decision-making mechanism that can proactively and forward-lookingly schedule charging behavior based on accurate remaining range prediction, future task requirements, and charging cost-benefit analysis.

[0007] 4. The functional modules are isolated and have not formed a collaborative optimization loop. More importantly, the aforementioned energy consumption estimation, route planning, and charging management functions are often isolated or weakly coupled modules in existing technologies. Coarse energy consumption estimation cannot provide effective input for route planning; route planning that does not consider energy consumption cannot provide an optimization basis for global energy scheduling; and passive charging strategies operate independently of the former two. This fragmented state prevents the system from achieving optimal global energy efficiency. For example, even with an accurate energy consumption model, its value cannot be realized if the planner does not use this information; even if an energy-saving route is planned, incorrect scheduling may still occur if charging decisions are not based on the updated and accurate range of this route.

[0008] In summary, existing technologies are insufficient to meet the demands of mobile robots for high reliability and high efficiency in continuous operation in complex and dynamic environments. Therefore, there is an urgent need for a systematic solution capable of real-time, high-precision energy consumption perception, deep integration of energy consumption into path decision-making, and proactive energy scheduling based on prediction. This would overcome the aforementioned limitations and fundamentally improve the endurance and overall energy efficiency of mobile robots. Summary of the Invention

[0009] The purpose of this invention is to overcome the shortcomings of the prior art and provide a system and method that integrates real-time energy consumption modeling, dynamic path planning and adaptive charging decision-making.

[0010] This invention is achieved through the following technical solution: A first aspect of the present invention provides a robot dynamic energy efficiency optimization system, the system comprising: The high-precision energy consumption modeling module is used to collect multi-dimensional data of the robot in real time, calculate the robot's real-time total power consumption and remaining battery life, and generate an energy cost map. The dynamic energy efficiency path planning module is used to receive the task objective and a real-time map containing obstacles, and to generate an energy-efficient path using the energy cost map. The adaptive charging decision module is used to make tiered charging decisions based on the robot's real-time battery level, remaining range, and task queue.

[0011] Furthermore, the multi-dimensional data includes: power system data, motion state data, and load and environmental data.

[0012] Furthermore, the real-time total power consumption and remaining range of the computing robot are calculated by using a physical power consumption model to decompose the total power consumption into basic fixed power consumption, motion power consumption and environmental power consumption, and the remaining range is dynamically calculated based on the real-time speed.

[0013] Furthermore, the calculation of the motion power consumption takes into account the following factors: The robot's mass parameters include its own weight and real-time dynamic load. Motion state parameters, including acceleration and velocity; Environmental parameters, including slope angle and ground friction coefficient; System efficiency parameters include transmission efficiency, motor efficiency, and air resistance.

[0014] Furthermore, the dynamic energy efficiency path planning module adopts a multi-objective cost function, which accumulates the energy consumption cost weights of all grids traversed by the path, and obtains the energy efficiency optimized path through forward simulation and integral calculation in combination with the speed curve. The energy cost is obtained by accumulating the weight values ​​of the areas traversed by the planned path in the energy cost map.

[0015] Furthermore, the dynamic energy efficiency path planning module also includes a velocity curve optimization module, which is used to generate a smooth velocity curve to match the generated energy efficiency optimized path, so as to limit acceleration and jerk.

[0016] Furthermore, the adaptive charging decision module executes hierarchical charging decision logic, including: performing safety charging judgment, task continuity assurance judgment, and energy efficiency optimization charging judgment in sequence through three layers of decision-making.

[0017] Furthermore, the energy efficiency optimization charging judgment includes: when there is a task gap time window, deciding whether to perform fragmented charging based on the comparison result of the energy consumption of going to and from the charging station and the energy that can be charged within the time window.

[0018] Another aspect of the present invention provides a robot endurance optimization method, which optimizes the endurance of a mobile robot through the robot endurance optimization system. The method includes the following steps: Step 1: Collect multi-dimensional data during robot operation in real time, build a physical power consumption model to calculate the robot's total power consumption and remaining battery life, and generate an energy cost map; Step 2: Based on the task objectives, real-time map, and energy cost map, perform dynamic energy efficiency path planning; Step 3: Make tiered charging decisions based on remaining range, task queue, and charging station location.

[0019] Furthermore, the operation of making tiered charging decisions includes: Determine if the robot's real-time battery level is below the safety threshold; if so, perform emergency charging. If not, determine whether the dynamic remaining battery life can support the completion of the next task. If not, perform preventative charging. If yes, determine whether there is a fragmented time window between tasks. If not, execute the next task. If so, assess whether the energy gain from fragmented charging is greater than the energy consumed in traveling to and from the charging station. If so, execute fragmented charging; otherwise, proceed to the next task.

[0020] The beneficial effects of this invention are: 1. Significantly improved range prediction accuracy: Due to the adoption of a real-time physical power consumption model based on multi-dimensional dynamic parameters, the range estimation error can be reduced compared with the traditional voltage estimation method, providing a reliable data foundation for planning.

[0021] 2. Overall operation energy consumption is effectively reduced: By deeply integrating energy consumption items into the path cost function and optimizing the speed curve, energy efficiency is optimized from both path selection and motion control levels.

[0022] 3. Maximizing Effective Working Time: The prediction-based proactive charging strategy utilizes fragmented charging during task intervals, avoiding long interruptions caused by low battery levels. This improves robot online rate and significantly increases task throughput. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the interaction of the core modules of this invention.

[0024] Figure 2 This is the execution flowchart of the adaptive charging decision module in this invention. Detailed Implementation

[0025] The present invention will now be described in further detail with reference to the accompanying drawings.

[0026] Example 1: Robot Endurance Optimization System The system of this invention runs on a robot local controller and a central server, with three core modules as follows: Figure 1 As shown, the system structure and workflow are as follows: 1. High-precision energy consumption modeling module The high-precision energy consumption modeling module is the core perception and computing foundation of this system. Its core invention lies in the real-time and dynamic estimation of the robot's precise power consumption and remaining range through multi-source sensor data fusion and refined physical modeling, thereby providing key data support for subsequent intelligent decision-making.

[0027] This module collects multi-dimensional data from the robot itself in real time, including: Powertrain data: Battery voltage / current (BMS), motor current (driver); Motion state data: triaxial acceleration / angular velocity (IMU), real-time velocity (encoder), used to quantify the robot's dynamic behavior.

[0028] Load and environmental data: load mass (pressure sensor), ground type and slope (visual / LiDAR or pre-stored map). The ground type is matched to the rolling friction coefficient using a predefined lookup table (Table 1).

[0029] Table 1

[0030] In the actual operation of the system, the first step is to input the above data into the physical energy consumption model for power consumption calculation.

[0031] The model decomposes total power consumption into basic power consumption, motion power consumption, and environmental power consumption.

[0032] Motion power consumption ( The model is defined by the following physical formulas: (1) in, Robot's own weight + load capacity, with the load capacity being a real-time dynamic load capacity; Acceleration (measured by IMU) quantifies the instantaneous high power consumption during acceleration / deceleration, explaining why sudden start-stop cycles consume power; Gravitational acceleration; :speed; Slope angle Accurately calculate the work done by the gravitational component when climbing / descending a slope; The coefficient of rolling friction on the ground transforms the abstract type of ground into a calculable physical quantity, explaining why energy consumption differs on different road surfaces. These are coefficients such as transmission efficiency, motor efficiency, air resistance coefficient, and comprehensive coefficient of loss factors related to high-order speeds, which were calibrated through experiments.

[0033] This model incorporates multiple dynamic variables, such as mass, acceleration, velocity, slope, and ground friction, into a unified calculation to achieve real-time and precise estimation of energy consumption during robot movement.

[0034] Basic fixed power consumption ( The specific component of this component is the power consumed by the robot's main control, sensors, and other basic circuits. In model processing, this power consumption can be regarded as a constant. That is, the model estimates the underlying hardware power consumption necessary for the robot to complete its task, which does not change with motion state (such as speed and acceleration) and working environment (such as slope and road surface), by uniformly reducing it to a fixed value.

[0035] Environmental power consumption ( ): This refers to the power consumption generated by the dedicated functional modules activated by the robot to perform specific tasks. It includes not only the air conditioning or heat dissipation system to maintain the internal environment (such as the temperature and humidity inside the cabinet), but also related equipment that performs core tasks, such as the servo motor of the robotic arm, the drive unit of the gripper, and the high-power sensing module (such as LiDAR) for precision detection.

[0036] Total power consumption is calculated using the following formula: (2) The second step is to estimate the remaining battery life. The physical energy consumption model achieves real-time, dynamic estimation of remaining range through the following steps and formulas: (3) Wherein, SOC: the real-time state of charge of the battery, from the BMS; Battery_Capacity: the total capacity of the battery; P_total: the total power consumption calculated in real time by the high-precision energy consumption modeling module; v: the real-time speed of the robot.

[0037] The core of this estimation lies in the fact that P_total is dynamically changing, integrating multi-dimensional sensor data such as the robot's real-time load, acceleration, ground friction, and slope. Therefore, when the robot is climbing a slope, accelerating, or traveling on a rough surface, P_total will increase, the energy consumption per unit distance will increase, and the remaining driving range calculated by the system will decrease accordingly; conversely, on downhill, constant speed, or smooth surfaces, the estimated remaining driving range will be more optimistic.

[0038] The real-time power consumption, dynamic remaining battery life, and energy cost map output by this module support the path planning module. This map, built on a basic grid map, assigns an energy cost weight to each grid cell. .

[0039] 2. Dynamic Energy Efficiency Path Planning Module The dynamic energy efficiency path planning module is the intelligent decision-making center of the system of this invention. Its core invention lies in the deep integration of real-time energy consumption prediction and traditional path planning to achieve global energy efficiency optimization in complex dynamic environments.

[0040] This module receives the task objective and a real-time map containing obstacles from upstream. The real-time map includes an energy cost map generated by the high-precision energy consumption modeling module, assigning a weight value representing the energy consumption level of each passable area.

[0041] Based on the above inputs, the module performs path optimization using a multi-objective cost function. This function, for the first time, treats "energy consumption cost" as a core optimization objective of equal importance to "path length" and "travel time." Its cost function is as follows:

[0042] in, Configurable weights, configured as follows: Dynamically adjust based on remaining power (SOC) When the SOC is low, the energy consumption weight is automatically increased. The value of makes the robot more inclined to choose energy-saving paths and prioritize returning to the charging station, even if the path is longer or slower.

[0043] Dynamically adjusted based on task urgency For high-priority tasks, the time weight β can be temporarily increased, allowing the selection of shorter paths.

[0044] Set as the baseline weight (e.g., fixed at 1), and and Dynamically scale around it. Simultaneously, set a total weight and constraints or... The minimum threshold. When or When task requirements increase significantly, the system automatically reduces them proportionally. But still guarantee The path length should not be lower than a certain minimum value to ensure that it always has a certain influence on decision-making and to prevent extremely unreasonable detours.

[0045] Furthermore, the system of this invention learns and optimizes based on historical data: the system can record different paths (different...) , , The actual energy consumption and time under the combination are automatically optimized by methods such as reinforcement learning to adapt to the specific working environment in the long term.

[0046] Users can flexibly set strategies by adjusting weighting coefficients: when Significantly greater than and When the system is in use, it prioritizes finding the "most energy-efficient" path rather than the "shortest" path.

[0047] Finally, all the grids the path passes through... The optimal path is obtained by weighting and combining the velocity curve with forward simulation and integral calculation. Specifically, after generating a smooth velocity curve for a candidate path, the planning algorithm determines the expected velocity and acceleration of the robot at each point on the path. Then, the spatial location information of each path point (corresponding to the friction coefficient and slope of the ground) and the motion state (velocity and acceleration) determined by the velocity curve at that point are input into formula (1). By calculating the instantaneous power consumption of all points on the path and accumulating (or integrating), the total estimated energy consumption cost of the path under the specific velocity curve planning can be obtained.

[0048] In the specific planning process, this invention employs an improved A-search algorithm for path searching. When evaluating each potential path node, this algorithm not only calculates the geometric distance but also queries an energy cost map, simultaneously accumulating the energy cost of the traversed areas. Therefore, it can proactively guide the robot to avoid high-friction surfaces, steep slopes, and sharp bends—high-energy-consuming areas—ultimately generating a path that achieves the optimal balance between distance, time, and energy consumption.

[0049] To further explore energy-saving potential, this invention also integrates a velocity curve optimization module: it generates smooth trapezoidal or S-curve velocity curves for the planned path and strictly limits the thresholds for acceleration and jerk (i.e., the rate of change of acceleration). This smooth control can directly reduce the acceleration value in the motion model, reducing dynamic losses at the control level, and forming a dual energy-saving guarantee with path-level optimization.

[0050] 3. Adaptive charging decision module The adaptive charging decision module establishes a hierarchical, proactive, and preventative intelligent charging decision-making logic. It monitors the current task status and predicts whether the remaining battery life is sufficient to complete the next task or N future tasks. Here, N is a dynamic value, determined dynamically by the current remaining battery life and the cumulative demand of the future task queue. For example, if the remaining battery life is sufficient, only the next task (N=1) might be considered; if battery life is limited, it might need to anticipate 3-5 future tasks (N=3~5) to ensure that execution is not interrupted due to insufficient battery power. Based on global real-time status and task planning, it proactively schedules charging behavior to maximize overall operational efficiency while ensuring task continuity.

[0051] like Figure 2 As shown, after each new task is received, the task queue, current SOC, remaining battery life, and charging station location are input into this module, which will initiate a multi-level decision sequence.

[0052] First layer: Emergency charging The decision-making process first checks whether the battery's real-time state of charge (SOC) is below a preset safety threshold. If so, the current plan is immediately interrupted, and an emergency charging command is initiated. This level serves as the highest priority safeguard mechanism, ensuring that the robot will never be paralyzed on the work site due to depleted power, thus guaranteeing the system's basic safety and reliability.

[0053] Second layer: Preventative charging If the battery's real-time state of charge (SOC) is not lower than the preset safety threshold, it enters the second layer of preventive charging. The module calls the real-time dynamic range data provided by the high-precision energy consumption modeling module to determine whether the predicted range is greater than the total round-trip mileage required to complete the next task.

[0054] If the predicted remaining range is not greater than the total round-trip distance required for the next task, a preventative charging command is proactively initiated (even if the battery is 50% depleted), and preventative charging is performed to ensure that the current task queue can be reliably and continuously executed; if the predicted remaining range is greater than the total round-trip distance required for the next task, the next task is executed directly.

[0055] Third layer: Fragmented charging If the predicted range is greater than the total round-trip distance required for the next task, the module will proceed to the third layer and further examine the task scheduling plan to determine if there is a task gap time window (such as 5 minutes).

[0056] If there is a time window between tasks, the module will perform a cost-benefit analysis: determine whether the power replenishment benefit (range gain) obtained by going to the charging station for quick power replenishment is greater than the energy consumed by going to and from the charging station; if the power replenishment benefit is greater than the round-trip energy consumption, the module will perform fragmented charging; if the power replenishment benefit is not greater than the round-trip energy consumption, the next task will be executed.

[0057] The detailed evaluation formula is as follows:

[0058] in, If the current energy gain is greater than 0, perform fragmented charging; otherwise, proceed to the next task. (Round trip energy consumption) is the energy consumed by the robot to travel to the charging station and return to the original starting point of the next task. It is calculated by the high-precision energy consumption model based on the planned round trip path.

[0059] (Charging revenue) is the amount gained during the time window between tasks. The formula for calculating the energy that can be added inside is as follows:

[0060] in, For charging power, and For the outbound and return journey times, This is the maximum effective charging time.

[0061] This mechanism makes full use of the idle time of tasks and increases the overall available power at near-zero cost, which is a key optimization strategy to maximize system energy efficiency.

[0062] Example 2: Robot Endurance Optimization Method This method aims to optimize the endurance of mobile robots through a closed-loop robot endurance optimization system based on perception, modeling, planning, and decision-making. Its core lies in deeply integrating real-time energy consumption estimation into the entire process of robot behavior decision-making, specifically including the following three collaborative steps: Step 1: Real-time high-precision energy consumption sensing and modeling This step forms the data foundation for all optimization decisions.

[0063] First, multi-dimensional data is collected in real time during the robot's operation, including: Powertrain data: Battery voltage / current (BMS), motor current (driver); Motion state data: triaxial acceleration / angular velocity (IMU), real-time velocity (encoder), used to quantify the robot's dynamic behavior.

[0064] Load and environmental data: load mass (pressure sensor), ground type and slope (visual / LiDAR or pre-stored map).

[0065] Subsequently, the above data is input into the physical power consumption model of the aforementioned system for calculation. The total power consumption (P_total) of this model consists of basic fixed power consumption (P_basic), motion power consumption (P_motion), and environmental power consumption (P_environment), as shown in the following formula:

[0066] Among them, the motion power consumption model is the key. It integrates multiple dynamic variables such as mass, acceleration, speed, ground friction and slope, and can accurately reflect the energy consumption changes in working conditions such as sudden start and stop, climbing, and driving on different road surfaces.

[0067] Finally, based on the real-time calculated total power consumption P_total and the current speed, the method dynamically calculates a more accurate remaining driving range and generates an "energy cost map" with energy consumption weights, providing spatialized energy consumption information for route planning.

[0068] Step 2: Dynamic Energy Efficiency Path Planning This step utilizes the task objective and real-time map containing obstacles output in Step 1 to perform intelligent path search.

[0069] After receiving the task objective, the system's dynamic energy efficiency path planning module does not only search for the geometric shortest path, but optimizes with the goal of "minimizing the total cost".

[0070] The total cost function is designed as follows:

[0071] in, Path geometric length; Estimated travel time; Estimated energy consumption costs Configurable weights.

[0072] The estimated energy cost is directly derived from querying and accumulating the "energy cost map". By adjusting the weights (especially the energy cost weight γ), the robot can actively choose to avoid high-friction surfaces, steep slopes, and other high-energy-consuming areas, resulting in lower total energy consumption even if the route is slightly longer. At the same time, the method generates a smooth velocity curve (such as an S-curve) for the planned path to limit rapid acceleration and deceleration, further reducing motion energy consumption from a control perspective.

[0073] Step 3: Adaptive Intelligent Charging Decision This step proactively manages energy at the global task scheduling level. The decision-maker continuously monitors the robot's remaining battery life and task queue, and executes a three-layer decision-making logic (such as...) when a new task arrives. Figure 2 (as shown) 1. Emergency charging: Determine if the battery charge (SOC) is below the safety threshold. If the battery charge is below the safety threshold, immediately trigger emergency charging; otherwise, if the battery charge is not below the safety threshold, proceed to preventive charging.

[0074] 2. Preventive Charging: Predicts whether the current battery range is sufficient to complete the next or several future tasks (including the cost of traveling to and from charging stations). If not, preventive charging is automatically triggered after the current task is completed to ensure that subsequent tasks are not interrupted due to power shortage; if so, it enters fragmented charging mode.

[0075] 3. Fragmented Charging: If the previous steps predict sufficient battery life, determine if there is a fragmented time window between tasks. If no fragmented time window exists, proceed to the next task. If a fragmented time window exists, assess whether the energy gain from "fragmented charging" at a charging station within this window exceeds the energy consumed traveling to and from the charging station. If the gain exceeds the energy consumed, it is considered a significant gain, and fragmented charging is performed to increase overall available battery power. If the gain is not greater than the energy consumed traveling to and from the charging station, proceed to the next task.

[0076] This method, through a closed-loop process of "real-time and accurate energy consumption perception - planning energy-saving driving paths - actively scheduling charging opportunities", enables robots to transform from passively responding to low battery alarms to actively predicting, planning and managing their own energy, thereby maximizing continuous working time and overall energy efficiency in complex dynamic environments.

[0077] The above technical solution is only one embodiment of the present invention. For those skilled in the art, based on the principles disclosed in the present invention, it is easy to make various types of improvements or modifications, and not limited to the technical solutions described in the specific embodiments of the present invention. Therefore, the foregoing description is only a preferred option and is not restrictive.

Claims

1. A robot dynamic energy efficiency optimization system, characterized in that, The system includes: The high-precision energy consumption modeling module is used to collect multi-dimensional data of the robot in real time, calculate the robot's real-time total power consumption and remaining battery life, and generate an energy cost map. The dynamic energy efficiency path planning module is used to receive the task objective and a real-time map containing obstacles, and to generate an energy-efficient path using the energy cost map. The adaptive charging decision module is used to make tiered charging decisions based on the robot's real-time battery level, remaining range, and task queue.

2. The robot dynamic energy efficiency optimization system according to claim 1, characterized in that, The multi-dimensional data includes: power system data, motion state data, load and environmental data.

3. The robot dynamic energy efficiency optimization system according to claim 1 or 2, characterized in that, The real-time total power consumption and remaining range of the computational robot are calculated by using a physical power consumption model to decompose the total power consumption into basic fixed power consumption, motion power consumption and environmental power consumption, and the remaining range is dynamically calculated based on the real-time speed.

4. The robot dynamic energy efficiency optimization system according to claim 3, characterized in that, The calculation of motion power consumption takes into account the following factors: The robot's mass parameters include its own weight and real-time dynamic load. Motion state parameters, including acceleration and velocity; Environmental parameters, including slope angle and ground friction coefficient; System efficiency parameters include transmission efficiency, motor efficiency, and air resistance.

5. The robot dynamic energy efficiency optimization system according to claim 4, characterized in that, The dynamic energy efficiency path planning module adopts a multi-objective cost function, which accumulates the energy consumption cost weights of all grids passed through the path, and obtains the energy efficiency optimized path through forward simulation and integral calculation in combination with the speed curve. The energy cost is obtained by accumulating the weight values ​​of the areas traversed by the planned path in the energy cost map.

6. The robot dynamic energy efficiency optimization system according to claim 5, characterized in that, The dynamic energy efficiency path planning module also includes a velocity curve optimization module, which generates a smooth velocity curve to match the generated energy efficiency optimized path, so as to limit acceleration and jerk.

7. The robot dynamic energy efficiency optimization system according to claim 6, characterized in that, The adaptive charging decision module executes a hierarchical charging decision logic, including: performing safety charging judgment, task continuity assurance judgment, and energy efficiency optimization charging judgment in sequence through three layers of decision-making.

8. The robot dynamic energy efficiency optimization system according to claim 7, characterized in that, The energy efficiency optimization charging judgment includes: when there is a task gap time window, deciding whether to perform fragmented charging based on the comparison between the energy consumption of going to and from the charging station and the energy that can be charged within the time window.

9. A robot endurance optimization method, which optimizes the endurance of a mobile robot using the robot endurance optimization system described in any one of claims 1-8, characterized in that, The method includes the following steps: Step 1: Collect multi-dimensional data during robot operation in real time, build a physical power consumption model to calculate the robot's total power consumption and remaining battery life, and generate an energy cost map; Step 2: Based on the task objectives, real-time map, and energy cost map, perform dynamic energy efficiency path planning; Step 3: Make tiered charging decisions based on remaining range, task queue, and charging station location.

10. The robot dynamic energy efficiency optimization method according to claim 9, characterized in that, The operation of making tiered charging decisions includes: Determine if the robot's real-time battery level is below the safety threshold; if so, perform emergency charging. If not, determine whether the dynamic remaining battery life can support the completion of the next task. If not, perform preventative charging. If yes, determine whether there is a fragmented time window between tasks. If not, execute the next task. If so, assess whether the energy gain from fragmented charging is greater than the energy consumed in traveling to and from the charging station. If so, execute fragmented charging; otherwise, proceed to the next task.

Images