Urban security-oriented robot dog autonomous obstacle avoidance method and system

Abstract

The present application relates to the technical field of intelligent robot motion control and state evaluation, and particularly relates to a robot dog autonomous obstacle avoidance method and system for urban security, the method comprising the following steps: identifying events that occur during the movement of the robot dog and cause small movement deviations of internal precision components, and adjusting parameters representing the movement execution capability state of the robot dog according to the identification results of the events; when the parameters indicate that the movement execution capability has deviations, switching the action mode of the robot dog, so that the robot dog performs exploratory actions before executing tasks that require accurate movement; during the execution of the exploratory actions, obtaining difference information between the movement instructions and the actual movement responses of the robot dog; through the above scheme, the obstacle avoidance behavior mode of the robot dog can be dynamically adjusted, the small movement deviations of internal precision components can be effectively identified and quantified, and the cumulative trajectory deviation and secondary collision risk can be avoided.

Description

Technical Field

[0001] This invention relates to the field of intelligent robot motion control and state assessment technology, and in particular to a method and system for autonomous obstacle avoidance by a robot dog for urban security. Background Technology

[0002] In urban security, the core value of robotic dogs lies in their ability to autonomously perform tasks in complex environments, with obstacle avoidance being the foundation for their safe and efficient operation. During routine patrols, obstacle avoidance relies on preset rules, and with training, they can effectively identify and bypass common static and dynamic objects. However, the complexity of urban environments far exceeds the coverage of preset rules. Sometimes, unexpected dynamic targets may appear, or when the robotic dog responds to emergencies, its internal precision components may suffer "silent" damage, which is difficult to detect through routine self-inspection procedures, ultimately affecting subsequent precise actions. For example, in "nighttime-only" commercial pedestrian streets, although the robotic dog's obstacle avoidance system has been adjusted and optimized, unexpected situations may still occur: a food delivery rider ignores the prohibition and speeds across on an electric scooter. This target is "high-speed, small in size, and has an irregular path," which is significantly different from the norm. The system will identify it as a high-risk object that needs to be avoided at the highest safe distance immediately to ensure safety.

[0003] In response to this highest-priority avoidance command, the robot dog's path planning module immediately calculates an obstacle avoidance path away from the rider's predicted trajectory. This often requires the robot dog to quickly move laterally towards the area in front of a shop to escape the potential collision risk zone in the shortest possible time. This rapid lateral movement places extremely high demands on the robot dog's motion control system, requiring it to precisely adjust leg posture and driving force in a very short time to achieve rapid and stable displacement. When performing rapid lateral movement, the robot dog may encounter difficult-to-detect ground conditions. For example, shallow depressions formed in the dark antique-style bricks in front of the shop due to use or subsidence, and water accumulation after rain, can be difficult for visual sensors and lidar to distinguish from wet brick surfaces under nighttime lighting, challenging its multimodal perception capabilities.

[0004] If the robot dog accidentally steps into a shallow puddle while performing a rapid lateral movement on one leg, it may slip instantly, causing a shift in its center of gravity and loss of balance. To prevent a fall, the motion control system will output a peak compensation torque to the leg drive motor within milliseconds, far exceeding the normal gait adjustment, to forcefully stabilize the robot's posture and restore balance. From a system perspective, this violent impact and compensation will be recorded as a successful "post-stress instability recovery" event, indicating that the robot dog successfully coped with the sudden situation and avoided falling. However, this seemingly successful posture recovery may cause "silent" damage to the robot dog's internal precision mechanical structure. Although the extreme compensation action in the slippage event does not directly cause hardware breakage or obvious failure, it causes impact overload to the precision transmission mechanism of the leg joint (such as the internal planetary gear reducer), accelerating gear backlash wear and introducing a minute transmission backlash that is difficult to detect in normal movement. This hysteresis is a tiny delay or deviation between the control command and the actual motion output. Its magnitude is extremely small, and during normal straight-line walking or sharp turns, its impact is masked by the control system's routine feedback adjustments, preventing the triggering of abnormal alarms for moving parts. For example, a command to move forward 10 centimeters would normally result in a precise 10-centimeter movement, but with a slight hysteresis, it might only move 9.98 centimeters. The control system will make minor adjustments to compensate, and neither external observers nor system self-checks can detect the internal wear. Ultimately, this "silent" physical damage creates a deep-seated technical dilemma: the robot dog's decision-making and perception systems function normally, but the accuracy of its leg movements has been "silently" impaired, undetectable by routine self-checks. Therefore, when performing tasks requiring precise movement, decisions based on the "health" model will generate cumulative trajectory deviations, making obstacle avoidance actions clumsy and dangerous, significantly increasing the risk of secondary collisions.

[0005] While existing technologies have made progress in addressing how to avoid unexpected external obstacles, they still present significant challenges in dynamically and tentatively adjusting obstacle avoidance behavior patterns to ensure the absolute safety of subsequent tasks, especially when the physical performance of the device may have declined due to wear and tear and this decline cannot be directly quantified.

[0006] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0007] The purpose of this invention is to address the shortcomings of existing technologies by proposing an autonomous obstacle avoidance method and system for robot dogs in urban security.

[0008] In a first aspect, the present invention provides a method for autonomous obstacle avoidance by a robotic dog for urban security, the method comprising the following steps: The system identifies events that occur during the robot dog's movement that cause minute motion deviations in its internal precision components, and adjusts parameters characterizing the robot dog's motion execution capability based on the identification results of these events. When the parameter indicates a deviation in the motion execution capability, the robot dog's action mode is switched so that the robot dog performs exploratory actions before performing tasks that require precise motion. During the execution of the exploratory action, the difference information between the robot dog's movement command and the actual movement response is obtained, and the minute movement deviation is quantified based on the difference information; Based on the quantified motion deviation, the robot dog's subsequent motion commands are corrected to ensure the accuracy of the robot dog when performing precise motion tasks; Based on the execution effect of the revised motion commands, the parameters characterizing the robot dog's motion execution capability are adjusted.

[0009] The above approach allows for dynamic and exploratory adjustment of the robot dog's obstacle avoidance behavior, effectively identifying and quantifying minute motion deviations in internal precision components, and correcting motion commands accordingly. This ensures that even when the robot dog's physical execution capabilities may decline, it can still guarantee the accuracy and safety of subsequent precise motion tasks, avoiding cumulative trajectory deviations and the risk of secondary collisions.

[0010] The core innovation of this application lies in the introduction of dynamically adjusted motion execution capability state parameters and a "confirmation-execution" separate action mode. This enables the robot dog to proactively perform probing actions to quantify these hidden deviations in real time when there may be minor motion deviations in the internal precision components. Based on this, subsequent motion commands are corrected, thus solving the problem that traditional methods cannot detect and compensate for the decrease in motion accuracy caused by "silent" internal damage. This achieves the effect of ensuring that the robot dog can maintain accuracy and reliability when performing precise motion tasks.

[0011] Secondly, a robotic dog autonomous obstacle avoidance system for urban security is provided, comprising: The event recognition module is used to identify events that occur during the robot dog's movement that cause minor motion deviations in its internal precision components, and to adjust the parameters characterizing the robot dog's motion execution capability based on the event recognition results. The action switching module is used to switch the action mode of the robot dog when the parameters indicate that there is a deviation in the motion execution capability, so that the robot dog performs exploratory actions before performing tasks that require precise motion. The deviation quantization module is used to acquire the difference information between the robot dog's motion command and the actual motion response during the execution of the trial action, and to quantify the small motion deviation based on the difference information. The instruction correction module is used to correct the robot dog's subsequent motion instructions based on the quantified motion deviation, so as to ensure the accuracy of the robot dog when performing precise motion tasks. The parameter adjustment module is used to adjust the parameters that characterize the robot dog's motion execution capability state based on the execution effect of the corrected motion commands.

[0012] The above scheme provides a system for implementing the above method. Through modular design, the method can be specifically implemented and deployed, and has good operability and integrability.

[0013] Compared with the prior art, the present invention has the following beneficial effects: This application provides an autonomous obstacle avoidance method and system for robot dogs in urban security applications. It overcomes the shortcomings of existing technologies that cannot effectively identify and quantify the minute motion deviations caused by wear and tear on the precision components inside the robot dog, which leads to cumulative trajectory deviations and secondary collision risks when performing precise tasks. This method and system can dynamically and tentatively adjust the obstacle avoidance behavior of the robot dog, effectively identify and quantify the minute motion deviations of the precision components, and correct the motion commands accordingly. This ensures that even when the physical execution capability of the robot dog may decline, it can still guarantee the accuracy and safety of subsequent precise motion tasks, and avoid cumulative trajectory deviations and secondary collision risks. Attached Figure Description

[0014] Figure 1 This is a flowchart of the method of the present invention.

[0015] Figure 2 This is a schematic diagram of the system structure of the present invention.

[0016] In the diagram: 201, Event Recognition Module; 202, Action Switching Module; 203, Deviation Quantification Module; 204, Command Correction Module; 205, Parameter Adjustment Module. Detailed Implementation

[0017] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0019] like Figure 1The method shown is an autonomous obstacle avoidance method for a robot dog for urban security, which includes the following steps: S101. Identify events that occur during the robot dog's movement that cause minor motion deviations in its internal precision components, and adjust the parameters characterizing the robot dog's motion execution capability based on the event identification results. S102. When the parameters indicate a deviation in the motion execution capability, switch the robot dog's action mode so that the robot dog can perform exploratory actions before performing tasks that require precise motion. S103. During the execution of the trial action, obtain the difference information between the robot dog's motion command and the actual motion response, and quantify the small motion deviation based on the difference information. S104. Based on the quantified motion deviation, correct the robot dog's subsequent motion commands to ensure the accuracy of the robot dog when performing precise motion tasks; S105. Based on the execution effect of the revised motion command, adjust the parameters characterizing the robot dog's motion execution capability.

[0020] This application aims to address the issue of autonomous obstacle avoidance accuracy in robotic dogs caused by minute motion deviations in their internal precision components during movement. Identifying events that cause these minor deviations refers to monitoring the robot's operational status to detect specific situations that may lead to subtle, imperceptible motion deviations in its internal mechanical structure. This can be achieved by analyzing the robot's sensor data, system logs, or motion control feedback information. For example, monitoring the instantaneous operating status of joint drive components, rapid changes in body posture, or instantaneous deviations between motion commands and actual responses is crucial for timely detection of potential damage, providing a trigger for subsequent self-diagnosis and adjustments. Adjusting parameters characterizing the robot's motion execution capability involves modifying one or more internal state variables based on the identified events. These variables reflect the accuracy and reliability of the robot's current motion execution. This can be done using preset rules, lookup tables, or algorithms based on event severity. For example, lowering a "physical credibility" parameter from a high value to a low value quantifies the potential decline in the robot's motion capability, serving as a basis for subsequent behavior mode switching. Switching the robot dog's action mode refers to changing the robot dog's basic behavioral logic when performing tasks, transitioning from one mode to another. This can be achieved through logical switching of the software control module, such as switching from a conventional "command-execution" mode to a separate "confirmation-execution" mode. This is primarily to introduce additional verification steps to reduce task execution risks when there is uncertainty in the robot dog's movement capabilities. Performing exploratory movements refers to the robot dog actively engaging in a series of small-range, controlled preparatory movements before performing tasks requiring high-precision motion. This can be achieved by controlling specific joints or limbs of the robot dog to reciprocate or slightly probe within a preset, minute range. This is mainly to provide observable physical processes for real-time measurement of the robot dog's movement deviations. Obtaining information on the difference between the robot dog's movement commands and actual movement responses refers to collecting deviation data between the commands executed and the actual movements performed during the execution of exploratory movements. This can be achieved by using high-precision encoders, inertial measurement units, or other motion sensors to acquire data in real time and calculate the instantaneous deviation between the expected and actual motion parameters. Its main purpose is to directly quantify motion hysteresis or backlash caused by internal damage. Quantifying minute motion deviations refers to converting the acquired difference information into specific numerical values ​​to accurately represent the degree of deviation in the robot dog's motion execution. This can be achieved using data processing algorithms, such as calculating average hysteresis, phase difference, or amplitude attenuation, primarily to provide accurate compensation values ​​for subsequent motion command corrections. Correcting the robot dog's subsequent motion commands refers to adjusting the motion commands the robot dog will execute next based on the quantified motion deviations.This can be achieved by modifying relevant parameters of the motion control strategy, such as changing the response intensity of joint movements, adjusting the degree of influence of error accumulation, or correcting the coefficients of predictive correction. The main purpose is to ensure, through a compensation mechanism, that the robot dog's final actual movement still achieves the expected accuracy even with internal biases. Adjusting the parameters characterizing the robot dog's motion execution capability refers to further modifying these parameters based on the execution effect of the corrected motion commands. This can be achieved by evaluating actual motion performance information and combining it with environmental information and sensor operating status information for a comprehensive judgment. For example, slightly increasing the "physical reliability" value after successfully completing a task. This is mainly to form a closed-loop feedback mechanism, enabling the robot dog to dynamically adjust its own capability assessment based on actual performance, thereby optimizing its long-term operational reliability.

[0021] This application's solution employs a closed-loop adaptive mechanism to enable the robot dog to dynamically respond to minute motion deviations that may arise from its internal precision components, thereby ensuring motion accuracy during urban security patrols. First, the system continuously identifies events occurring during the robot dog's movement that could potentially cause minute motion deviations in its internal precision components. The identification results of these events, such as monitoring the instantaneous operating state of joint drive components, rapid changes in body posture, or instantaneous deviations between motion commands and actual responses, are immediately used to adjust parameters characterizing the robot dog's motion execution capability. This parameter adjustment, for example, lowering a "physical confidence" value, reflects a potential decline in the robot dog's motion capability and serves as the basis for subsequent behavior mode switching. When this parameter indicates a deviation in the robot dog's motion execution capability, the system intelligently switches the robot dog's action mode. This switching transforms the robot dog from a conventional direct "command-execution" mode to a "confirmation-execution" separated mode. In this new mode, when the robot dog needs to perform a precise motion task, it does not directly execute the complete action but first performs a series of small, controlled, exploratory movements. These exploratory actions, such as controlling specific joints to swing back and forth within a preset range, aim to provide an observable physical process for acquiring information on the difference between the robot dog's motion commands and actual motion responses. During the execution of these exploratory actions, the system acquires real-time information on the difference between the robot dog's commanded motion and its actual completed motion. This difference information, such as the instantaneous deviation between the expected and actual motion parameters collected and calculated by high-precision sensors, forms the basis for directly quantifying minute motion deviations. Subsequently, the system precisely quantifies these minute motion deviations based on this difference information, converting them into specific values, such as average hysteresis or phase difference. Based on this quantified motion deviation, the system immediately corrects the robot dog's subsequent motion commands. This means that when the robot dog performs actual, complete obstacle avoidance or path tracking actions, its motion control system pre-calculates this measured deviation and compensates for the original motion commands, for example, by modifying relevant parameters of the motion control strategy, to ensure the accuracy of the robot dog when performing precise motion tasks. Finally, to form a complete feedback loop, the system readjusts the parameters characterizing the robot dog's motion execution capability based on the execution effect of the corrected motion commands. For example, by evaluating actual motion performance information and combining it with environmental information and the sensor's own operating status information for a comprehensive judgment, if the corrected command execution effect meets expectations, the parameter value can slightly increase. This dynamic adjustment mechanism enables the robot dog to continuously learn and adapt, ensuring that it maintains the reliability of its movement and the completion of its tasks in complex and ever-changing environments.

[0022] As one embodiment of the present invention, during the execution of the exploratory action, the step of acquiring the difference information between the robot dog's motion command and the actual motion response, and quantifying the minute motion deviation based on the difference information, includes: During the execution of exploratory actions, information on the difference between the robot dog's movement commands and actual movement responses is obtained; at the same time, data on the sensor's own working status in the current environment is acquired. Based on its own work status data, the discrepancy information is calibrated; Then, based on the calibrated difference information, the minute motion deviations are quantified.

[0023] The term "sensor" refers to a device used to sense the robot dog's motion state or environmental parameters. Specifically, it can include, but is not limited to, encoders, gyroscopes, and accelerometers used to measure joint angles, speeds, and accelerations, as well as environmental sensors used to measure ambient temperature, humidity, air pressure, or electromagnetic field strength. Its purpose is to provide basic data on the robot dog's motion response and environmental conditions. "Self-operating state data" refers to the internal or external influencing parameters exhibited by the sensor under the current operating environment. Specifically, it can include the sensor's internal temperature, supply voltage, operating current, and the electromagnetic field strength or humidity of its environment. Its purpose is to reflect potential sources of deviation in the sensor's current output data. "Calibration of discrepancy information" refers to correcting or compensating for the discrepancy between the originally acquired motion command and the actual motion response based on the sensor's own operating state data. Specifically, this can be achieved by adjusting the original discrepancy information using a pre-established performance model or calibration curve of the sensor under different operating states. For example, if the sensor has a known deviation at a specific temperature, the discrepancy information is corrected accordingly based on the current temperature data. Its purpose is to eliminate or reduce the influence of the sensor's own characteristics or environmental factors on the accuracy of the discrepancy information, thereby obtaining more accurate motion deviation data.

[0024] This application's solution, during the execution of exploratory actions, not only acquires information on the difference between the robot dog's motion commands and actual motion responses, but also simultaneously acquires the sensor's own operating status data in the current environment. Based on this operating status data, the system calibrates the difference information and finally quantifies minute motion deviations using the calibrated difference information. This coordinated operation of steps enables the robot dog to overcome measurement biases that may arise from sensors in different environments, thereby obtaining more accurate motion deviation quantification results. Specifically, when the robot dog's motion execution capability may have deviations, requiring exploratory actions to quantify its minute motion deviations, the system first collects the raw difference between the motion command and the actual response. However, considering that the sensor's performance is affected by environmental factors such as temperature, voltage, and electromagnetic interference, these factors may lead to inaccurate sensor output data, thus preventing the raw difference information from accurately reflecting the robot dog's motion deviation. Therefore, this solution introduces the acquisition of the sensor's own operating status data, such as real-time monitoring of the sensor's internal temperature or power supply voltage. Subsequently, the system uses this operating status data to intelligently calibrate the previously acquired raw difference information. This calibration process, based on a pre-established sensor characteristic model or calibration curve, compensates for or corrects the original difference information, effectively filtering out or reducing measurement errors caused by changes in the sensor's own state. The accuracy of the calibrated difference information is improved, more accurately reflecting the robot dog's actual motion deviation. Ultimately, based on this precisely calibrated difference information, the system can accurately quantify minute motion deviations. This accurate quantification result directly provides a reliable basis for subsequent correction of the robot dog's motion commands, ensuring the accuracy of the robot dog when performing precise motion tasks. This consideration and calibration mechanism of the sensor's own state refines and enhances the robustness of the motion deviation quantification link in the robot dog's autonomous obstacle avoidance method. This allows the entire obstacle avoidance system to maintain a high level of motion accuracy and reliability in complex and changing environments, effectively solving the problem of inaccurate deviation quantification caused by unstable sensor operating states in different environments.

[0025] In one embodiment of the present invention, after executing the step of correcting the subsequent movement commands of the robot dog, the method further includes: Continuously acquire the location information and motion status of moving targets in the current environment; Based on location information and motion status, detect whether the position or motion status of a moving target changes instantaneously; When a transient change is detected, the motion parameters in the corrected motion command are adjusted in real time based on the transient change, taking into account the small motion deviations that were previously quantified.

[0026] Continuously acquiring the position and motion status of moving targets in the current environment refers to the continuous collection of real-time position and motion trend data of moving objects in the surrounding environment, such as pedestrians and vehicles, through various sensors on the robot dog, such as vision sensors, lidar, or ultrasonic sensors. This includes their speed, direction, and acceleration, aiming to provide the robot dog with the latest perception information of the dynamic environment. Detecting whether the position or motion status of a moving target has undergone instantaneous changes involves analyzing the continuously acquired moving target data to determine whether its position or motion status has undergone significant and unexpected changes in a short period of time, such as sudden acceleration, abrupt stop, turning, or the appearance of a new target. This detection can be based on preset thresholds or pattern recognition algorithms, aiming to identify sudden situations requiring immediate response from the robot dog. Real-time adjustment of motion parameters in the corrected motion commands means that after detecting an instantaneous change in the moving target, the robot dog's motion control system immediately modifies the currently executing or about-to-be-executed motion commands. These motion parameters can include the robot dog's speed, direction, gait frequency, joint angles, or waypoints, aiming to ensure that the robot dog's trajectory can quickly adapt to the latest changes in the environment, ensuring the effectiveness and safety of obstacle avoidance. The adjustment process takes into account previously quantified minor motion deviations. This means that when making real-time adjustments to motion commands, not only should the instantaneous changes of the moving target be considered, but also the minor motion deviations caused by internal wear and tear, which were previously quantified through the robot dog's trial movements. This implies that in the calculation of new motion parameters, this known self-deviation will be superimposed or compensated. The purpose is to ensure that even during dynamic adjustments, the robot dog's motion accuracy can be guaranteed, avoiding new risks caused by the accumulation of its own deviations.

[0027] This application's solution, based on the robot dog's correction of its own motion commands, further introduces a real-time response mechanism to dynamic changes in the urban environment. Specifically, the robot dog continuously acquires the position information and motion status of surrounding moving targets, providing real-time data for subsequent decision-making. Based on this, the system detects whether the position or motion status of the moving targets has undergone instantaneous changes. This detection mechanism is key to triggering subsequent adjustments, avoiding unnecessary calculations and adjustments, thereby improving system efficiency. When such an instantaneous change is detected, the robot dog immediately adjusts the motion parameters in its corrected motion commands. This immediate adjustment allows the robot dog to dynamically adjust its trajectory based on the latest environmental information, ensuring the accuracy and safety of obstacle avoidance. Notably, the solution also considers previously quantified minor motion deviations during this immediate adjustment. This means that while responding to changes in the external environment, the robot dog also takes into account its own potential motion deviations, thus achieving more precise motion control. In this way, the solution not only solves the problem of command failure caused by instantaneous changes in moving targets in the urban environment when the robot dog executes the corrected movement commands, but also enables the robot dog to exhibit higher adaptability and robustness in the complex and ever-changing urban security environment by combining real-time perception of the external environment with precise compensation of its own internal state, effectively reducing the risk of secondary collisions.

[0028] As one embodiment of the present invention, the step of adjusting the parameters characterizing the motion execution capability state of the robot dog based on the execution effect of the modified motion command includes: Acquire motion performance information of the robot dog when executing the revised motion commands; At the same time, it acquires the lighting information and ground material information of the robot dog's current environment, as well as the working status information of the sensors used to acquire motion performance information; Based on the lighting information, ground material information, and the sensor's own operating status information, the motion performance information is processed to obtain the execution effect of the corrected motion command; Then, based on the execution results, adjust the parameters that characterize the robot dog's motion execution capabilities.

[0029] Motion performance information refers to data such as the actual movement trajectory, speed, acceleration, posture, torque output, and energy consumption of various parts of the robot dog's body when executing specific motion commands. This data can be characterized using real-time data acquired by various sensors, such as inertial measurement units (IMUs), joint encoders, vision sensors, force sensors, or motor current sensors, and its purpose is to directly reflect the robot dog's actual motion state. Sensor operating status information refers to the data on the operating status of the sensors used to acquire motion performance information in the current environment. This can include sensor calibration status, noise level, data integrity, operating temperature, power supply voltage, or internal error flags, and its purpose is to evaluate the reliability and accuracy of the sensor data. Processing motion performance information involves performing a series of analyses, calibrations, filtering, or weighting operations on the raw motion performance data. This can be achieved using data fusion algorithms, environmental adaptability models, or machine learning-based correction models. Its purpose is to eliminate or reduce the influence of environmental factors and the sensor's own state on the motion performance data, thereby obtaining a more realistic and accurate picture of the motion execution. The execution effect of the corrected motion command refers to the consistency or deviation between the robot dog's actual movement and the corrected motion command after correction for environmental factors and sensor status. This can be expressed as the deviation between the actual trajectory and the target trajectory, the difference between the actual speed and the target speed, or the matching degree between the actual posture and the target posture. Its purpose is to provide an accurate motion performance evaluation result unaffected by external interference. The parameters characterizing the robot dog's motion execution capability state refer to one or a set of internal state variables used to quantify the robot dog's current physical motion capability and reliability. These can be a physical reliability value, a motion accuracy index, a wear factor, or a dynamic response coefficient, and their purpose is to guide the robot dog's subsequent action mode switching and motion command adjustments.

[0030] This application's solution refines the adjustment process of the robot dog's motion execution capability state parameters, enabling the robot dog to more accurately assess its motion performance in complex environments, thereby improving its autonomous obstacle avoidance reliability. Specifically, when the robot dog executes a modified motion command, the system first acquires its motion performance information, which directly reflects the robot dog's actual movement. However, relying solely on this raw motion performance information for parameter adjustment is insufficient, as environmental lighting conditions, ground material, and the operating status of the sensors used to acquire this information can all interfere with the accuracy of the motion performance information. For example, in insufficient lighting or on slippery ground, even if the robot dog's motion execution capability does not actually decrease, its motion performance information may still show deviations; similarly, if the sensors themselves have noise or calibration problems, their output data may also be inaccurate. Therefore, this solution further acquires the lighting and ground material information of the robot dog's current environment, as well as the operating status information of the sensors used to acquire motion performance information. This auxiliary information is used to process the raw motion performance information, calibrating, weighting, or correcting it to obtain a more realistic and accurate modified motion command execution effect. For example, in poor lighting conditions, the weight of visual sensor data can be reduced; when ground friction is insufficient, compensation can be made for potential deviations in motion performance caused by leg slippage; and when the sensor itself is in poor condition, its output data can be calibrated or discarded. This processing ensures that the evaluation of motion execution is no longer affected by external interference or the limitations of the sensor itself, but truly reflects the robot dog's motion execution capabilities. Ultimately, the system adjusts the parameters characterizing the robot dog's motion execution capabilities based on this precisely processed and corrected execution effect. This adjustment is based on more reliable evaluation results, thus more reasonably reflecting the robot dog's actual motion capabilities, such as its physical reliability or motion accuracy index. This refined parameter adjustment mechanism forms a close logical connection with other steps in the robot dog's autonomous obstacle avoidance method. In the robot dog's autonomous obstacle avoidance method, when an event that may cause minor motion deviations in internal precision components is identified, the parameters characterizing the robot dog's motion execution capabilities are initially adjusted, and the system may switch action modes to perform tentative actions to quantify the deviation, thereby correcting subsequent motion commands. The precise evaluation and parameter adjustment of the corrected motion command execution effect provided by this solution constitutes the closed-loop feedback loop of the entire adaptive cycle. It ensures that after the robot dog experiences internal damage and performs self-correction, it can accurately evaluate the correction effect and continuously optimize its perception of its own motion capabilities accordingly.This accurate feedback mechanism enables the robot dog to dynamically and continuously adjust its "confidence" in its own physical capabilities. As a result, when faced with obstacle avoidance tasks that require precise positioning, it can make decisions based on its more realistic self-state, effectively avoiding cumulative trajectory deviations and significantly improving the robot dog's continuous safe operation capability and task reliability in complex urban security environments.

[0031] As one embodiment of the present invention, the step of identifying events that occur during the movement of the robot dog, causing minute movement deviations in its internal precision components, includes: Acquire instantaneous operating status information of the joint drive components; Obtain information on the attitude changes of the robot dog's body; Acquire instantaneous motion response information of the joint; Based on instantaneous operating status information, attitude change information, and instantaneous motion response information, determine whether the instantaneous operating status information indicates the presence of a current peak, the attitude change information indicates the presence of rapid instability and recovery of the fuselage attitude, and the instantaneous motion response information indicates the presence of an instantaneous deviation between the motion command and the actual motion, and whether these conditions are simultaneously met. When the judgment result meets the preset conditions, it is identified as an event that causes a slight movement deviation in the internal precision components.

[0032] The instantaneous operating status information refers to the operating data of the joint drive components at a specific moment, specifically parameters such as current, voltage, temperature, or torque of the joint drive components. Its purpose is to reflect the immediate load or abnormal conditions of the joint drive components during movement. The attitude change information refers to the changes in angle, angular velocity, or angular acceleration of the robot dog's body during movement. Specifically, it can be the body tilt angle, roll speed, or yaw rate obtained through the inertial measurement unit (IMU). Its purpose is to reflect whether the robot dog's body has experienced drastic changes in motion state or impacts. The instantaneous motion response information refers to the actual motion performance data of the joints within a very short time after receiving a motion command. Specifically, it can be the actual angle, angular velocity, position, or torque feedback of the joints. Its purpose is to reveal whether there is an instantaneous deviation or lag between the motion command and the actual motion. The preset conditions refer to the logical rules or numerical threshold combinations used to determine whether an event has occurred. Specifically, it can be a multi-parameter joint judgment standard set through experimental data, historical fault mode analysis, or expert experience. Its purpose is to ensure accurate identification of events that cause minute motion deviations in internal precision components, avoiding misjudgments or omissions.

[0033] This solution achieves accurate identification of events causing minute motion deviations in internal precision components by comprehensively analyzing multi-dimensional instantaneous data during the robot dog's movement. First, it acquires instantaneous operating status information of the joint drive components, such as peak current, which directly reflects whether the joint has experienced abnormal impact or overload. Simultaneously, it acquires information on the robot dog's posture changes, such as rapid instability and recovery, which helps determine whether the robot dog has experienced severe external impact or internal stress. Furthermore, it acquires instantaneous motion response information of the joints, such as the instantaneous deviation between the motion command and the actual motion, which directly reveals whether there is minute hysteresis or backlash in the transmission mechanism. It is precisely by organically combining these three seemingly independent types of instantaneous information and determining whether they simultaneously meet preset conditions that misjudgments or omissions that might occur from a single information source can be effectively avoided. For example, a peak current alone might be caused by a normal large load, and instability in the robot dog's posture might be caused by external thrust, but when all three occur simultaneously, it strongly indicates that the internal precision components may have suffered impact damage. This multi-dimensional, real-time data fusion and judgment mechanism enables the system to accurately capture subtle damage events that are difficult to detect through routine self-inspection. This accurate event identification provides a reliable basis for subsequent adjustments to parameters characterizing the robot dog's motion execution capabilities, ensuring that the robot dog can promptly perceive changes in its own state during urban security patrols. This lays the foundation for subsequent action mode switching and motion command correction, thereby enhancing the robot dog's ability to operate safely and continuously in complex environments.

[0034] As one embodiment of the present invention, the step of the robot dog performing exploratory movements before performing a task requiring precise movement includes: Control specific leg joints of the robot dog to make it swing back and forth within a preset small angle range in order to perform exploratory movements; During the reciprocating swing, the difference between the motion command of a specific leg joint and the actual swing angle is recorded.

[0035] In this context, "specific leg joint" refers to the joint in the robot dog's legs selected for performing exploratory movements, such as the hip, knee, or ankle joints. This can be achieved using any one or more leg joints. "Preset micro-angle range" refers to the range of joint swing angles during exploratory movements. This range is predetermined and has a small amplitude; it can be achieved using a fixed angle range smaller than the robot dog's normal gait swing angle. "Reciprocating swing" refers to the repetitive back-and-forth movement of the joint within the preset micro-angle range, which can be achieved by driving the joint motor with periodic sine or square wave control signals. "Motion command" refers to the command issued by the control system to the specific leg joint, specifying the desired motion state or position. This can be achieved using a target angle value or a target angular velocity value. "Actual swing angle" refers to the real-time motion angle of the specific leg joint when executing motion commands, measured in real-time by sensors. This can be achieved using real-time data obtained from a high-precision encoder or angle sensor. The difference refers to the numerical deviation between the motion command and the actual swing angle. This deviation reflects the accuracy of the joint motion execution and can be expressed as the difference or relative percentage between the command value and the actual measured value.

[0036] This application's solution controls specific leg joints of a robotic dog, causing them to reciprocate within a preset, minute angle range. During this reciprocating motion, the difference between the commanded movement and the actual angle is recorded, effectively quantifying deviations in the robotic dog's motion execution capability. Specifically, specific leg joints are chosen for this exploratory movement because they are critical components for the robotic dog's motion execution; their movement deviations directly affect the overall motion accuracy. Targeted testing of specific joints avoids unnecessary testing of all joints, improving testing efficiency. Furthermore, the small-amplitude reciprocating motion within the preset minute angle range avoids excessive interference with the robotic dog's normal tasks while stimulating joint movement responses, thus detecting minute motion deviations. By recording the difference between the commanded movement and the actual angle during the reciprocating motion, the joint's motion accuracy can be accurately assessed, providing a basis for subsequent motion command corrections. Combined with the robotic dog's autonomous obstacle avoidance method, when deviations are indicated in the robotic dog's motion execution capability parameters, the robot switches to a cautious action mode and performs self-evaluation using the exploratory movements provided by this solution before executing tasks requiring precise movement. This probing action can precisely reveal, in a non-invasive manner, the minute motion lags or backlashes caused by wear and tear on the robot dog's internal precision components. In this way, the robot dog no longer relies solely on external performance or cursory self-checks, but can proactively and in real-time acquire the true state of its physical execution capabilities. This precise quantification of deviations allows the robot dog to make compensatory corrections to subsequent motion commands based on reliable data, ensuring that even with minor internal damage, the robot dog can still accurately and reliably complete its obstacle avoidance tasks, avoiding cumulative trajectory deviations and significantly improving its robustness and mission reliability in complex urban security environments.

[0037] As one embodiment of the present invention, the step of obtaining the difference information between the robot dog's motion command and the actual motion response during the execution of the trial action includes: During the execution of exploratory actions, the expected movement position information corresponding to the robot dog's movement commands is acquired in real time; At the same time, the actual movement position information corresponding to the actual movement response of the robot dog is obtained in real time; Based on the expected motion position information and the actual motion position information, the instantaneous deviation between the two is calculated to obtain the difference information.

[0038] The expected motion position information refers to the spatial position or posture data that the robot dog should ideally achieve after receiving a specific motion command. It can be generated using kinematic models or preset trajectory data, and its purpose is to provide an accurate benchmark for subsequent motion deviation calculations. The actual motion position information refers to the actual spatial position or posture data detected in real time by the robot dog's own sensors during the execution of motion commands. It can be acquired using high-precision encoders, inertial measurement units, or visual positioning systems, and its purpose is to accurately reflect the robot dog's motion state in the current environment. The instantaneous deviation refers to the quantitative difference between the expected motion position information and the actual motion position information at a specific moment. It can be calculated using Euclidean distance, angle difference, or posture error matrix, and its purpose is to accurately measure the instantaneous deviation of the robot dog's motion. The difference information refers to the data set that reflects the robot dog's motion deviation, obtained by calculating the instantaneous deviation between the expected and actual motion position information. It can include the value, direction, or trend of the deviation, and its purpose is to provide an accurate data basis for subsequent motion deviation quantification and command correction.

[0039] This application's solution first acquires the expected motion position information corresponding to the robot dog's motion command in real time during the execution of exploratory actions, setting a clear ideal benchmark for subsequent deviation calculations. Simultaneously, it acquires the actual motion position information corresponding to the robot dog's actual motion response in real time, accurately reflecting the robot dog's motion performance in the current environment. Then, based on the expected and actual motion position information, the instantaneous deviation between the two is calculated, thus obtaining the difference information. This method enables the robot dog to accurately quantify the minute deviations between its motion commands and actual motion responses, overcoming the potential real-time and accuracy deficiencies of traditional methods in acquiring difference information. It is precisely this precise, real-time difference information acquisition mechanism that allows the robot dog to accurately detect minute motion lags or backlashes caused by wear of internal precision components during exploratory actions. This capability, combined with the steps of acquiring difference information and quantifying minute motion deviations in the basic solution, makes the subsequent process of quantifying minute motion deviations based on difference information more accurate and reliable. In this way, the robot dog is no longer just making a preliminary assessment of motion deviations, but can accurately quantify minute motion deviations, providing a solid data foundation for subsequent correction of motion commands and improvement of motion accuracy. This ensures the accuracy of the robot dog when performing precise motion tasks and effectively avoids the occurrence of cumulative trajectory deviations.

[0040] As one embodiment of the present invention, the step of correcting the robot dog's subsequent motion commands based on the quantified motion deviation includes: Based on the quantified motion deviation, adjust the relevant parameters in the robot dog's motion control strategy.

[0041] Among them, the quantified motion deviation refers to the specific value that reflects the wear or lag of internal precision components, accurately measured by comparing the motion command with the actual motion response during the robot dog's exploratory actions. It can be characterized by angular lag, response time extension, or torque output attenuation. The motion control strategy refers to the set of algorithms and control logic used to guide the robot dog to achieve specific motion behaviors. It can adopt various control frameworks such as PID control, torque control, impedance control, or trajectory tracking control. Relevant parameters refer to the values ​​or coefficients in the motion control strategy that can be adjusted to affect the robot dog's motion performance. They can be control parameters such as proportional gain, integral time, derivative time, feedforward coefficient, or joint stiffness.

[0042] This application's solution uses the precisely quantified motion deviation from previous steps as the basis for adjusting relevant parameters in the robot dog's motion control strategy, thereby achieving effective correction of the robot dog's subsequent motion commands. Specifically, when the robot dog's internal precision components experience minute motion deviations due to wear and other factors, and these deviations are accurately quantified, traditional direct compensation methods may not be fully adaptable to complex and changing environments or continuously accumulating deviations. This solution delves into the core of the robot dog's motion control system, dynamically adjusting key parameters in the control algorithm based on the quantified deviation information. For example, if the quantified deviation indicates a response lag in a certain joint, the system can correspondingly increase the proportional gain or feedforward coefficient of the joint controller to improve its response speed and accuracy; if the deviation exhibits oscillation or instability, it may be necessary to adjust the integral or derivative parameters to enhance stability. This fine-grained adjustment of control strategy parameters based on actual deviations enables the robot dog's motion control system to adaptively optimize its performance, rather than simply superimposing or scaling commands. It is precisely this deep optimization of motion control strategies that enables the robot dog to execute corrected motion commands more accurately, effectively overcoming cumulative trajectory deviations caused by internal wear. This significantly improves the accuracy and stability of its movement in various complex environments, such as precise obstacle avoidance in confined spaces or maintaining a stable gait on uneven ground, ensuring reliable mission completion. This approach avoids the limitations that may arise from relying solely on external command corrections, fundamentally enhancing the robot dog's motion execution capabilities.

[0043] As one embodiment of the present invention, the step of adjusting the relevant parameters in the motion control strategy of the robot dog includes: Based on the quantified motion deviation, the response intensity of the controlled joint motion is changed, the degree of influence of error accumulation is adjusted, and the coefficient of predictive correction is modified to adjust the relevant parameters in the robot dog's motion control strategy.

[0044] The response intensity of joint movement control refers to the sensitivity or force with which a joint reacts to received motion commands or deviation signals in the robot dog's motion control system. This can be achieved by adjusting the proportional gain (P gain) in a proportional-integral-derivative (PID) controller, or by adjusting the upper or lower limit of the joint torque output. The degree of error accumulation refers to the weight or cumulative effect of historical motion deviations on the correction of current or future motion commands during the robot dog's motion control process. This can be controlled by adjusting the weight or upper limit of the integral term (I term), or by introducing a forgetting factor or sliding window averaging to limit the cumulative effect of historical errors. The predictive correction coefficient refers to the weight or proportional factor used by the robot dog's motion control system to adjust the output of the predictive model when predicting future motion states and making corrections in advance. This can be achieved by adjusting the gain parameter in feedforward control, or by adjusting the prediction step size or model parameters in algorithms based on Kalman filtering or model predictive control (MPC).

[0045] This application's solution refines the relevant parameters in the robot dog's motion control strategy, thereby more effectively correcting the robot dog's motion commands. Specifically, during the robot dog's movement, when a quantified motion deviation exists between its motion command and the actual motion response, the system adjusts the response intensity of the controlled joints according to the magnitude and characteristics of the deviation. This means that when the deviation is large, the joint response intensity can be appropriately reduced to avoid overreaction and instability caused by the control system; while when the deviation is small, the response intensity can be increased, allowing the system to make more sensitive fine-tuning, thus improving motion accuracy. Simultaneously, the solution adjusts the degree of influence of error accumulation to reasonably handle errors generated during movement. By adjusting the weight of error accumulation according to the magnitude of the deviation, excessive or insufficient error accumulation can be effectively suppressed, preventing the motion trajectory from deviating from the predetermined path. Furthermore, the solution corrects the predictive correction coefficients, enabling the robot dog's control system to more accurately predict future motion states and make corrections in advance. By correcting the prediction coefficients based on the quantified deviation, the accuracy of prediction can be improved, thereby ensuring the accuracy of the robot dog when performing precise motion tasks. It is precisely because of this multi-dimensional, adaptive parameter adjustment mechanism that the robot dog can dynamically optimize its motion control behavior based on its actual motion deviation. Thus, even when there may be minor damage to its physical execution capabilities, it can still complete tasks accurately and reliably, effectively avoiding the occurrence of cumulative trajectory deviations and significantly improving the robot dog's autonomous obstacle avoidance ability and task reliability in complex environments.

[0046] like Figure 2 The system shown is an autonomous obstacle avoidance system for robotic dogs designed for urban security. The system includes: The event recognition module 201 is used to recognize events that occur during the movement of the robot dog, causing minor motion deviations in its internal precision components, and to adjust the parameters characterizing the robot dog's motion execution capability based on the event recognition results. The action switching module 202 is used to switch the robot dog's action mode when the parameters indicate that there is a deviation in the motion execution capability, so that the robot dog can perform exploratory actions before performing tasks that require precise motion. The deviation quantization module 203 is used to acquire the difference information between the robot dog's motion command and the actual motion response during the execution of the trial action, and to quantify the small motion deviation based on the difference information. The instruction correction module 204 is used to correct the robot dog's subsequent motion commands based on the quantized motion deviation, so as to ensure the accuracy of the robot dog when performing precise motion tasks. The parameter adjustment module 205 is used to adjust the parameters that characterize the robot dog's motion execution capability based on the execution effect of the corrected motion commands.

[0047] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A method for autonomous obstacle avoidance by a robotic dog for urban security, characterized in that, The method includes the following steps: The system identifies events that occur during the robot dog's movement that cause minute motion deviations in its internal precision components, and adjusts parameters characterizing the robot dog's motion execution capability based on the identification results of these events. When the parameter indicates a deviation in the motion execution capability, the robot dog's action mode is switched so that the robot dog performs exploratory actions before performing tasks that require precise motion. During the execution of the exploratory action, the difference information between the robot dog's movement command and the actual movement response is obtained, and the minute movement deviation is quantified based on the difference information; Based on the quantified motion deviation, the robot dog's subsequent motion commands are corrected to ensure the accuracy of the robot dog when performing precise motion tasks; Based on the execution effect of the revised motion commands, adjust the parameters characterizing the robot dog's motion execution capability. The steps for identifying events that occur during the robot dog's movement that cause minute motion deviations in its internal precision components include: Acquire instantaneous operating status information of the joint drive components; Obtain information on the attitude changes of the robot dog's body; Acquire instantaneous motion response information of the joint; Based on the instantaneous operating status information, the attitude change information, and the instantaneous motion response information, determine whether the instantaneous operating status information indicating the presence of a current peak, the attitude change information indicating rapid instability and recovery of the fuselage attitude, and the instantaneous motion response information indicating an instantaneous deviation between the motion command and the actual motion simultaneously meet preset conditions. When the judgment result is that the preset condition is met, it is identified as the event that causes a slight movement deviation in the internal precision components; The step of obtaining the difference information between the robot dog's movement command and the actual movement response during the execution of the exploratory action includes: During the execution of the exploratory action, the expected movement position information corresponding to the movement command of the robot dog is acquired in real time; At the same time, the actual movement position information corresponding to the actual movement response of the robot dog is acquired in real time; Based on the expected motion position information and the actual motion position information, the instantaneous deviation between the two is calculated to obtain the difference information; The steps for correcting subsequent motion commands of the robot dog based on the quantified motion deviation include: changing the response intensity of the controlled joint movement, adjusting the degree of influence of error accumulation, and correcting the coefficient of predictive correction according to the quantified motion deviation, so as to adjust the relevant parameters in the motion control strategy of the robot dog.

2. The autonomous obstacle avoidance method for a robot dog for urban security according to claim 1, characterized in that, The step of acquiring the difference information between the robot dog's movement command and the actual movement response during the execution of the exploratory action, and quantifying the minute movement deviation based on the difference information, includes: During the execution of the exploratory actions, information on the difference between the robot dog's movement commands and actual movement responses is acquired; simultaneously, data on the sensor's own operating status in the current environment is acquired. The discrepancy information is calibrated based on the self-operating status data; Then, based on the calibrated difference information, the minute motion deviation is quantified.

3. The autonomous obstacle avoidance method for a robotic dog for urban security as described in claim 1, characterized in that, After executing the step of correcting the subsequent movement commands of the robot dog, the method further includes: Continuously acquire the location information and motion status of moving targets in the current environment; Based on the location information and motion state, detect whether the position or motion state of the moving target has undergone instantaneous changes; When the instantaneous change is detected, the motion parameters in the corrected motion command are adjusted in real time according to the instantaneous change, and the adjustment process takes into account the small motion deviations that were previously quantized.

4. The autonomous obstacle avoidance method for a robot dog for urban security according to claim 1, characterized in that, The step of adjusting the parameters characterizing the robot dog's motion execution capability based on the execution effect of the modified motion command includes: Obtain motion performance information of the robot dog when executing the modified motion command; Simultaneously, the system acquires the lighting information and ground material information of the current environment in which the robot dog is located, as well as the working status information of the sensors used to acquire the motion performance information. Based on the illumination information, the ground material information, and the sensor's own operating status information, the motion performance information is processed to obtain the execution effect of the corrected motion command; Then, based on the execution results, adjust the parameters that characterize the robot dog's motion execution capability.

5. The autonomous obstacle avoidance method for a robot dog for urban security according to claim 1, characterized in that, Before performing tasks requiring precise movement, a robot dog will take the following steps in its trial-and-error actions: Control specific leg joints of the robot dog to make it swing back and forth within a preset small angle range in order to perform the exploratory action; During the reciprocating swing, the difference between the motion command of the specific leg joint and the actual swing angle is recorded.

6. A robotic dog autonomous obstacle avoidance system for urban security, used to execute the robotic dog autonomous obstacle avoidance method for urban security as described in any one of claims 1 to 5, characterized in that, The system includes: The event recognition module is used to identify events that occur during the robot dog's movement that cause minor motion deviations in its internal precision components, and to adjust the parameters characterizing the robot dog's motion execution capability based on the event recognition results. The action switching module is used to switch the action mode of the robot dog when the parameters indicate that there is a deviation in the motion execution capability, so that the robot dog performs exploratory actions before performing tasks that require precise motion. The deviation quantization module is used to acquire the difference information between the robot dog's motion command and the actual motion response during the execution of the trial action, and to quantify the small motion deviation based on the difference information. The instruction correction module is used to correct the robot dog's subsequent motion instructions based on the quantified motion deviation, so as to ensure the accuracy of the robot dog when performing precise motion tasks. The parameter adjustment module is used to adjust the parameters that characterize the robot dog's motion execution capability state based on the execution effect of the corrected motion commands.

Images