A battery thermal management method for a cleaning robot

Abstract

The application provides a battery thermal management method of a cleaning robot, comprising: acquiring environmental state data in a main control system of the cleaning robot; determining an estimated temperature rise curve by using a thermal load prediction model according to the environmental state data, a remaining cleaning area and a navigation path plan; adjusting a maximum output power limit of a lithium battery according to a relationship between the estimated temperature rise curve and a preset thermal safety threshold to obtain a corrected power parameter; starting a heat dissipation system by using the corrected power parameter to determine an advanced heat dissipation execution plan; acquiring real-time temperature data of a battery management system, judging a relationship between the real-time temperature data and the thermal safety threshold, and generating an optimized cleaning path sequence; recalculating an output of the thermal load prediction model according to the optimized cleaning path sequence, judging whether a predicted thermal load decreases, and determining a stable temperature control state.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a battery thermal management method for a cleaning robot. Background Technology

[0002] Current battery management systems for cleaning robots primarily rely on real-time monitoring of battery temperature and current, triggering frequency reduction or heat dissipation measures based on fixed thresholds. This approach often lags behind actual changes in thermal load, leading to overheating risks in high-intensity tasks due to delayed response. Furthermore, existing systems typically focus only on the battery's operational status, neglecting the complexity of the cleaning robot's working environment and the dynamic nature of task planning. For example, the impact of floor material on motor load or the varying battery requirements due to different cleaning path lengths. This passive management approach struggles to adapt to diverse working scenarios. The core technical challenge in solving this problem lies in enabling the battery management system to be proactive, capable of anticipating and responding to potential thermal load fluctuations based on changes in the cleaning robot's task and environment. Battery temperature control is not only related to the current output but also influenced by the combined effects of the external environment, task intensity, and duration. For instance, when a cleaning robot operates in high-power mode on a long-pile carpet, the motor requires higher power, causing a surge in battery current output and potentially leading to a rapid temperature increase within a short period. If the management system fails to anticipate such high-load scenarios and take appropriate measures, the battery may be damaged by sustained high heat, potentially even posing a safety hazard. Therefore, how to dynamically predict the future thermal load of the battery and adjust the operating strategy in advance based on real-time environmental and task information when cleaning robots perform diverse tasks has become a key issue in improving battery safety and task reliability. Summary of the Invention

[0003] This invention provides a battery thermal management method for a cleaning robot, mainly comprising: The system acquires environmental status data from the main control system of the cleaning robot; based on the environmental status data, remaining cleaning area, and navigation path planning, it uses a heat load prediction model to determine the estimated temperature rise curve; based on the relationship between the estimated temperature rise curve and a preset thermal safety threshold, it adjusts the maximum output power limit of the lithium battery to obtain corrected power parameters; it starts the cooling system using the corrected power parameters to determine an advance cooling execution plan; it acquires real-time temperature data from the battery management system, determines the relationship between the real-time temperature data and the thermal safety threshold, and generates an optimized cleaning path sequence; it recalculates the output of the heat load prediction model based on the optimized cleaning path sequence, determines whether the predicted heat load has decreased, and establishes a stable temperature control state; based on the stable temperature control state, it uses a feedback loop mechanism to update the parameters of the heat load prediction model and determines the final adaptive control result of the thermal safety boundary.

[0004] Furthermore, acquiring environmental state data from the cleaning robot's main control system includes: identifying the floor material of the cleaning area using a camera or ultrasonic sensor, and determining whether the floor material is a high-friction carpet or a low-friction floor; acquiring the current cleaning mode from the main control system; generating environmental state data characterizing the cleaning area based on the floor material and the current cleaning mode, wherein the environmental state data includes the floor material type, friction coefficient, and cleaning mode parameters; determining the operating resistance characteristics of the cleaning robot in the current cleaning area using the environmental state data; and generating environmental state parameters compatible with the input of the heat load prediction model based on the operating resistance characteristics.

[0005] Furthermore, the step of determining the estimated temperature rise curve using a heat load prediction model based on the environmental state data, remaining cleaning area, and navigation path planning includes: obtaining the remaining cleaning area and the path length and turning frequency in the navigation path planning; determining the power consumption characteristics of the cleaning robot on different ground materials based on the environmental state data; calculating the battery current output demand over a future period using the heat load prediction model, combined with the power consumption characteristics, the path length, and the turning frequency; and generating an estimated temperature rise curve characterizing battery temperature changes based on the battery current output demand.

[0006] Furthermore, the step of adjusting the maximum output power limit of the lithium battery based on the relationship between the estimated temperature rise curve and the preset thermal safety threshold to obtain the corrected power parameters includes: comparing the estimated temperature rise curve with the preset thermal safety threshold to determine whether there is a risk of exceeding the thermal safety threshold; if the risk exists, reducing the maximum output power limit through the dynamic control mechanism of the lithium battery; and calculating the corrected power parameters based on the reduced maximum output power limit.

[0007] Furthermore, the step of starting the heat dissipation system using the corrected power parameters and determining the advance heat dissipation execution plan includes: determining the required speed of the cooling fan based on the corrected power parameters; pre-starting the heat dissipation system using the required speed and turning on the cooling fan at a higher speed; generating the advance heat dissipation execution plan based on the operating parameters of the cooling fan; determining the operating time and intensity of the heat dissipation system based on the advance heat dissipation execution plan; and generating an expected temperature drop curve characterizing the heat dissipation effect based on the operating time and intensity.

[0008] Furthermore, the step of acquiring real-time temperature data from the battery management system, determining the relationship between the real-time temperature data and the thermal safety threshold, and generating an optimized cleaning path sequence includes: acquiring real-time temperature data from the battery management system; comparing the real-time temperature data with the preset thermal safety threshold to determine whether it is close to the thermal safety boundary; if it is close to the thermal safety boundary, sending a task scheduling suggestion to the main control system; and adjusting the navigation path planning according to the task scheduling suggestion to generate an optimized cleaning path sequence.

[0009] Furthermore, the step of recalculating the output of the heat load prediction model based on the optimized cleaning path sequence, determining whether the predicted heat load has decreased, and establishing a stable temperature control state includes: updating the input parameters of the heat load prediction model using the optimized cleaning path sequence; recalculating the output of the heat load prediction model based on the updated input parameters to generate a new predicted heat load; comparing the new predicted heat load with the previous predicted heat load to determine whether the heat load has decreased; if the heat load has decreased, maintaining the current power parameters and heat dissipation system settings; and determining a temperature control state characterizing stable battery operation based on the maintained settings.

[0010] Furthermore, the step of updating the parameters of the heat load prediction model using a feedback loop mechanism through the stable temperature control state to determine the final adaptive control result of the thermal safety boundary includes: acquiring real-time operating data of the battery through the stable temperature control state; adjusting the parameters of the heat load prediction model using a feedback loop mechanism based on the real-time operating data; recalculating the output of the heat load prediction model using the adjusted parameters to generate an updated thermal safety boundary; determining an adaptive control strategy based on the updated thermal safety boundary; and generating the final adaptive control result of the thermal safety boundary through the adaptive control strategy.

[0011] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention acquires environmental status data, including ground material type, friction coefficient, and cleaning mode parameters, through a main control system. It generates operational resistance characteristics and, combined with the remaining cleaning area and navigation path planning, uses a heat load prediction model to calculate the estimated temperature rise curve. Based on this, it compares the temperature rise curve with a thermal safety threshold, dynamically adjusts the maximum output power limit of the lithium battery to obtain corrected power parameters, and simultaneously pre-activates the cooling system to formulate an advance cooling plan, ensuring the heat load remains within a safe range. Furthermore, it collects real-time temperature data from the battery management system; if the temperature approaches the threshold, it optimizes the cleaning path sequence to reduce the running time in high-heat-load areas. Based on the optimized path, it recalculates the predicted heat load, confirms a decrease, maintains stable temperature control, and updates model parameters through feedback loops to generate adaptive thermal safety boundary control results. This method achieves accurate heat load prediction and dynamic control, significantly improving battery thermal management efficiency and ensuring the long-term stable operation of the cleaning robot. Attached Figure Description

[0012] Figure 1 This is a flowchart of a battery thermal management method for a cleaning robot according to the present invention. Detailed Implementation

[0013] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] like Figure 1 This embodiment of a battery thermal management method for a cleaning robot may specifically include: Acquire environmental status data from the main control system of the cleaning robot; Based on the environmental status data, remaining clean area, and navigation path planning, a heat load prediction model is used to determine the estimated temperature rise curve.

[0015] Based on the relationship between the estimated temperature rise curve and the preset thermal safety threshold, the maximum output power limit of the lithium battery is adjusted to obtain the corrected power parameters.

[0016] The cooling system is activated using the revised power parameters to determine an advance cooling execution plan. The system acquires real-time temperature data from the battery management system, determines the relationship between the real-time temperature data and the thermal safety threshold, and generates an optimized cleaning path sequence.

[0017] The output of the heat load prediction model is recalculated based on the optimized cleaning path sequence to determine whether the predicted heat load has decreased and to establish a stable temperature control state.

[0018] By using the stable temperature control state, the parameters of the heat load prediction model are updated using a feedback loop mechanism to determine the final adaptive control result of the thermal safety boundary.

[0019] S1. The step of acquiring environmental state data in the main control system of the cleaning robot includes: identifying the ground material of the cleaning area through a camera or ultrasonic sensor, and determining that the ground material is a high-friction carpet or a low-friction floor; acquiring the current cleaning mode from the main control system; generating environmental state data characterizing the characteristics of the cleaning area based on the ground material and the current cleaning mode, wherein the environmental state data includes the ground material type, friction coefficient, and cleaning mode parameters; determining the operating resistance characteristics of the cleaning robot in the current cleaning area based on the environmental state data; and generating environmental state parameters compatible with the input of the heat load prediction model based on the operating resistance characteristics.

[0020] S101, the floor material of the cleaning area is identified by a camera or ultrasonic sensor, and it is determined that the floor material is a high-friction carpet or a low-friction floor.

[0021] In one embodiment, after the cleaning robot is started, the camera captures images of the cleaning area, and the texture and color features are analyzed by image processing algorithms to classify the ground as a high-friction carpet or a low-friction floor. For example, carpets show dense fiber textures, while floors show smooth reflections.

[0022] S1011, activate the camera to collect real-time image data.

[0023] S1012 uses preset image recognition rules to compare data and determine the material type.

[0024] S102, Obtain the current cleaning mode from the main control system.

[0025] The cleaning robot reads the current settings from the main control system, such as vacuuming mode or mopping mode, and extracts the parameters directly.

[0026] S103, Based on the ground material and the current cleaning mode, generate environmental state data characterizing the characteristics of the cleaning area, wherein the environmental state data includes ground material type, friction coefficient and cleaning mode parameters.

[0027] In one embodiment, environmental state data, including floor material type, friction coefficient, and cleaning mode, is generated by combining the floor material and high-friction carpet friction coefficient determined in S101 with the vacuuming mode parameters in S102. This generation process ensures that the data comprehensively characterizes the area's properties, which is beneficial for subsequent resistance assessment and improves robot operating efficiency.

[0028] S1031, extract the floor material type and the coefficient of friction of high-friction carpet (0.8) or low-friction floor (0.3).

[0029] The coefficient of friction is calculated by measuring the resistance ratio between different ground materials and robot wheels through experiments, such as by testing the wheel sliding distance under laboratory conditions.

[0030] S1032 integrates current cleaning mode parameters, such as suction intensity levels 1 to 5.

[0031] S1033, combine and generate environmental status data to ensure data structure consistency.

[0032] For example, when running on a high-friction carpet at mode level 3, the data highlights the potential impact of high resistance.

[0033] S104, using the environmental state data, determine the operating resistance characteristics of the cleaning robot in the current cleaning area.

[0034] Specifically, the operating resistance characteristics are calculated based on the environmental condition data of S103. The resistance characteristics refer to the total resistance encountered by the robot when it moves forward. The formula is: Resistance = Coefficient of friction × Robot weight × Speed ​​factor. This ensures that the characteristics accurately reflect the actual operating conditions and is beneficial to the accuracy of heat load prediction.

[0035] S1041, Input the friction coefficient and cleaning mode parameters from the environmental condition data.

[0036] S1042, calculate the resistance characteristic value. For example, when the friction coefficient is 0.8, the weight is 10kg, and the speed is 0.5m / s, the resistance is approximately 4N.

[0037] The speed factor is a preset empirical value, such as 0.5 corresponding to low speed mode.

[0038] S1043 outputs the running resistance characteristics as a numerical sequence.

[0039] In one embodiment, for high-friction carpets, the resistance characteristics show a higher peak value than those for low-friction floors, which helps to avoid overload.

[0040] For example, in the living room carpet area, the calculated resistance characteristic is 5N, which is significantly different from the 2N of the kitchen floor.

[0041] S105, Based on the operating resistance characteristics, generate environmental state parameters compatible with the input of the heat load prediction model.

[0042] For example, parameters can be generated using the resistance characteristics of S104, such as converting resistance values ​​into current demand estimates, ensuring that the parameter format matches the input requirements of the heat load prediction model, such as parameters including resistance level and expected power consumption.

[0043] S1051, convert the operating resistance characteristics to standardized parameters, for example, resistance 5N corresponds to parameter value 50.

[0044] S1052, Adjust parameters to be compatible with the model input interface.

[0045] In one embodiment, environmental state parameters generated in high-friction regions cause the model to predict higher heat loads, thereby adjusting the power in advance.

[0046] For example, parameters can be directly input into the model after generation to optimize path planning.

[0047] Specifically, for different ground surfaces, the parameters cover a range of resistance from 2N to 5N to ensure that the model accurately predicts the temperature rise curve.

[0048] S2. The step of determining the estimated temperature rise curve using a heat load prediction model based on the environmental state data, remaining cleaning area, and navigation path planning includes: obtaining the remaining cleaning area and the path length and turning frequency in the navigation path planning; determining the power consumption characteristics of the cleaning robot on different ground materials based on the environmental state data; calculating the battery current output demand over a future period using the heat load prediction model, combined with the power consumption characteristics, the path length, and the turning frequency; and generating an estimated temperature rise curve characterizing battery temperature changes based on the battery current output demand.

[0049] S201, obtain the remaining cleaning area and the path length and turning frequency in the navigation path planning.

[0050] The main control system of the cleaning robot reads the remaining area data of the current cleaning task and extracts the total length of the planned path and the frequency of turning in the path from the navigation module.

[0051] S202, Based on the environmental status data, determine the power consumption characteristics of the cleaning robot on different ground materials.

[0052] S2021 identifies the current ground material type, such as high-friction carpet or low-friction floor, based on environmental status data collected by cameras or ultrasonic sensors. S2022, based on the identified ground material, query the pre-stored power consumption table, which records the power requirements of the robot during movement and cleaning under different materials. For example, power consumption is higher on carpets to overcome friction, and lower on floors. S2023, combined with the remaining cleaning area, calculate the distribution ratio of specific material areas to determine the overall power consumption characteristics.

[0053] For example, in a living room cleaning scenario, if environmental data shows that 80% of the floor is low-friction and 20% is high-friction carpet, the power consumption characteristics will be biased towards the lower average value.

[0054] S203, using the heat load prediction model, combined with the power consumption characteristics, the path length, and the turning frequency, calculate the battery current output demand over a future period.

[0055] S2031, power consumption characteristics are used as input parameters, and path length and turning frequency are used as time series variables to input into the heat load prediction model. The heat load prediction model is based on the linear regression method. Linear regression is a statistical model that fits the relationship between independent and dependent variables using the least squares method. Here, the independent variables include path length, turning frequency and power characteristics, and the dependent variable is current demand. S2032, the model calculation process involves first standardizing the input data, and then applying regression coefficients for calculation. For example, current demand = a Path length + b Steering frequency + c Power characteristics, where a, b, and c are coefficients obtained through training with historical data; S2033 outputs a sequence of current output requirements for the next 30 minutes, ensuring that the forecast covers the entire remaining cleaning area.

[0056] For example, in a scenario with a path length of 100 meters, a turning frequency of 0.1 times per meter, and a moderate power characteristic, the model predicts a peak current demand of 5 amps, which is beneficial for early intervention.

[0057] In one embodiment, S203 can be applied to high-load scenarios, extending the model by adding a time window function to perform segmented prediction of the path; using a variant of linear regression such as ridge regression to handle multicollinearity, with ridge regression adding a regularization term in least squares to stabilize the coefficients; calculating and summing the segmented current demands to obtain the total demand.

[0058] It should be noted that this extension can handle complex paths and is beneficial to the thermal stability of the robot during long-term operation.

[0059] For example, if the path includes multiple areas with dense turning, the current demand forecast will increase, indicating that power needs to be adjusted.

[0060] S204, Based on the battery current output requirements, generate a predicted temperature rise curve characterizing the battery temperature change.

[0061] S2041 uses the Joule heating formula to estimate heat generation based on current demand. Joule heating is the heat generated when current passes through a resistor, and the formula is Q=I. 2 Rt, where I is the current, R is the resistance, and t is the time; S2042, combined with the battery thermal capacity, generates a curve showing the temperature change over time.

[0062] For example, when the current requirement is 4 amps, the predicted temperature rise curve shows the temperature rising from 25 degrees Celsius to 35 degrees Celsius.

[0063] S3, adjusting the maximum output power limit of the lithium battery based on the relationship between the estimated temperature rise curve and the preset thermal safety threshold to obtain the corrected power parameters includes: comparing the estimated temperature rise curve with the preset thermal safety threshold to determine whether there is a risk of exceeding the thermal safety threshold; if the risk exists, reducing the maximum output power limit through the dynamic control mechanism of the lithium battery; and calculating the corrected power parameters based on the reduced maximum output power limit.

[0064] S301. Compare the estimated temperature rise curve with the preset thermal safety threshold to determine whether there is a risk of exceeding the thermal safety threshold.

[0065] S3011, obtains data points for the predicted temperature rise curve. These data points are calculated based on the heat load prediction model and cover the temperature change trend over a future period of time.

[0066] S3012 compares each data point with a preset thermal safety threshold point by point. For example, if the threshold is set to 60 degrees Celsius, any data point exceeding this value is marked as having a risk.

[0067] For example, when cleaning robots handle high-friction carpets, the predicted temperature rise curve may show a rapid increase. By comparing and judging the risk, intervention and control can be carried out in time to extend battery life.

[0068] S302, If the aforementioned risk exists, the maximum output power limit is reduced through the dynamic regulation mechanism of the lithium battery.

[0069] S3021 activates the dynamic control mechanism, which dynamically adjusts the upper limit of power based on real-time temperature data and the estimated temperature rise curve.

[0070] S3022 calculates the ratio of the current battery current output demand to the risk level. For example, if the risk is high, the maximum output power is reduced from the standard 100 watts to 80 watts.

[0071] S3023 applies the reduced limits to the battery management system to ensure that the output does not exceed the new upper limit.

[0072] Specifically, when environmental data of high-friction carpets or low-friction floors are identified, the dynamic control mechanism combines the remaining cleaning area and navigation path planning to reduce power limitations.

[0073] In one possible implementation, a robot cleans a large area of ​​high-friction carpet. The predicted temperature rise curve indicates the risk, and the mechanism reduces the power to 70 watts, which reduces the heat load and maintains stable operation.

[0074] In another possible implementation, for low-friction flooring with complex paths, the mechanism progressively reduces power to 85 watts to avoid sudden overheating.

[0075] It should be noted that this dynamic control updates parameters through a feedback loop mechanism to achieve self-adaptation, which can effectively cope with the temperature rise caused by different ground materials and improve the overall thermal safety boundary.

[0076] S303, calculate the corrected power parameters based on the reduced maximum output power limit.

[0077] The new current requirement is calculated using the reduced power limit as input.

[0078] S4, the step of starting the heat dissipation system using the corrected power parameters and determining the advance heat dissipation execution plan includes: determining the required speed of the cooling fan based on the corrected power parameters; pre-starting the heat dissipation system using the required speed and turning on the cooling fan at a higher speed; generating the advance heat dissipation execution plan based on the operating parameters of the cooling fan; determining the operating time and intensity of the heat dissipation system based on the advance heat dissipation execution plan; and generating an expected temperature drop curve characterizing the heat dissipation effect based on the operating time and intensity.

[0079] S401, Based on the corrected power parameters, determine the required speed of the cooling fan.

[0080] S4011, obtain the corrected power parameters.

[0081] S4012 inputs the corrected power parameters into the speed mapping function to calculate the required speed of the cooling fan. The speed mapping function is a linear relationship fitted from experimental data to ensure that the speed is proportional to the power.

[0082] In one embodiment, the corrected power parameter is 80 watt-hours, and the speed mapping function can be expressed as speed = 10. Power +500, speed requirement calculated as 1300 rpm.

[0083] S402, based on the stated speed requirement, pre-start the heat dissipation system by turning on the cooling fan at a high speed.

[0084] S4021 receives the speed demand signal and sends a pre-start command to the cooling system.

[0085] S4022 controls the cooling fan to start at a speed higher than the standard speed, for example, if the standard speed is 1000 rpm, it starts at 1200 rpm.

[0086] In one embodiment, when the required speed is 1300 rpm, the fan immediately reaches that speed during pre-start to avoid heat buildup caused by delay.

[0087] S403, Generate an advance heat dissipation execution plan based on the operating parameters of the cooling fan.

[0088] S4031 collects the operating parameters of the cooling fan, including the current speed, power consumption, and airflow output, where the airflow output is measured by a wind speed sensor.

[0089] S4032 integrates operating parameters, uses time series analysis to predict short-term heat dissipation needs, and generates an execution plan, which includes the start-up timing and duration.

[0090] Construct an operating parameter vector, using engine speed, power consumption, and airflow output as vector elements. Process the parameter vector using an autoregressive integral moving average model to predict the heat dissipation demand curve for the next 5 minutes. Based on the heat dissipation demand curve, develop an execution plan to ensure coverage of periods of high heat load.

[0091] In one embodiment, for a cleaning robot operating on a high-friction carpet, the operating parameters show a rotation speed of 1500 rpm, a power consumption of 5 watts, and an airflow output of 2 cubic meters per minute. The demand curve peak is predicted to appear after 3 minutes using an autoregressive integral moving average model, and an execution plan is generated to start immediately and continue for 4 minutes.

[0092] In one embodiment, in a low-friction floor scenario, the operating parameters are a rotation speed of 1200 rpm, a power consumption of 4 watts, and an airflow output of 1.5 cubic meters per minute. The model predicts stable demand, and the execution plan is adjusted to intermittent start-up, running for 1 minute every 2 minutes.

[0093] For example, when there is a large remaining cleaning area, the execution plan generated after integrating the operating parameters can prioritize covering the high-load segments in the navigation path, which is beneficial to extending battery life.

[0094] S404, through the aforementioned advance heat dissipation execution plan, determines the operating time and intensity of the heat dissipation system.

[0095] S4041 extracts the time period and intensity level from the execution plan, with the intensity level based on fan speed grading.

[0096] S4042, set the running time to the duration specified in the plan, and the intensity to the corresponding rotational speed.

[0097] In one embodiment, the execution plan specifies a runtime of 5 minutes and an intensity of high speed, i.e., 1500 revolutions per minute.

[0098] S405, based on the running time and intensity, generate a expected temperature drop curve characterizing the heat dissipation effect.

[0099] S4051, input the running time and intensity into the heat conduction simulation model. The heat conduction simulation model is based on the Fourier heat conduction equation and simulates the heat diffusion process. The Fourier heat conduction equation describes the propagation of heat over time and space. The calculation process includes discretizing the time step and spatial grid, and iteratively solving for the temperature distribution.

[0100] S4052 simulates temperature changes at different time points and generates a decreasing curve, which represents the gradual decrease of temperature from the initial value to the safe threshold.

[0101] S40521, establish the initial temperature distribution grid with a grid resolution of 0.1 cm.

[0102] S40522, apply boundary conditions, including fan intensity as a cooling source, and iteratively calculate the temperature update for each grid point.

[0103] S40523, the fitting calculation result is a continuous curve, representing the expected downward trend.

[0104] In one embodiment, the running time is 5 minutes and the intensity is 1500 revolutions per minute. The heat conduction simulation model calculates an initial temperature of 50 degrees Celsius. The temperature drop curve shows that the temperature drops to 45 degrees Celsius within 2 minutes and stabilizes at 40 degrees Celsius within 5 minutes, which helps prevent the battery from overheating.

[0105] In one embodiment, for a runtime of 3 minutes and an intensity of 1200 revolutions per minute, the model simulates an initial temperature of 48 degrees Celsius, with the temperature decreasing to 44 degrees Celsius within 1 minute and finally to 42 degrees Celsius, ensuring the continuity of the cleaning task.

[0106] For example, in a high-friction carpet environment, the generated descent curve can be used to verify thermal safety boundaries, emphasizing the need for early intervention to achieve temperature stability.

[0107] S5, the step of acquiring real-time temperature data from the battery management system, determining the relationship between the real-time temperature data and the thermal safety threshold, and generating an optimized cleaning path sequence includes: acquiring real-time temperature data from the battery management system; comparing the real-time temperature data with the preset thermal safety threshold to determine whether it is close to the thermal safety boundary; if it is close to the thermal safety boundary, sending a task scheduling suggestion to the main control system; adjusting the navigation path planning according to the task scheduling suggestion to generate an optimized cleaning path sequence; and reducing the running time of the cleaning robot in high heat load areas through the optimized cleaning path sequence.

[0108] S501, Obtain real-time temperature data from the battery management system.

[0109] Specifically, the temperature value of the lithium battery cell is collected every 5 seconds by the temperature sensor built into the battery management system, and the collected temperature value is transmitted to the control module as real-time temperature data.

[0110] S502, compare the real-time temperature data with the preset thermal safety threshold to determine whether it is close to the thermal safety boundary.

[0111] S5021 compares the real-time temperature data with the thermal safety threshold, where the thermal safety threshold is preset to 45 degrees Celsius. Approaching the thermal safety boundary is defined as the real-time temperature data reaching 90% of the thermal safety threshold, that is, above 40.5 degrees Celsius but not exceeding 45 degrees Celsius.

[0112] S5022, based on the comparison results, if the real-time temperature data is between 40.5 degrees Celsius and 45 degrees Celsius, it is determined to be close to the thermal safety boundary and a boundary approach signal is generated.

[0113] For example, when a cleaning robot is handling a high-friction carpet area, the increased output power of the lithium battery causes the temperature to rise. If the real-time temperature data is 42 degrees Celsius, it is determined that it is approaching the thermal safety boundary. Doing so can trigger the adjustment mechanism in time to avoid battery overheating and damage, and improve the overall operational safety.

[0114] S503, If the thermal safety boundary is approached, a task scheduling suggestion is sent to the main control system.

[0115] Specifically, once a boundary approach signal is generated, a suggestion signal is immediately sent to the main control system via the internal communication interface. The suggestion includes prioritizing avoiding areas with high heat loads.

[0116] S504, Based on the task scheduling suggestion, adjust the navigation path planning to generate an optimized clean path sequence.

[0117] S5041, after receiving the task scheduling suggestion, analyze the high heat load area in the current navigation path, for example, identify the carpet area as a high heat load area.

[0118] S5042, based on the remaining cleaning area and environmental condition data, uses A The path planning algorithm recalculates the path.

[0119] A Path planning algorithms are a heuristic search method that calculates the cost function f(n) = g(n) + h(n) from the starting point to the target, where g(n) is the actual cost from the starting point to the current node and h(n) is the estimated cost from the current node to the target. Here, the cost weight of high heat load areas is set to a higher value, such as 1.5 times the original path cost, in order to avoid these areas.

[0120] S5043, Generate an optimized cleaning path sequence that adjusts the cleaning order to treat low-friction floor areas first, followed by high-friction carpet areas, and shortens the time spent on carpets.

[0121] For example, when a cleaning robot faces a mixed floor environment, if the original path plan directly passes through the carpet, causing the temperature to approach the limit, the adjusted path sequence will bypass the carpet and clean 80% of the floor area first. This can reduce the predicted heat load by 15% and keep the battery temperature within a safe range.

[0122] Step S6, recalculating the output of the heat load prediction model based on the optimized cleaning path sequence, determining whether the predicted heat load has decreased, and determining a stable temperature control state, includes: updating the input parameters of the heat load prediction model using the optimized cleaning path sequence; recalculating the output of the heat load prediction model based on the updated input parameters to generate a new predicted heat load; comparing the new predicted heat load with the previous predicted heat load to determine whether the heat load has decreased; if the heat load has decreased, maintaining the current power parameters and heat dissipation system settings; and determining a temperature control state characterizing stable battery operation based on the maintained settings.

[0123] S601, Update the input parameters of the heat load prediction model using the optimized cleaning path sequence.

[0124] S6011, Obtain the optimized cleaning path sequence, which is generated based on real-time temperature data and task scheduling suggestions previously obtained from the battery management system, and includes adjusted waypoints and cleaning order.

[0125] S6012, based on this sequence, extract the remaining cleaning area and updated values ​​of the navigation path planning. These values ​​reflect the results of path optimization, such as reducing the dwell time in high-friction areas.

[0126] S6013 combines the extracted values ​​with the current cleaning mode and floor material information to form new input parameters for the heat load prediction model.

[0127] S602, based on the updated input parameters, recalculate the output of the heat load prediction model to generate a new predicted heat load.

[0128] S6021 imports the updated input parameters into the heat load prediction model to calculate the battery current output demand over a future period. For example, higher demand is required on high-friction carpets, leading to increased heat generation.

[0129] S6022 determines the predicted temperature rise curve based on calculation requirements and extracts a new predicted heat load value from it, which represents the amount of heat accumulation per unit time.

[0130] For example, when a cleaning robot is handling a low-friction floor, the input parameters after path sequence optimization show that the remaining area is reduced by 20%. The model recalculates and the heat load is reduced from the original 15 joules / second to 12 joules / second. This can lead to more accurate heat prediction and help prevent overheating.

[0131] S603, compare the new predicted heat load with the previous predicted heat load to determine whether the heat load has decreased.

[0132] S6031, read the predicted heat load previously calculated based on the original path sequence and compare it numerically with the new value.

[0133] S6032, if the new value is less than the old value, it is judged as a decrease, which ensures that the positive impact of the optimized path on the heat load is confirmed.

[0134] S604, if the heat load decreases, the current power parameters and heat dissipation system settings are maintained.

[0135] S6041, after confirming the decrease in heat load, retain the corrected power parameters, which were previously adjusted through a dynamic control mechanism, such as reducing the maximum output power limit to match the predicted temperature rise curve.

[0136] S6042, while keeping the heat dissipation system in a pre-start state, such as the cooling fan running at a high speed, which helps to continuously reduce heat buildup.

[0137] For example, in scenarios where robots clean high-friction carpets, maintaining the setting after the heat load decreases can stabilize the battery temperature below 40 degrees Celsius, which is beneficial for extending battery life; in another scenario, for mixed floors, maintaining the setting can avoid the increased energy consumption caused by frequent adjustments.

[0138] S605, by maintaining the settings, a temperature control state characterizing stable battery operation is determined.

[0139] The system monitors and maintains real-time temperature data under the specified settings, comparing it to a thermal safety threshold. If the temperature remains within the threshold, a stable temperature control state is generated. This state is used by the subsequent feedback loop mechanism to update model parameters, ensuring adaptive regulation.

[0140] For example, if the remaining cleaning area is large, maintaining the temperature fluctuation below 2 degrees Celsius after setting indicates that a steady state has been achieved, which is beneficial to overall operational safety. In another implementation, for low-friction floors, the steady state can be confirmed through multiple iterations, supporting the dynamic adjustment of thermal safety boundaries.

[0141] S7, the step of updating the parameters of the heat load prediction model using a feedback loop mechanism through the stable temperature control state to determine the final adaptive control result of the thermal safety boundary includes: acquiring real-time operating data of the battery through the stable temperature control state; adjusting the parameters of the heat load prediction model using a feedback loop mechanism based on the real-time operating data; recalculating the output of the heat load prediction model using the adjusted parameters to generate an updated thermal safety boundary; determining an adaptive control strategy based on the updated thermal safety boundary; and generating the final adaptive control result of the thermal safety boundary through the adaptive control strategy.

[0142] S701, by means of the stable temperature control state, obtains the real-time operating data of the battery.

[0143] During the operation of the cleaning robot, once the system enters a stable temperature control state, it reads data such as current temperature, current output, and voltage from the battery management system. This data is directly used for subsequent adjustments to ensure the real-time nature of the model input.

[0144] S702, Based on the real-time operating data, the parameters of the heat load prediction model are adjusted using a feedback loop mechanism.

[0145] S7021 calculates the deviation between the actual heat load and the predicted heat load based on real-time operating data.

[0146] For example, when a robot is cleaning a high-friction carpet, the real-time temperature data shows that the battery temperature is 45 degrees Celsius, while the model predicts 40 degrees Celsius, which is a deviation of 5 degrees Celsius.

[0147] S7022 uses the deviation value to update model parameters using the least squares method, such as adjusting the weight of the current output coefficient.

[0148] Using real-time operational data as observations, iterative calculations are performed to update the parameters in the heat load prediction model. For example, the current output coefficient is adjusted from an initial 0.8 to 0.9, allowing the model to more accurately reflect actual heat generation. This adjustment forms a feedback loop, with parameters gradually converging after each loop, improving prediction accuracy.

[0149] S7023, verify whether the model output after updating parameters reduces the bias. If it does not reduce the bias, repeat the loop.

[0150] In one possible implementation, for low-friction floor cleaning scenarios, the initial deviation is 3 degrees Celsius. After one cycle of parameter adjustment, the deviation drops to 1 degree Celsius, and the verification is successful. If the deviation remains high under complex path planning, the cycle continues until the threshold is met. This mechanism ensures the model's adaptability, especially when the remaining cleaning area is large, and can dynamically respond to the thermal risks caused by changes in battery load.

[0151] For example, when a robot is dealing with mixed ground materials, the feedback loop can cover multiple iterations. The first iteration is based on the initial data for adjustment, and the second iteration incorporates navigation path data to further refine the parameters and achieve accurate prediction of heat load.

[0152] S703, using the adjusted parameters, recalculate the output of the heat load prediction model to generate an updated thermal safety boundary.

[0153] S7031 inputs the adjusted parameters into the heat load prediction model to calculate the predicted temperature rise curve for a future period of time.

[0154] For example, using the adjusted current coefficient in the formula, the temperature rise curve can be predicted when the remaining cleaning area is 50 square meters.

[0155] S7032 determines the upper limit of the thermal safety boundary based on the predicted temperature rise curve, such as updating the boundary from 60 degrees Celsius to 55 degrees Celsius.

[0156] This update process involves comparing the predicted curve with historical data, and if the curve indicates potential overheating, the boundary value is lowered to increase the safety margin. In cleaning robot applications, this helps prevent battery damage under high-load conditions.

[0157] S7033 generates updated thermal safety boundaries as the basis for subsequent regulation.

[0158] In one embodiment, for low-friction flooring identified by ultrasonic sensors, updating the boundary can support higher power output, while on high-friction carpets, the boundary is conservatively set to ensure overall thermal safety.

[0159] For example, assuming the initial boundary is 58 degrees Celsius, if the predicted temperature rise is accelerated, it is updated to 52 degrees Celsius. This generation method connects the previous parameter adjustment through a logical chain, which directly affects the boundary accuracy and is beneficial for real-time thermal management.

[0160] S704, determine an adaptive control strategy based on the updated thermal safety boundary.

[0161] Based on the updated boundaries, strategies such as reducing maximum output power or optimizing the cleaning path sequence are developed.

[0162] For example, if the boundary is 55 degrees Celsius, the strategy might include increasing the fan speed to a moderate level. This determination process ensures that the strategy matches the boundary, forming a closed-loop control.

[0163] S705, through the aforementioned adaptive control strategy, generates the final adaptive control result of the thermal safety boundary.

[0164] After applying the strategy, the final result is output, such as confirming that the battery temperature is stable below 50 degrees Celsius, thus achieving adaptive regulation.

[0165] In one embodiment, the feedback loop mechanism in S702 can quickly adjust parameters when the cleaning robot handles different floor materials.

[0166] For example, in high-friction carpet scenarios, real-time data displays high current demand, and the feedback loop quickly adjusts parameters to reduce deviations; on low-friction floors, fewer loop iterations are needed to efficiently generate boundaries and optimize energy consumption.

[0167] It should be noted that the principle behind this mechanism is to continuously compare actual and predicted values, and through iterative optimization, ensure that the model adapts to the dynamic environment, thereby improving battery safety and robot operating efficiency.

[0168] For example, the recalculation process of S703 can be understood in principle as rerunning the thermal balance calculation using adjusted parameters to generate boundaries. In practice, if the initial predicted temperature rise is 8 degrees Celsius / minute, and it is adjusted to 5 degrees Celsius / minute, the updated boundary is more stringent, which is beneficial to extending the cleaning task time.

[0169] In one embodiment, when determining the strategy in S704, path sequence optimization is considered, such as placing high-load areas later to support overall control; this is connected with the result generated in S705 to form a complete chain, ensuring the reliability of the final thermal safety boundary adaptive control result.

[0170] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A battery thermal management method for a cleaning robot, characterized in that, include: Acquire environmental status data from the main control system of the cleaning robot; Based on the environmental status data, remaining clean area, and navigation path planning, a heat load prediction model is used to determine the estimated temperature rise curve. Based on the relationship between the estimated temperature rise curve and the preset thermal safety threshold, the maximum output power limit of the lithium battery is adjusted to obtain the corrected power parameters. The cooling system is activated using the revised power parameters to determine an advance cooling execution plan. Obtain real-time temperature data from the battery management system, determine the relationship between the real-time temperature data and the thermal safety threshold, and generate an optimized cleaning path sequence; The output of the heat load prediction model is recalculated based on the optimized cleaning path sequence to determine whether the predicted heat load has decreased and to determine a stable temperature control state. By using the stable temperature control state, the parameters of the heat load prediction model are updated using a feedback loop mechanism to determine the final adaptive control result of the thermal safety boundary.

2. The method as described in claim 1, characterized in that, The acquisition of environmental status data from the main control system of the cleaning robot includes: The floor material of the cleaning area is identified by a camera or ultrasonic sensor, and it is determined whether the floor material is a high-friction carpet or a low-friction floor. Obtain the current cleaning mode from the main control system; Based on the ground material and the current cleaning mode, environmental state data characterizing the cleaning area is generated, wherein the environmental state data includes ground material type, friction coefficient and cleaning mode parameters; The environmental condition data is used to determine the operating resistance characteristics of the cleaning robot in the current cleaning area; Based on the operating resistance characteristics, environmental state parameters compatible with the input of the heat load prediction model are generated.

3. The method as described in claim 1, characterized in that, The step of determining the estimated temperature rise curve using a heat load prediction model based on the environmental status data, remaining clean area, and navigation path planning includes: Obtain the remaining cleaning area and the path length and turning frequency in the navigation path planning; Based on the environmental condition data, the power consumption characteristics of the cleaning robot on different ground materials are determined; Using the heat load prediction model, combined with the power consumption characteristics, the path length, and the turning frequency, the battery current output demand in the future period is calculated. Based on the battery current output requirements, a predicted temperature rise curve characterizing the battery temperature change is generated.

4. The method as described in claim 1, characterized in that, The step of adjusting the maximum output power limit of the lithium battery based on the relationship between the estimated temperature rise curve and the preset thermal safety threshold to obtain the corrected power parameters includes: Compare the estimated temperature rise curve with the preset thermal safety threshold to determine whether there is a risk of exceeding the thermal safety threshold; If the aforementioned risks exist, the maximum output power limit will be reduced through the dynamic regulation mechanism of the lithium battery. Calculate the corrected power parameters based on the reduced maximum output power limit.

5. The method as described in claim 4, characterized in that, The step of activating the heat dissipation system using the corrected power parameters and determining an advance heat dissipation execution plan includes: Based on the corrected power parameters, determine the required speed of the cooling fan; Based on the required rotation speed, the heat dissipation system is pre-started, and the cooling fan is turned on at a high speed. Based on the operating parameters of the cooling fan, a pre-heating execution plan is generated; The operating time and intensity of the cooling system are determined by the aforementioned advance cooling execution plan; Based on the running time and intensity, a temperature drop curve characterizing the heat dissipation effect is generated.

6. The method as described in claim 1, characterized in that, The process of acquiring real-time temperature data from the battery management system, determining the relationship between the real-time temperature data and the thermal safety threshold, and generating an optimized cleaning path sequence includes: Obtain real-time temperature data from the battery management system; Compare the real-time temperature data with the preset thermal safety threshold to determine whether it is close to the thermal safety boundary; If the thermal safety boundary is approached, a task scheduling suggestion is sent to the main control system; Based on the task scheduling suggestions, the navigation path planning is adjusted to generate an optimized clean path sequence; The optimized cleaning path sequence reduces the operating time of the cleaning robot in high-heat-load areas.

7. The method as described in claim 1, characterized in that, The step of recalculating the output of the heat load prediction model based on the optimized cleaning path sequence, determining whether the predicted heat load has decreased, and identifying a stable temperature control state includes: The input parameters of the heat load prediction model are updated using the optimized cleaning path sequence. Based on the updated input parameters, the output of the heat load prediction model is recalculated to generate a new predicted heat load; Compare the new predicted heat load with the previous predicted heat load to determine whether the heat load has decreased; If the heat load decreases, the current power parameters and heat dissipation system settings are maintained. The temperature control state characterizing stable battery operation is determined by the maintained settings.

8. The method as described in claim 1, characterized in that, The step of updating the parameters of the heat load prediction model using a feedback loop mechanism through the stable temperature control state to determine the final adaptive control result of the thermal safety boundary includes: Real-time operating data of the battery is obtained through the stable temperature control state; Based on the real-time operating data, the parameters of the heat load prediction model are adjusted using a feedback loop mechanism; By recalculating the output of the heat load prediction model using the adjusted parameters, an updated thermal safety boundary is generated. Based on the updated thermal safety boundary, an adaptive control strategy is determined; The adaptive control strategy is used to generate the final adaptive control result of the thermal safety boundary.

Images