Robotic autonomous navigation method, device, and medium

Abstract

The application relates to the technical field of robot navigation, and discloses a robot autonomous navigation method, equipment and medium. The method comprises the following steps: when a robot travels along a preset navigation path, acquiring a space obstacle feature of a position where the robot is located, a water flow speed feature and a travel deviation feature of the robot; based on the preset navigation path and the space obstacle feature, performing obstacle avoidance control on the robot, so that the robot passes the current position in a corresponding preset obstacle avoidance strategy; based on the water flow speed feature and the travel deviation feature, performing deviation correction control on the robot, so that the travel deviation feature of the robot is constrained by a preset deviation correction constraint condition. The embodiment of the application can reduce the risk of collision of the robot during autonomous navigation, so as to adapt to the navigation scene of a tailings pond drainage tunnel.

Description

Technical Field

[0001] This application relates to the field of robot navigation technology, and in particular to a robot autonomous navigation method, device and medium. Background Technology

[0002] Tailings dam drainage tunnels are core facilities ensuring the safe operation of tailings dams, and their structural integrity directly affects the ecology and personnel safety of downstream areas. Currently, the inspection of tailings dam drainage tunnels mainly relies on manual labor, fixed sensors, or traditional hydraulic tunnel inspection equipment. However, due to the unique obstacles within tailings dam drainage tunnels, such as high-impact water flow and fixed drainage pipes, existing robotic navigation systems are easily affected by water flow impacts, causing them to veer off course, and they struggle to accurately avoid the serpentine layout of the pipes, making collisions a frequent occurrence. Summary of the Invention

[0003] The purpose of this application is to provide a robot autonomous navigation method, device and medium, which aims to reduce the risk of collisions during robot autonomous navigation, so as to adapt to the navigation scenario of tailings dam drainage tunnel.

[0004] This application provides a robot autonomous navigation method, including: While the robot is traveling along the preset navigation path, the spatial obstacle characteristics, water flow velocity characteristics, and travel deviation characteristics of the robot's location are acquired. Based on the preset navigation path and the spatial obstacle features, the robot is controlled to bypass obstacles, so that the robot passes through the current position with a corresponding preset obstacle bypass strategy. Based on the water flow velocity characteristics and the travel deviation characteristics, the robot is subjected to correction control so that the robot's travel deviation characteristics are constrained by preset correction constraints.

[0005] In some embodiments, the obstacle avoidance control of the robot based on the preset navigation path and the spatial obstacle features includes: Based on the preset navigation path, the first access risk characteristic of the robot's location is determined; Based on the spatial obstacle characteristics, a second passage risk characteristic of the robot's location is determined; When the first passage risk feature meets the first obstacle avoidance condition, the robot is controlled to execute the first preset obstacle avoidance strategy; when the second passage risk feature meets the second obstacle avoidance condition, the robot is controlled to execute the second preset obstacle avoidance strategy.

[0006] In some embodiments, the first obstacle avoidance condition includes at least one of the following: The turning angle of the robot's position exceeds the turning angle threshold; The radius of curvature of the path at the robot's location exceeds the radius of curvature threshold. The first preset obstacle avoidance strategy is to control the robot to retract its robotic arm and remain in the middle area of ​​its current position, traveling along the preset navigation path.

[0007] In some embodiments, the second obstacle avoidance condition includes at least one of the following: The width of the robot's location exceeds the first width threshold but does not exceed the second width threshold; The width of the robot's location does not exceed the first width threshold; The second preset obstacle avoidance strategy is as follows: when the width exceeds the first width threshold but does not exceed the second width threshold, the robot performs obstacle avoidance path planning based on the first safe width threshold and the width value of the robot's position; when the width does not exceed the first width threshold, the robot performs obstacle avoidance path planning based on the second safe width threshold and the width value of the robot's position, and controls the robot to move along the planned target obstacle avoidance path; the first safe width threshold is determined based on the width value of the robot when its robotic arm is extended, and the second safe width threshold is determined based on the width value of the robot when its robotic arm is retracted.

[0008] In some embodiments, the obstacle avoidance path planning for the robot includes: Based on a target safe width threshold, a passable area is determined in the robot's location; the target safe width threshold is either the first safe width threshold or the second safe width threshold. The A* path search method is used to search for the path with the minimum travel cost in each of the passable areas to obtain the initial obstacle bypass path; The initial obstacle avoidance path is smoothed to obtain the target obstacle avoidance path.

[0009] In some embodiments, the travel deviation characteristics include lateral sideslip deviation and yaw rate of change, and the deviation correction control of the robot based on the water flow velocity characteristics and the travel deviation characteristics includes: When the lateral sideslip offset exceeds the lateral sideslip offset threshold and / or the yaw angle change rate exceeds the yaw angle change threshold, the lateral wheel speed difference of the robot is determined based on the water flow velocity characteristics and the travel offset characteristics. The robot is controlled to travel at the lateral wheel speed difference.

[0010] In some embodiments, determining the lateral wheel speed difference of the robot based on the water flow velocity characteristics and the travel offset characteristics includes: Based on the lateral sideslip offset and the yaw angle change rate, a comprehensive deviation coefficient is determined; Based on the comprehensive deviation coefficient and the water flow velocity characteristics, the corresponding proportional coefficient, integral coefficient and differential coefficient are determined; Based on the proportional coefficient, the integral coefficient, and the differential coefficient, the lateral sideslip offset is subjected to proportional-integral-differential calculations to obtain the lateral wheel speed difference.

[0011] In some embodiments, the robot autonomous navigation method further includes: A bidirectional path matching method is used to reverse-parse the preset navigation path to obtain the corresponding return navigation path, so that the robot returns to its home position according to the return navigation path. When the robot travels a first preset mileage along the preset navigation path and / or a second preset mileage along the return navigation path, the robot is controlled to reduce its speed to a preset speed threshold.

[0012] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described robot autonomous navigation method.

[0013] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described robot autonomous navigation method.

[0014] The beneficial effects of this application are as follows: During the robot's journey along the preset navigation path, the characteristics of spatial obstacles, water flow velocity, and robot deviation in the travel space are acquired. Based on these characteristics, the robot can be controlled to avoid obstacles and correct its course. This allows the robot to be constrained by preset conditions, effectively cope with high-speed water flow and complex spatial obstacles, improve the robot's navigation accuracy and operational safety in complex environments such as tailings dam drainage tunnels, and reduce the risk of collisions during autonomous navigation, thus adapting to the navigation scenarios in tailings dam drainage tunnels. Attached Figure Description

[0015] Figure 1 This is a diagram illustrating the application environment of the robot autonomous navigation method provided in the embodiments of this application.

[0016] Figure 2 This is a flowchart of the robot autonomous navigation method provided in the embodiments of this application.

[0017] Figure 3 This is a flowchart of a method for obstacle avoidance control of a robot provided in an embodiment of this application.

[0018] Figure 4 This is a flowchart of a method for corrective control of a robot provided in an embodiment of this application.

[0019] Figure 5 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0021] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and drawings are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0023] The robot autonomous navigation method provided in this application can be executed by a computer device, which can be a terminal device or a server. The terminal device includes, but is not limited to, mobile phones, computers, smart home appliances, vehicle terminals, and aircraft. The server can be a standalone physical server, a server cluster consisting of multiple physical servers, a distributed system, or a cloud server. Furthermore, all information, data, and signals involved in this application's embodiments are authorized by the relevant parties or have been fully authorized by all parties, and the collection, use, and processing of related data comply with the relevant laws, regulations, and standards of the relevant countries and regions.

[0024] In traditional tailings dam drainage tunnel inspections, robot navigation is disrupted by high-impact water flow, causing deviations in its trajectory. Simultaneously, the identification and avoidance of spatial obstacles are hindered by the serpentine layout of fixed drainage pipes. Key performance indicators such as navigation accuracy and obstacle avoidance reliability are significantly reduced, further leading to insufficient robot path stability and increased collision risks. For example, in the actual operating environment of tailings dam drainage tunnels, the water flow velocity is monitored at a high level, and spatial obstacles manifest as complex geometric shapes of serpentine pipes. Lateral slippage during robot movement is triggered by the impact of water flow, and path adjustment is restricted by the position of obstacles. This results in frequent physical contact events between the mechanical structure and the pipes, interrupting the continuity of the inspection task and increasing the risk of equipment damage.

[0025] If the above problems are not resolved, the robot's autonomous navigation function may be disabled, the risk of equipment damage will be increased, the safety of the tunnel structure will be threatened, and the ecological stability and personnel safety of the downstream area will be weakened, thus highlighting the necessity of improving navigation and control methods.

[0026] Based on this, embodiments of this application provide a robot autonomous navigation method, device, and medium. Based on the spatial obstacle characteristics, water flow velocity characteristics, and robot travel deviation characteristics in the travel space, obstacle avoidance control and deviation correction control are performed on the robot, so that the robot is constrained by preset constraints, which can reduce the risk of collision during robot autonomous navigation and adapt to the navigation scenario of tailings dam drainage tunnel.

[0027] Figure 1 This diagram illustrates the application environment of the robot autonomous navigation method provided in the embodiments of this application. (See also...) Figure 1 This method is applied to an autonomous navigation system for robots. The system includes a terminal 110 and a server 120. The terminal 110 and server 120 are connected via a network. The terminal 110 can be a desktop terminal or a mobile terminal; the mobile terminal can be at least one of a mobile phone, tablet, or laptop. The server 120 can be a standalone server or a server cluster consisting of several servers. The terminal 110 sends the spatial obstacle characteristics, water flow velocity characteristics, and robot deviation characteristics of the robot's current location to the server 120 when the robot is traveling along a preset navigation path. The server 120, while the robot is traveling along the preset navigation path, acquires the spatial obstacle characteristics, water flow velocity characteristics, and robot deviation characteristics of the robot's current location. Based on the preset navigation path and spatial obstacle characteristics, it performs obstacle avoidance control on the robot, enabling the robot to pass through its current location using a corresponding preset obstacle avoidance strategy. Based on the water flow velocity characteristics and deviation characteristics, it performs deviation correction control on the robot, ensuring that the robot's deviation characteristics are constrained by preset deviation correction constraints.

[0028] It should be understood that Figure 1 The application scenarios shown are merely examples. In practical applications, the robot autonomous navigation method provided in this application embodiment can also be applied to other scenarios. For example, the above-described robot autonomous navigation method can be directly applied to terminal 110. Terminal 110 is used to acquire the spatial obstacle features, water flow velocity features, and robot travel deviation features of the robot's current position when the robot travels along a preset navigation path. Based on the preset navigation path and spatial obstacle features, it performs obstacle avoidance control on the robot, enabling the robot to pass through its current position using a corresponding preset obstacle avoidance strategy. Based on the water flow velocity features and travel deviation features, it performs deviation correction control on the robot, making the robot's travel deviation features subject to preset deviation correction constraints.

[0029] See Figure 2 In one embodiment, a robot autonomous navigation method is provided. The execution subject of the method may be a robot or a server, including but not limited to steps S201 to S203.

[0030] Step S201: While the robot is traveling along the preset navigation path, acquire the spatial obstacle features, water flow velocity features, and robot travel deviation features of the robot's location.

[0031] A preset navigation path refers to a route that the robot must follow before it begins to move, planned in advance based on task requirements and environmental information. This path can consist of a series of discrete waypoints or continuous curve segments. The preset navigation path can be generated before robot deployment through manual teaching, map editing software, or automatically based on environmental scan data. For example, the operator can directly draw a route the robot needs to follow on a map interface, or define the path by inputting a series of coordinate points. The robot's internal motion controller receives this path information and drives the robot's actuators (such as wheels, propellers, etc.) to move along the path.

[0032] In one implementation, the robot can employ path tracking algorithms, such as pure tracking algorithms or proportional-integral-derivative (PID) control, to make its center point fit as closely as possible to the preset navigation path.

[0033] Spatial obstacle features refer to the geometric and topological properties of physical entities or areas in the environment surrounding the robot's current location that may hinder the robot's normal movement. These features can include the size, shape, location, density of obstacles, and the width of passable areas.

[0034] Water flow velocity characteristics refer to the velocity and direction of water flow in the robot's environment. In underwater or surface environments, water flow velocity characteristics have a significant impact on the robot's stable movement and may cause the robot to deviate from its intended path.

[0035] Travel deviation refers to the deviation of a robot's position or orientation from a preset navigation path or desired state during actual travel. This feature can quantify the degree to which the robot deviates from the predetermined trajectory, such as lateral displacement or heading angle deviation.

[0036] During the robot's movement, it is necessary to acquire real-time information on spatial obstacle characteristics, water flow velocity characteristics, and the robot's deviation characteristics. Acquiring spatial obstacle characteristics can be achieved using various sensors. For example, the robot can be equipped with ultrasonic sensors or lidar to detect obstacles in the surrounding environment by emitting and receiving signals, and to calculate the distance and position of obstacles based on the signal's time of flight or reflection intensity. Acquiring water flow velocity characteristics can rely on water flow sensors. For instance, the robot can be equipped with a Doppler-effect velocimeter to determine the speed and direction of local water flow by measuring the velocity of particles moving in the water. Acquiring deviation characteristics can be achieved using an inertial measurement unit (IMU) combined with visual odometry or a global positioning system (GPS). The IMU provides the robot's attitude and angular velocity information, while visual odometry or GPS provides the robot's position information. By comparing this real-time data with a preset navigation path, the robot's lateral, longitudinal, or angular deviations can be calculated.

[0037] Step S202: Based on the preset navigation path and spatial obstacle characteristics, the robot is controlled to bypass obstacles, so that the robot passes through the current position with the corresponding preset obstacle bypass strategy.

[0038] Obstacle avoidance control refers to the process by which a robot adjusts its path and posture according to a preset strategy to safely avoid obstacles when it detects them during its movement. This control aims to ensure that the robot can navigate smoothly in complex environments.

[0039] Pre-set obstacle avoidance strategies refer to a series of pre-defined responses based on different obstacle characteristics and environmental conditions during obstacle avoidance control. These strategies may include deceleration, steering, changing attitude, or selecting alternative paths.

[0040] During the robot's movement, when an obstacle is detected in front of the robot or the passage space is insufficient, an obstacle avoidance mechanism needs to be activated. In one implementation, the robot can simply choose to turn left or right at a fixed angle based on the distance and size of the obstacle to avoid it. For example, if the obstacle is directly in front of the robot, it can be controlled to turn 30° to the left, travel a certain distance, and then turn back 30° to the right to bypass the obstacle. In another implementation, the robot can choose the shortest path to bypass the obstacle within the passable area based on its shape and position. For example, if the obstacle is a long, narrow pipe, the robot can be controlled to move parallel to one side of the pipe until it passes through the obstacle area. These obstacle avoidance strategies are predefined and are triggered upon detection of an obstacle.

[0041] Step S203: Based on the water flow velocity characteristics and travel deviation characteristics, perform deviation correction control on the robot so that the robot's travel deviation characteristics are constrained by preset deviation correction constraints.

[0042] Corrective control refers to the process by which a system automatically adjusts the robot's motion state to return it to a preset navigation path or desired state when the robot's deviation characteristics exceed the allowable range. This control aims to maintain the robot's stability and path accuracy.

[0043] Preset correction constraints refer to the allowable range or adjustment limits set for the robot's travel deviation characteristics during correction control. When the robot's deviation meets or exceeds these conditions, the corresponding correction action will be triggered.

[0044] During its movement, the robot is susceptible to being impacted by water flow and deviating from its preset path. To address this, a correction control system is introduced. In one implementation, when the robot's lateral deviation exceeds a preset threshold, the system simply increases or decreases the rotational speed of the robot's two thrusters based on the current water flow velocity and direction. This generates a counter-thrust to counteract the water flow's influence, returning the robot to its preset path. For example, if the water flow pushes the robot to the left, causing its lateral deviation to exceed the threshold, the robot can be controlled to increase the rotational speed of the right thruster while decreasing the rotational speed of the left thruster, thereby generating a rightward corrective force. The preset correction constraint can be a maximum allowable deviation distance or a maximum allowable yaw angle; once the actual deviation exceeds these limits, the correction control is activated.

[0045] The following example will provide a more detailed explanation of the above technical solution: Imagine a robot performing an inspection task inside an underwater drainage tunnel. The tunnel contains irregularly shaped fixed drainage pipes and a continuous flow of water.

[0046] First, before the mission begins, the operators have planned a preset navigation path for the robot, which is stored in the robot's navigation system. Once the robot is deployed into the tunnel, it is controlled to travel along this preset navigation path.

[0047] During its movement, the robot's onboard sensor system continuously operates, acquiring environmental information in real time. Specifically, lidar or sonar sensors are used to scan the tunnel's internal structure to obtain spatial obstacle features at the robot's location, such as the shape and position of pipes and the available width of the tunnel. Simultaneously, a water flow velocity meter is used to measure the water flow velocity characteristics within the tunnel, including the direction and magnitude of the flow. Furthermore, an inertial measurement unit and visual odometry work together to continuously monitor the robot's actual position and attitude, comparing it with a pre-set navigation path to calculate the robot's deviation characteristics, such as lateral sideslip and yaw angle.

[0048] When the robot enters a narrow section of the tunnel with serpentine pipes, its sensor system detects an obstacle ahead. At this point, based on the pre-defined navigation path and the characteristics of the obstacle, the obstacle avoidance control module is activated. The robot assesses the passage risk at the current location based on the pipe geometry and the available width of the tunnel. For example, if the pipe occupies most of the passage space and the path requires a large turn, the robot will select a pre-defined obstacle avoidance strategy. This strategy might instruct the robot to adjust its posture, such as retracting its robotic arm to reduce its width and carefully moving laterally along the edge of the pipe to avoid collisions, thus safely navigating the narrow area.

[0049] Meanwhile, due to the influence of water flow within the tunnel, the robot may deviate from the preset path. When the travel deviation characteristic (e.g., lateral slip) exceeds the preset correction constraint (e.g., the maximum allowable lateral deviation distance), the correction control module is triggered based on the water flow velocity characteristic and the travel deviation characteristic. The actuator calculates the required corrective force based on the current water flow velocity and direction, as well as the robot's actual deviation. For example, if the water flow pushes the robot to the left, causing it to deviate from the preset path to the right, the correction control module instructs the robot's right-side thrusters to increase thrust and the left-side thrusters to decrease thrust, thereby generating a leftward lateral thrust to counteract the water flow's influence. This gradually returns the robot to the preset navigation path and keeps its travel deviation characteristic within the preset correction constraint.

[0050] Through the aforementioned collaborative work, the robot can effectively avoid irregular obstacles and resist the impact of water flow in complex underwater tunnel environments, maintaining stable path tracking, thereby achieving autonomous and safe inspection tasks.

[0051] Based on the above examples, the robot autonomous navigation method provided in this embodiment demonstrates significant technical contributions. In existing technologies, inspection robots for tailings dam drainage tunnels often face problems such as yaw caused by water flow impact and difficulty in accurately avoiding complex obstacles. This embodiment effectively solves the problem of robots struggling to accurately avoid complex obstacles such as serpentine pipes by introducing real-time acquisition of spatial obstacle characteristics and obstacle avoidance control based on preset navigation paths. Unlike the simple obstacle avoidance logic (such as stopping or fixing the direction upon encountering an obstacle) that may be used in existing technologies, this method can dynamically select and execute corresponding preset obstacle avoidance strategies based on the specific characteristics of the obstacle and the preset navigation path, thereby achieving more refined and efficient obstacle avoidance. For example, in the above example, the robot can adjust its posture and move along the edge according to the shape of the pipe and the width of the tunnel, rather than simply stopping or colliding, which significantly improves the robot's mobility and safety in complex environments. Furthermore, this embodiment effectively overcomes the challenge of existing robots being susceptible to yaw caused by water flow impact by acquiring water flow velocity characteristics and travel deviation characteristics in real time and performing correction control based on these characteristics. Traditional robots may rely solely on internal inertial navigation or simple path tracking, making it difficult to maintain path accuracy in strong water flow environments. This method, however, proactively senses the impact of water flow and makes targeted correction adjustments based on actual deviations, ensuring that the robot's movement deviation characteristics are always constrained by preset correction constraints. This guarantees the robot's path stability in water flow environments, avoiding path deviations or task failures caused by water flow impacts, thereby improving the reliability and success rate of inspection tasks.

[0052] In summary, the robot autonomous navigation method of this embodiment organically combines obstacle avoidance control and deviation correction control, and makes decisions based on multi-source environmental characteristics, forming a comprehensive and robust navigation system. This system can not only cope with complex and ever-changing physical obstacles, but also effectively resist external environmental interference, significantly improving the robot's autonomous navigation capability and operational efficiency in special working environments.

[0053] In some of the embodiments described above in this application, obstacle avoidance control of the robot is proposed based on a preset navigation path and spatial obstacle features. However, in actual travel, the robot may face the dual challenges of the complexity of the preset path itself (such as sharp turns) and real-time obstacles. If a single obstacle avoidance strategy is adopted without differentiation, it may lead to low obstacle avoidance efficiency or insufficient safety. In this regard, this application further proposes a method for obstacle avoidance control of the robot.

[0054] See Figure 3 In one embodiment, the method for controlling the robot to avoid obstacles includes, but is not limited to, steps S301 to S303.

[0055] Step S301: Based on the preset navigation path, determine the first passage risk characteristics of the robot's location.

[0056] Step S302: Based on the spatial obstacle characteristics, determine the second passage risk characteristics of the robot's location.

[0057] Step S303: When the first passage risk feature meets the first obstacle avoidance condition, control the robot to execute the first preset obstacle avoidance strategy; when the second passage risk feature meets the second obstacle avoidance condition, control the robot to execute the second preset obstacle avoidance strategy.

[0058] The first traversal risk characteristic indicates the difficulty or potential risk a robot may face while traversing a pre-defined navigation path due to the path's geometric characteristics (such as curvature and width variations). This characteristic can be determined by analyzing geometric parameters such as the local curvature, slope, and rate of width change of the pre-defined navigation path. For example, areas with greater curvature, steeper slopes, and abruptly narrowing path widths have higher first traversal risk characteristic values. Alternatively, it can be determined by pre-assessing and labeling the pre-defined navigation path with risks. For instance, during the path planning phase, high-risk areas in the path can be identified manually or by algorithms, and assigned corresponding risk levels or characteristic values.

[0059] The second passage risk characteristic indicates the degree of impact or potential hazard posed by real-time spatial obstacles (such as underwater rocks, shipwrecks, other underwater robots, etc.) around the robot's location to its passage. This characteristic can be obtained by real-time detection of the surrounding environment using sensors onboard the robot (such as sonar, vision sensors, lidar, etc.), acquiring information such as the distance, size, shape, and speed of movement of obstacles, and calculating the degree of threat posed by the obstacles to the robot's passage. For example, the closer the obstacle is to the robot, the larger its size, and the faster its movement speed, the higher its second passage risk characteristic value. It can also be determined by combining historical data or pre-loaded map information; for example, knowing in advance that certain areas have fixed obstacles, and updating and assessing the risk based on real-time sensor data.

[0060] The first obstacle avoidance condition is the condition that triggers the robot to execute a first preset obstacle avoidance strategy, which is usually related to the risk of the preset navigation path itself. This condition may be that a certain indicator of the first passage risk characteristic (such as path curvature or turning angle) exceeds a preset threshold. For example, when the turning angle of the path exceeds a certain safety threshold, it is considered to meet the first obstacle avoidance condition. Alternatively, it may be based on path type or area division. For example, when the robot enters a preset "narrow passage" or "complex terrain" area, the first obstacle avoidance condition is triggered. The second obstacle avoidance condition is the condition that triggers the robot to execute a second preset obstacle avoidance strategy, which is usually related to the risk of obstacles in real-time space. This condition may be that a certain indicator of the second passage risk characteristic (such as obstacle distance or obstacle density) exceeds a preset threshold. For example, when an obstacle is detected within a certain range in front of the robot, and the passage space occupied by the obstacle is less than a certain safe width, it is considered to meet the second obstacle avoidance condition. Alternatively, it may be based on the type or threat level of the obstacle. For example, when a moving obstacle or a high-threat-level obstacle is detected, the second obstacle avoidance condition is triggered.

[0061] The first preset obstacle avoidance strategy addresses the risks inherent in the preset navigation path itself, with pre-defined countermeasures taken by the robot. This strategy can involve adjusting the robot's posture, speed, or robotic arm status, such as reducing travel speed, tightening the robotic arm, or adjusting the center of gravity to improve stability when navigating complex paths. Alternatively, it can employ specific path tracking algorithms, for example, using a smoother trajectory tracking algorithm at sharp bends to avoid over-turning or deviation from the path. The second preset obstacle avoidance strategy addresses the risks posed by real-time spatial obstacles, with pre-defined countermeasures taken by the robot. This strategy can involve dynamically planning a new detour path to avoid obstacles, for example, calculating a safe and efficient detour trajectory in real time based on the obstacle's position, size, and movement trend. Alternatively, it can involve taking avoidance actions such as emergency braking, hovering, or reversing; for example, when an obstacle suddenly appears and is too close, the robot can immediately stop or perform a small-scale avoidance maneuver.

[0062] The solution in this application refines obstacle avoidance control during the robot's autonomous navigation process. First, the robot analyzes the geometric characteristics of its preset navigation path, such as curvature and width variations, to determine the first obstacle avoidance characteristic at its current location. Simultaneously, the robot uses its sensors to perceive the surrounding environment in real time, acquiring information such as the distribution, size, and distance of spatial obstacles, thereby determining the second obstacle avoidance characteristic at its current location. By distinguishing between these two risk characteristics, the robot can more comprehensively assess the challenges it faces. When the assessment results show that the first obstacle avoidance characteristic matches the preset first obstacle avoidance condition, it indicates that the preset path itself has a high level of difficulty, such as being too narrow or curved. In this case, the robot will execute the first preset obstacle avoidance strategy, which typically aims to optimize the robot's stability or safety when navigating complex paths, such as adjusting its posture or retracting its robotic arm to reduce its size. On the other hand, when the second obstacle avoidance characteristic matches the preset second obstacle avoidance condition, it indicates that there are real-time spatial obstacles around the robot that need to be avoided. In this case, the robot will execute the second preset obstacle avoidance strategy, which typically involves dynamically planning a new detour path or taking emergency obstacle avoidance actions to ensure that the robot can safely avoid obstacles. In this way, the proposed solution decomposes obstacle avoidance control into independent assessment and response to inherent path risks and real-time obstacle risks, enabling the robot to flexibly select and execute the most suitable obstacle avoidance strategy based on different types of risk sources. This layered and refined risk assessment and strategy selection mechanism avoids the limitations that may arise from a single obstacle avoidance strategy, improving the adaptability and safety of the robot's autonomous navigation.

[0063] As a specific implementation method, when performing obstacle avoidance control, the robot's autonomous navigation method can first calculate the radius of curvature of each point on the preset navigation path by analyzing the digital map data. When the radius of curvature of a point is less than a preset "sharp bend radius threshold," the first passage risk characteristic at that point is considered to meet the first obstacle avoidance condition. At this time, the robot can be controlled to reduce its travel speed and retract its robotic arm to a minimum occupied space state to reduce the inertial impact and collision risk when passing through sharp bends. Simultaneously, the robot can use its front-mounted sonar sensor to scan the area ahead in real time to obtain the distance and orientation information of obstacles. When the sonar data shows that there is a narrow passage ahead with a width smaller than the width of the robot's extended robotic arm but larger than the width of the robot's retracted robotic arm, and the length of the passage exceeds a preset "safe passage length," the second passage risk characteristic of that area is considered to meet the second obstacle avoidance condition. At this time, the robot can be controlled to activate the local path planning module, and based on the current environmental information and its own size, plan a precise trajectory to pass through the narrow passage and proceed cautiously along the trajectory.

[0064] Through the above technical solutions, the robot can distinguish between the complexity of the preset navigation path itself and the risks posed by real-time spatial obstacles, and adopt different obstacle avoidance strategies accordingly. This refined risk assessment and strategy selection mechanism enables the robot to avoid obstacles more flexibly and efficiently, effectively avoiding the reduced passage efficiency or safety hazards caused by using a single strategy to deal with all situations, and significantly improving the adaptability and reliability of the robot's autonomous navigation.

[0065] In some embodiments, the first obstacle avoidance condition includes at least one of the following: the turning angle value of the robot's location exceeds a turning angle threshold; the path curvature radius value of the robot's location exceeds a curvature radius threshold. The first preset obstacle avoidance strategy is to control the robot to retract its robotic arm and remain in the middle area of ​​its location, traveling along a preset navigation path.

[0066] The condition that the robot's current turning angle exceeds a steering angle threshold is designed to identify sharp turns within a pre-defined navigation path. The robot can calculate the local turning angle of the pre-defined navigation path ahead of its current position in real time using its internal path planning module or navigation system. For example, the turning angle can be obtained by analyzing the angle formed by three consecutive points on the path, or by fitting the curve equation of a local path segment. This condition is considered met when the calculated turning angle is greater than a pre-set steering angle threshold. The steering angle threshold can be set based on the robot's kinematic characteristics, maximum turning capability, and safety margin. Furthermore, the actuator can pre-analyze the entire pre-defined navigation path, identifying and marking all path segments where the turning angle exceeds the steering angle threshold. This condition is triggered when the robot enters these pre-marked path segments during its movement. The condition that the robot's current path curvature radius exceeds a curvature radius threshold is used to identify areas with a large curvature radius (i.e., relatively straight or gently curved) within the pre-defined navigation path. These areas may require specific travel strategies to ensure stability and efficiency. The actuator can calculate the local curvature radius of the current travel path in real time. For example, curve fitting (such as B-splines or cubic splines) can be performed on the path points, and then the radius of curvature of the fitted curve at the robot's current position can be calculated. This condition is met when the calculated radius of curvature value is greater than a preset radius of curvature threshold. The radius of curvature threshold can be set according to the robot's size, stability requirements, and expected characteristics of the path. Alternatively, the preset navigation path can be preprocessed during planning to identify all path segments with radius of curvature values ​​exceeding a specific threshold and mark them as areas where a first preset obstacle avoidance strategy needs to be implemented.

[0067] In the embodiments of this application, to effectively address the inherent complexity of the preset navigation path, such as sharp turns or relatively straight areas requiring careful passage, this solution specifically defines the first passage risk characteristic. When the turning angle value of the preset navigation path at the robot's location exceeds a preset turning angle threshold, it indicates that there is a sharp turn in the current path segment. If the robot is in an extended state or deviating from the centerline, it is highly likely to collide or deviate from the path. Simultaneously, when the path curvature radius value exceeds a preset curvature radius threshold, it may indicate that the path is relatively straight or has a small curvature. However, to ensure stability and safety in these areas, specific travel strategies are still required. Therefore, once any of the above conditions is detected, the first passage risk characteristic is considered to meet the first obstacle avoidance condition, and the executing entity will immediately trigger the first preset obstacle avoidance strategy. The core of this strategy is to control the robot to retract its robotic arm to reduce its overall lateral size and decrease the risk of interference with environmental obstacles. Simultaneously, the robot is instructed to remain in the central area of ​​its location and travel along the preset navigation path. This is achieved through precise path tracking control, ensuring that the robot can stably and centrally pass through these potentially risky path segments. In this way, the solution can proactively identify and address the inherent geometric challenges of the preset navigation path, significantly improving the navigation reliability and safety of the robot in complex aquatic environments.

[0068] The following is a concrete example. Suppose an underwater inspection robot is traveling through an underwater pipe according to a preset navigation path. When the robot is about to enter a pipe section that requires a 90° turn, its navigation system, through geometric analysis of the preset navigation path, calculates that the turning angle at this point is 95°, which exceeds the system's preset turning angle threshold (e.g., 85°). At this moment, the robot detects a first obstacle avoidance feature that meets the first obstacle avoidance condition. In response, the robot's control system immediately sends a command to its robotic arm, causing the previously extended arm to quickly retract to a folded position close to the robot body, thus reducing the robot's lateral dimensions. Simultaneously, the robot's motion controller activates a high-precision path tracking mode, using the distance information between the robot and the pipe wall acquired in real time by its onboard sonar sensors, combined with the centerline data of the preset navigation path, to accurately calculate the lateral deviation of the robot from the centerline. Based on these deviations, the controller adjusts the speed difference between the robot's two thrusters, enabling the robot to accurately maintain itself in the middle of the pipe and smoothly pass through the sharp turn at a stable speed.

[0069] Through the above technical solution, this application can clearly define the first obstacle avoidance condition as the turning angle or radius of curvature of the path, enabling the robot to identify and predict geometric risks in the preset navigation path in advance, such as sharp turns or straight sections that need to be passed stably. When these risks are identified, by implementing the first preset obstacle avoidance strategy of retracting the robotic arm and staying in the middle area of ​​the path, the robot can effectively reduce its lateral size in the underwater environment, reducing the risk of scratching or colliding with pipe walls, underwater structures, or other obstacles. At the same time, the strategy of staying in the middle area of ​​the path enhances the robot's path tracking accuracy and stability, especially in environments with water flow disturbances or low visibility, ensuring that the robot can pass through complex path sections smoothly and safely. This proactive risk identification and strategy adjustment mechanism significantly improves the reliability and safety of the robot's autonomous navigation, reduces unexpected situations caused by the geometric complexity of the path, and thus improves the success rate and efficiency of task execution.

[0070] In some embodiments, the second obstacle avoidance condition includes at least one of the following: the width value of the robot's location exceeds a first width threshold but does not exceed a second width threshold; the width value of the robot's location does not exceed the first width threshold. The second preset obstacle avoidance strategy is as follows: when the width exceeds the first width threshold but does not exceed the second width threshold, obstacle avoidance path planning is performed on the robot based on a first safe width threshold and the width value of the robot's location; when the width does not exceed the first width threshold, obstacle avoidance path planning is performed on the robot based on the second safe width threshold and the width value of the robot's location, and the robot is controlled to move along the planned target obstacle avoidance path. The first safe width threshold is determined based on the width value of the robot when its robotic arm is extended, and the second safe width threshold is determined based on the width value of the robot when its robotic arm is retracted.

[0071] The second obstacle avoidance condition aims to finely determine whether a specific obstacle avoidance strategy needs to be executed based on the actual width of the robot's location. For example, when the width of the robot's location exceeds a first width threshold but does not exceed a second width threshold, it indicates that the area is relatively narrow but still has some passage space; while when the width does not exceed the first width threshold, it indicates that the area is very narrow, and the robot may need to adopt a more compact posture to pass through. These width thresholds can be preset according to the robot's physical dimensions, the extended / retracted state of the robotic arm, and the required safety margin. The width of the robot's location can be obtained in real time by sensors on the robot (such as LiDAR, ultrasonic sensors, or vision sensors) to acquire information about the surrounding environment, and the minimum or average width of the currently passable area can be calculated by data processing algorithms.

[0072] The second preset obstacle avoidance strategy provides differentiated path planning schemes for areas with varying degrees of narrowness. When the robot's location is in a moderately narrow range (i.e., exceeding the first width threshold but not exceeding the second width threshold), obstacle avoidance path planning is performed based on the first safe width threshold. This first safe width threshold typically considers the maximum width of the robot arm when extended, and adds a certain safety margin to ensure the robot can safely pass through while maintaining the extended arm. When the robot's location is in an extremely narrow range (i.e., not exceeding the first width threshold), obstacle avoidance path planning is performed based on the second safe width threshold. This second safe width threshold typically considers the minimum width of the robot arm when retracted, and adds a safety margin to ensure the robot can pass through the narrowest area after retracting the arm. Path planning can be implemented using various algorithms, such as the A* algorithm, RRT algorithm, or D* Lite algorithm. These algorithms can search for a feasible, collision-free path in an obstacle environment, taking into account robot size and safety margins. Once the target obstacle avoidance path is planned, the robot control system instructs the robot to travel along that path.

[0073] This application's solution classifies the width of the robot's location and uses corresponding safe width thresholds for obstacle avoidance path planning based on different width levels. This allows the robot to dynamically adjust its obstacle avoidance strategy according to the actual narrowness of the environment. When the robot encounters a moderately narrow area, a relatively lenient first safe width threshold can be used for path planning, potentially allowing the robotic arm to remain extended, thus maintaining a certain operational capability while passing through obstacles. When the robot encounters an extremely narrow area, a stricter second safe width threshold is used for path planning. This typically means the robot needs to retract its robotic arm to reduce its width, thereby enabling it to pass through areas that would otherwise be impassable. This tiered strategy allows the robot to cope with various complex spatial obstacles more flexibly and efficiently, avoiding the difficulties or inefficiencies that may result from a single strategy. In this way, the robot can more intelligently choose the optimal passage method, improving its autonomous navigation capabilities and task execution efficiency in complex environments.

[0074] For example, suppose an underwater robot needs to inspect a body of water filled with weeds and rocks. The robot is equipped with a retractable robotic arm for grasping samples or clearing obstacles. As the robot moves, its onboard sonar sensors or vision system detects a narrow passage ahead. The system first measures the width of the passage. If the passage width is 1.3 meters, and the preset first width threshold is 1.2 meters, and the second width threshold is 1.5 meters, then the passage width falls within the range of "exceeding the first width threshold but not exceeding the second width threshold." In this case, the actuator will use the first safe width threshold (e.g., the width of the robot's robotic arm when extended plus a safety margin of 0.1 meters, totaling 1.1 meters) for path planning. The robot will plan a path to carefully pass through the passage while keeping its robotic arm extended. If the passage width is only 1.0 meter, then this width falls within the range of "not exceeding the first width threshold." In this case, the actuator will use the second safe width threshold (e.g., the width of the robot's robotic arm when retracted plus a safety margin of 0.1 meters, totaling 0.9 meters) for path planning. The robot plans a path and may automatically retract its robotic arm before crossing the passage to ensure a smooth passage. The planned path is then transmitted to the robot's motion controller, which instructs it to move precisely along that path.

[0075] Through the aforementioned technical solution, the robot can adaptively select appropriate obstacle avoidance strategies and safe width thresholds for path planning based on the width characteristics of its location. This enables the robot to more effectively traverse obstacle areas of varying narrowness, avoiding collisions or obstructions caused by inappropriate strategies, and significantly improving the robot's mobility and navigation efficiency in complex environments. Furthermore, by differentiating the safe width between the extended and retracted states of the robotic arm, the robot can maximize its operational capabilities while ensuring safety, thereby optimizing the flexibility and efficiency of task execution.

[0076] In some embodiments, obstacle avoidance path planning for the robot includes: determining a passable area in the robot's location based on a target safe width threshold; using the A* path search method to search for the path with the minimum travel cost from each passable area to obtain an initial obstacle avoidance path; and smoothing the initial obstacle avoidance path to obtain a target obstacle avoidance path. The target safe width threshold is either a first safe width threshold or a second safe width threshold.

[0077] In one specific embodiment, the formula for calculating the target security width threshold is: , in, The target safe width threshold, Based on the safe distance, To ensure a safe braking distance, To maintain a safe distance for obstacle avoidance, Correct the distance for the robotic arm's status.

[0078] Determining the passable area within a robot's location based on a target safe width threshold aims to identify areas in the environment that the robot can safely traverse, according to the robot's current state (e.g., whether the robotic arm is deployed). Identifying passable areas is fundamental to path planning, simplifying a complex environment into a series of discrete or continuous spaces where the robot can move. This step involves acquiring environmental point cloud data or depth maps using sensors (such as LiDAR and depth cameras), combining this data with the robot's dimensions and the target safe width threshold to rasterize the environment. In the raster map, all obstacles can be expanded beyond the target safe width threshold; grids not covered by these expanded obstacles are marked as passable areas. Alternatively, a topological or semantic map of the environment can be created to pre-label the passability attributes and width information of different areas. When the robot enters a certain area, the sub-areas within that area that the robot can safely traverse are dynamically calculated based on the area's width information and the target safe width threshold.

[0079] The A* path search method is used to search for the path with the minimum travel cost in each of the traversable areas, obtaining the initial obstacle avoidance path. This step utilizes the A* algorithm to find the shortest path or the path with the minimum travel cost from the starting point to the destination within the determined traversable areas. The A* algorithm is an efficient heuristic search algorithm that evaluates the priority of each node by combining the actual cost (cost from the starting point to the current node) and the heuristic cost (estimated cost from the current node to the destination). In a grid map, each traversable grid can be considered as a node, and the movement cost between adjacent grids can be set according to distance or terrain complexity. The algorithm will prioritize exploring the node with the minimum total cost until the target node is found. In addition to the A* algorithm, other path search algorithms such as Dijkstra's algorithm, RRT (Rapidly-exploring Random Tree) algorithm, or PRM (Probabilistic RoadMap) algorithm can also be used, each with its own advantages in different scenarios.

[0080] Path smoothing is performed on the initial obstacle avoidance path to obtain the target obstacle avoidance path. This step aims to eliminate any sharp corners or discontinuities in the initial path, making it more consistent with the robot's kinematic constraints and facilitating stable and efficient robot execution. A smoothed path reduces the frequency of acceleration, deceleration, and turning, improving travel efficiency and comfort. Path smoothing can be achieved by using mathematical methods such as spline interpolation (e.g., B-splines, cubic splines) or Bézier curves to fit discrete points on the initial obstacle avoidance path, generating a continuous and smoothly curvature-controlled curve. These methods can generate paths that meet the robot's maximum curvature and maximum acceleration limits. Alternatively, signal processing methods such as Gaussian filtering and moving averages can be used to process path points, or optimization algorithms (e.g., gradient descent) can be used to adjust path points, minimizing curvature changes and path length while maintaining obstacle avoidance capabilities.

[0081] The obstacle avoidance path planning method of this application first accurately identifies all areas in the environment that the robot can safely traverse, based on the target safe width threshold required for the robot's current state. This step lays the foundation for subsequent path search, ensuring that the planned path is safe and feasible. Subsequently, using the A* path search method, an initial obstacle avoidance path with the minimum travel cost from the current position to the target position is efficiently searched within these identified traversable areas. The A* algorithm, by comprehensively considering the actual travel cost and the estimated residual cost, can quickly converge to a near-optimal path. However, the path generated by the A* algorithm often consists of a series of discrete straight line segments and may contain sharp corners, making it unsuitable for direct robot execution. Therefore, after obtaining the initial obstacle avoidance path, further path smoothing processing is performed to eliminate discontinuities in the path, generating a target obstacle avoidance path with continuous curvature that conforms to the robot's kinematic characteristics. Through the above synergistic effect, the solution of this application, while inheriting obstacle avoidance planning based on a safe width threshold, further solves the problems of path planning efficiency and path quality. It not only ensures the safety of the planned path but also guarantees the efficiency of the path through the A* algorithm, and improves the executability of the path and the stability of the robot's movement through path smoothing. This phased and progressive planning strategy enables the robot to complete obstacle avoidance tasks in a more intelligent and stable manner in complex and ever-changing environments, avoiding frequent adjustments and energy consumption caused by path irregularities, thereby improving the overall autonomous navigation performance.

[0082] The following is a concrete example to illustrate this. When a robot needs to plan an obstacle avoidance path, the execution unit determines whether to use a first or second safe width threshold as the target safe width threshold based on whether the robot's robotic arm is currently deployed. For example, if the robot retracts its robotic arm, the second safe width threshold is used. Next, the robot uses its onboard LiDAR sensor to acquire real-time point cloud data of the surrounding environment and, combined with the target safe width threshold, inflates all obstacles on the internally constructed grid map, thus clearly identifying all grid areas that the robot can safely pass through. Subsequently, the navigation module calls the A* path search algorithm. This algorithm starts from the robot's current position and ends at the target point after obstacle avoidance, searching within the previously determined passable grid areas. During the search, the travel cost of each grid can be set as a weighted sum of the travel distance and terrain complexity, and the heuristic function can be Manhattan distance or Euclidean distance. The A* algorithm will prioritize exploring nodes with the minimum total cost (actual cost plus heuristic cost) until an initial obstacle avoidance path connecting the start and end points is found. Finally, to make this initial obstacle-avoidance path more suitable for robot execution, the system smooths this path, which consists of discrete grid points. For example, cubic spline interpolation can be used to use key points on the initial obstacle-avoidance path as control points to generate a continuous curve with smooth curvature. This smoothed curve is the target obstacle-avoidance path that the robot will ultimately follow, ensuring the smoothness and stability of the robot's trajectory during actual movement.

[0083] Through the above technical solution, this application further optimizes the path generation process based on obstacle avoidance path planning using a safe width threshold. First, by determining the passable area based on the target safe width threshold, the safety of the planned path is ensured, avoiding collisions between the robot and obstacles. Second, the A* path search method is used to efficiently search for the initial obstacle avoidance path with the minimum travel cost from the passable area, guaranteeing the effectiveness and efficiency of the path. Finally, the initial obstacle avoidance path is smoothed, eliminating sharp corners and enabling the robot to perform obstacle avoidance tasks more smoothly and in a manner more consistent with its own kinematics, reducing unnecessary acceleration, deceleration, and turning, thereby improving the robot's travel efficiency, stability, and energy utilization. This allows the robot to complete obstacle avoidance tasks more reliably and efficiently when navigating autonomously in complex aquatic environments.

[0084] In some embodiments described above, this application proposes a method for corrective control of a robot based on water flow velocity characteristics and travel deviation characteristics, so that the robot's travel deviation characteristics are constrained by preset corrective constraints. However, in actual underwater or surface environments, the complexity and dynamics of water flow may cause the robot to exhibit various forms of travel deviation. If these deviations are not accurately identified and addressed, the corrective control may be untimely or inaccurate, affecting the robot's navigation stability and task execution efficiency. Therefore, this application further proposes a method for corrective control of a robot. See Figure 4 In one embodiment, the travel deviation characteristics include lateral sideslip deviation and yaw rate of change, and the method for correcting the deviation of the robot includes, but is not limited to, steps S401 to S402.

[0085] Step S401: When the lateral sideslip offset exceeds the lateral sideslip offset threshold and / or the yaw angle change rate exceeds the yaw angle change threshold, the lateral wheel speed difference of the robot is determined based on the water flow velocity characteristics and the travel offset characteristics.

[0086] Step S402: Control the robot to move by the lateral wheel speed difference.

[0087] Lateral sideslip offset refers to the lateral deviation between the robot's actual direction of motion and its longitudinal axis during movement; that is, the lateral drift distance or velocity of the robot relative to its forward direction. This offset can reflect the lateral thrust effect of water flow on the robot, for example, by fusing data from an inertial measurement unit (IMU) with data from a Doppler velocimeter (DVL) or visual odometry, or by calculating data from the robot's attitude and position sensors.

[0088] Yaw angle change rate refers to the rate at which a robot's heading angle changes over time during movement. This rate reflects the stability of the robot's posture and the influence of water flow or its own control system on the robot's heading. The yaw angle change rate can usually be obtained directly from the robot's internal gyroscope or inertial measurement unit (IMU), or it can be obtained by differentiating continuous heading angle data. Lateral sideslip threshold is a preset critical value used to determine whether lateral sideslip requires corrective control. This threshold can be set based on factors such as the robot's design characteristics, task accuracy requirements, environmental water flow intensity, and safety margins, for example, through experimental testing, simulation analysis, or empirical values. When the actual measured lateral sideslip exceeds this threshold, it indicates that the robot has experienced significant lateral drift and requires corrective action.

[0089] The yaw angle change threshold rate is a preset critical value used to determine whether a change in yaw angle requires the initiation of corrective control. This threshold rate can be set based on factors such as the robot's dynamic response characteristics, heading stability requirements, water flow disturbance frequency, and the robustness of the control system. It can be calibrated, for example, through system identification, dynamic model analysis, or actual operational data. When the measured yaw angle change rate exceeds this threshold, it indicates that the robot's heading stability is significantly affected, necessitating corrective control.

[0090] Lateral wheel speed difference refers to the speed difference between the left and right drive wheels (or thrusters) of a robot. Adjusting this difference generates a steering torque, thereby changing the robot's heading or counteracting lateral drift. Determining this difference is a control decision-making process that can employ various control algorithms, such as proportional-integral-derivative (PID) control, fuzzy logic control, adaptive control, or model predictive control. The required wheel speed difference is calculated based on the characteristics of the water flow velocity (such as direction and magnitude) and specific travel deviation characteristics (lateral sideslip deviation, yaw rate of change). Controlling the robot to travel at the lateral wheel speed difference means sending the calculated lateral wheel speed difference command to the robot's drive system (such as a motor controller or thruster controller), causing its left and right drive wheels or thrusters to operate with the corresponding speed difference. This is typically achieved by adjusting the speed of the drive motor or the thrust of the thruster, thereby generating the expected steering or lateral correction force to counteract travel deviation and return the robot to the preset navigation path.

[0091] The proposed solution continuously acquires the water flow velocity characteristics and refined travel deviation characteristics of the robot as it travels along a preset navigation path. These deviation characteristics specifically include lateral sideslip deviation and yaw rate of change. When the lateral sideslip deviation exceeds a preset threshold, or the yaw rate of change exceeds a preset threshold, the system determines that the robot has significantly deviated from its course and requires corrective control. At this point, based on the acquired water flow velocity characteristics and the specific travel deviation characteristics (lateral sideslip deviation and yaw rate of change), a specific control algorithm precisely calculates the required lateral wheel speed difference between the robot's left and right drive wheels or propellers. This calculated lateral wheel speed difference command is then sent to the robot's drive system, causing the robot to travel at this difference. In this way, the robot can generate a corrective force or torque opposite to the direction of deviation, effectively counteracting the lateral drift and heading deviation caused by the water flow, thus bringing the robot's travel deviation characteristics back within the preset corrective control constraints. This scheme improves the stability of the robot's autonomous navigation and the accuracy of path tracking by refining the identification of travel deviation characteristics and setting threshold conditions for triggering correction.

[0092] The following example illustrates this: During movement, the robot utilizes its onboard Inertial Measurement Unit (IMU) and Doppler Velocimetry (DVL) to acquire its motion status in real time. The DVL provides the robot's velocity information, which, combined with the attitude data from the IMU, allows for precise calculation of the robot's lateral sideslip. Simultaneously, the gyroscope within the IMU provides high-precision angular velocity information, directly enabling the acquisition of the robot's yaw rate of change. The actuator pre-sets threshold values ​​for lateral sideslip and yaw rate of change. For example, when the lateral sideslip exceeds a certain safe range, or the yaw rate of change exceeds a certain stable range, corrective control is triggered. At this point, the robot's internal controller comprehensively considers the current water flow velocity characteristics (e.g., obtained through an Acoustic Doppler Current Profiler (ADCP)) and the real-time lateral sideslip and yaw rate of change, employing a control strategy, such as a Proportional-Integral-Derivative (PID) controller, to calculate the required speed difference between the left and right thrusters, i.e., the lateral wheel speed difference. Subsequently, the controller sends these differential commands to the motor drivers of the left and right thrusters, causing the two wheels to operate at different speeds, thereby generating a steering torque to counteract lateral slippage and yaw angle changes, enabling the robot to move stably along the preset navigation path.

[0093] The above technical solution refines the travel deviation characteristics into lateral sideslip deviation and yaw rate of change, and sets corresponding thresholds as trigger conditions for correction control, enabling the robot to more accurately identify its specific deviation state in a flowing water environment. Based on these refined deviation characteristics and water flow velocity characteristics, correction can be achieved by determining and applying lateral wheel speed differences, enabling rapid and accurate correction of the robot's lateral drift and heading deviation. This significantly improves the robot's path tracking accuracy and navigation stability under complex water flow conditions, avoiding over-correction or under-correction caused by ambiguous deviation judgments, thereby improving the reliability of the robot's autonomous navigation and task execution efficiency.

[0094] In some embodiments, determining the lateral wheel speed difference of the robot based on water flow velocity characteristics and travel deviation characteristics includes: determining a comprehensive deviation coefficient based on the lateral sideslip deviation and yaw angle change rate; determining corresponding proportional coefficients, integral coefficients, and differential coefficients based on the comprehensive deviation coefficients and water flow velocity characteristics; and performing proportional-integral-differential operations on the lateral sideslip deviation based on the proportional coefficients, integral coefficients, and differential coefficients to obtain the lateral wheel speed difference.

[0095] The purpose of determining the comprehensive deviation coefficient is to integrate the two key travel deviation characteristics of the robot—lateral sideslip deviation and yaw rate of change—into a unified index. This comprehensive deviation coefficient can more comprehensively reflect the degree and trend of the robot's deviation from the preset path, providing more accurate input for subsequent corrective control. For example, the comprehensive deviation coefficient can be calculated as a weighted sum of the lateral sideslip deviation and the yaw rate of change, where the weights can be set according to the sensitivity of lateral and angular deviations in the actual application scenario; alternatively, a fuzzy logic system can be used, taking the lateral sideslip deviation and the yaw rate of change as inputs, and outputting the comprehensive deviation coefficient through preset fuzzy rules and membership functions to handle nonlinearity and uncertainty.

[0096] Determining the appropriate proportional, integral, and derivative coefficients is crucial for achieving adaptive control. These coefficients are core parameters in a proportional-integral-derivative (PID) controller, corresponding to the response strength to the current error, historical error accumulation, and error rate of change, respectively. By dynamically adjusting these coefficients in conjunction with the comprehensive deviation coefficient and water flow velocity characteristics, the correction control strategy can better adapt to current environmental conditions and the robot's deviation state. For example, a parameter lookup table can be pre-established to store the corresponding proportional, integral, and derivative coefficients based on different comprehensive deviation coefficients and water flow velocity characteristic ranges; alternatively, adaptive algorithms, such as gain scheduling or model reference adaptive control, can be used to optimize these coefficients in real time based on system performance and environmental changes.

[0097] Performing proportional-integral-derivative (PID) calculations on the lateral sideslip offset is the core step in generating precise control commands. PID is a feedback control method that generates the control output by calculating the proportional, integral, and derivative terms of the error. The proportional term responds to the current error, the integral term eliminates the steady-state error, and the derivative term predicts the error's trend. The combination of these three terms provides stable, fast, and overshoot-free control. For example, this calculation can be performed according to the standard PID control formula: the control output equals the proportional coefficient multiplied by the current lateral sideslip offset, plus the cumulative sum of the integral coefficient multiplied by the lateral sideslip offset, plus the derivative coefficient multiplied by the rate of change of the lateral sideslip offset. Alternatively, an incremental PID algorithm can be used to calculate the increment of the control output, achieving smoother control and avoiding integral saturation.

[0098] This application's solution integrates lateral sideslip offset and yaw angle change rate into a comprehensive deviation coefficient, enabling a more comprehensive assessment of the robot's deviation state. Based on this, and combined with water flow velocity characteristics, the proportional, integral, and derivative coefficients are dynamically determined, allowing the proportional-integral-derivative (PID) calculations to adaptively adjust according to the actual environment and the degree of deviation. This adaptive PID control strategy generates more accurate and stable lateral wheel speed differences, effectively correcting the robot's deviation. Compared to fixed-parameter correction methods, this solution significantly improves the robot's navigation accuracy and stability in complex water flow environments, ensuring the robot can more accurately travel along the preset path.

[0099] The following is a concrete example. Suppose an underwater robot is performing an autonomous navigation task. When the robot's sensors detect a lateral sideslip deviation (e.g., 0.3 meters) and a yaw rate of change (e.g., 1 degree / second), the control system first inputs these two values ​​into the comprehensive deviation coefficient calculation module. This module may use a weighted average algorithm to calculate a comprehensive deviation coefficient, for example, 0.6, from the 0.3-meter lateral sideslip deviation and the 1-degree / second yaw rate of change (after appropriate normalization). Simultaneously, the robot's onboard water flow sensor detects that the current water flow velocity is moderately fast (e.g., 0.8 meters / second). Based on this comprehensive deviation coefficient of 0.6 and the water flow velocity of 0.8 meters / second, the control system consults a preset adaptive gain table or calculates it through an algorithm to dynamically determine the most suitable proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd, for example, Kp=0.7, Ki=0.15, and Kd=0.1. Subsequently, the control system uses these dynamically determined Kp, Ki, and Kd values ​​to perform proportional-integral-differential calculations on the current lateral slip offset of 0.3 meters. For example, if the historical error accumulation term is 0.1 m / s and the error change rate is -0.05 m / s, the calculated lateral wheel speed difference might be 0.7 * 0.3 + 0.15 * 0.1 + 0.1 * (-0.05) = 0.21 + 0.015 - 0.005 = 0.22 m / s. Finally, the robot adjusts the rotation speed of its left and right thrusters based on this lateral wheel speed difference of 0.22 m / s, thereby generating a lateral thrust that pushes the robot back onto the preset navigation path.

[0100] In a specific embodiment, the formula for calculating the comprehensive deviation coefficient is: , , in, This is the comprehensive deviation coefficient. This is the lateral sideslip offset. The rate of change of yaw angle. and These are the weighting coefficients. This represents the actual lateral movement position of the robot. The baseline position; The formula for calculating the lateral wheel speed difference is: , in, For the first The lateral wheel speed difference at each moment. This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients. ; The proportional coefficient, integral coefficient, and differential coefficient are adapted based on the characteristics of water flow velocity and the deviation level. Since tailings dam tunnel drainage is divided into valve closed, 1 / 2 open, and fully open, the adaptation rules are as follows: When the valve is closed, if , Take 0.8-1.0, Take 0.1-0.2. Take 0.3-0.4; if , Take 1.0-1.2, Take 0.2-0.3. Take 0.4-0.5; When the valve is opened 1 / 2, if , Take 1.0-1.2, Take 0.2-0.3. Take 0.4-0.5; if , Take 1.2-1.4. Take 0.3-0.4. Take 0.5-0.6; When the valve is fully open, if , Take 1.2-1.4. Take 0.3-0.4. Take 0.5-0.6; if , Take 1.4-1.5. Take 0.4-0.5. Take 0.6-0.7; If the lateral wheel speed difference ( (This refers to the current robot speed), and the distance between the robot and the cave wall / pipe is greater than or equal to the target safe width threshold. At that time, no action is taken if the lateral wheel speed difference is... If the distance between the robot and the cave wall or pipe violates any constraint, reduce the lateral wheel speed difference by 50%. .

[0101] Through the above technical solution, when the robot performs deviation control, it can comprehensively consider the lateral sideslip offset and the rate of change of yaw angle to form a more comprehensive deviation assessment. Simultaneously, by dynamically adjusting the proportional, integral, and derivative coefficients based on the water flow velocity characteristics, the proportional-integral-differential calculations can be performed more accurately and adaptably to the current environment, resulting in a more reasonable and effective lateral wheel speed difference. This significantly improves the accuracy and stability of the robot's deviation control, effectively avoiding problems such as overshoot, oscillation, or response lag that may occur in complex water flow environments, ensuring that the robot can more accurately maintain the preset navigation path and improving the reliability of autonomous navigation.

[0102] In some embodiments, the robot autonomous navigation method provided in this application further includes: using a bidirectional path matching method to reverse analyze a preset navigation path to obtain a corresponding return navigation path, so that the robot returns according to the return navigation path; and controlling the robot to reduce its speed to a preset speed threshold when the robot travels a first preset mileage along the preset navigation path and / or travels a second preset mileage along the return navigation path.

[0103] Bidirectional path matching aims to generate a corresponding, reversed return path based on an existing pre-defined navigation path. Its purpose is to provide the robot with an optimized path suitable for return, rather than simply reversing the forward path. Implementation methods can include: using graph search algorithms (e.g., A* algorithm, Dijkstra's algorithm, or variants) based on the topological information of the pre-defined navigation path for path planning, but searching from the end point to the starting point of the pre-defined navigation path; or, analyzing the geometric features of the pre-defined navigation path and combining it with environmental map information to generate a geometrically symmetrical or complementary path to the pre-defined navigation path.

[0104] Reverse analysis is a key step in bidirectional path matching methods. It involves analyzing and processing a pre-defined navigation path to extract the necessary information for generating a return navigation path. Specifically, this can involve reversing the sequence of waypoints in the pre-defined navigation path and reassessing the connectivity and transit costs between them; or, based on the geometric curve of the pre-defined navigation path, calculating a reverse curve from the endpoint to the starting point, and then optimizing the path accordingly.

[0105] The return navigation path is a path obtained through bidirectional path matching and reverse analysis, allowing the robot to return from the task endpoint to the starting point. This path may differ geometrically from the preset navigation path because it may take into account specific conditions during the return journey, such as energy consumption, environmental changes, or obstacle avoidance strategies.

[0106] This application's solution effectively addresses the efficiency and safety issues of robot return after task completion by introducing a complete return-to-home mechanism and safe speed control in key areas. Specifically, before or during task execution, the system uses a bidirectional path matching method to reverse-engineer the preset navigation path, generating a dedicated return-to-home navigation path. This optimized path guides the robot to return efficiently and safely from the endpoint of the preset path to the starting point. Once the robot successfully reaches the endpoint and completes its forward task, its control system switches to return-to-home mode, precisely controlling the robot to travel along the pre-planned path. Furthermore, to further enhance operational safety, especially when the robot approaches the task's starting or ending point—areas with potentially complex environments or requiring precise manipulation—this solution also includes a mileage-triggered deceleration mechanism. Specifically, when the robot travels a distance along the preset navigation path reaching a first preset mileage, or along the return-to-home navigation path reaching a second preset mileage, the system automatically reduces the robot's speed to a preset speed threshold. This phased speed control strategy enables the robot to not only autonomously plan and track its path during the entire mission cycle (including outward and return journeys), but also to obtain additional safety guarantees at key nodes. This significantly reduces the potential risks caused by high-speed travel in complex waters or narrow passages, ensuring the successful completion of the mission and the safe recovery of the robot.

[0107] As a specific implementation method, the return-to-home and deceleration control in robot autonomous navigation can be achieved in the following way. First, before the robot starts its task, its navigation system can generate a return-to-home navigation path based on a preset navigation path and using a graph search-based path planning algorithm (e.g., a variant of the A* algorithm) to perform a reverse search and optimization from the endpoint of the preset navigation path towards the starting point. This return-to-home navigation path can be stored as a series of ordered waypoints, each containing position coordinates and direction information. When the robot travels along the preset navigation path and detects that the distance between its current position and the endpoint of the preset navigation path is less than a preset threshold (e.g., 1 meter), the system determines that the robot has reached the endpoint. At this time, the robot will stop forward navigation and begin to perform the return-to-home task, and its motion controller will load and track the previously generated return-to-home navigation path. To ensure safety during the return journey, for example, when the robot travels along the preset navigation path to within 5 meters of the starting point (i.e., the first preset distance is 5 meters), or when the robot travels along the return navigation path to within 5 meters of the return destination (i.e., the second preset distance is 5 meters), the robot's speed controller will automatically limit the maximum travel speed to a preset speed threshold of 0.3 meters per second. This speed limit will remain until the robot has completely stopped or left the safe area, thus effectively avoiding potential dangers caused by excessive speed in narrow or busy areas.

[0108] Through the above technical solution, this application provides a more complete autonomous navigation method for robots. This method not only guides the robot to perform forward tasks in complex environments, but also ensures efficient and orderly return after task completion through a pre-planned return navigation path. Especially when the robot approaches critical areas such as the task start or end point, the implementation of automatic deceleration control significantly improves the robot's safety in these sensitive areas, reduces operational risks, and makes the entire task cycle of the robot more reliable and controllable.

[0109] Figure 5 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application.

[0110] The following reference Figure 5 To describe an electronic device 500 according to such an embodiment of the present disclosure. Figure 5 The electronic device 500 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0111] like Figure 5As shown, the electronic device 500 is presented in the form of a general-purpose computing device. The components of the electronic device 500 may include, but are not limited to: at least one processing unit 510, at least one storage unit 520, a bus 530 connecting different system components (including storage unit 520 and processing unit 510), a display unit 540, etc.

[0112] The storage unit stores program code, which can be executed by the processing unit 510, causing the processing unit 510 to perform the steps described in the above section of the robot autonomous navigation method according to various exemplary embodiments of this disclosure.

[0113] Storage unit 520 may include a readable medium in the form of a volatile storage unit, such as random access memory (RAM) 5201 and / or cache memory 5202, and may further include a read-only memory (ROM) 5203.

[0114] Storage unit 520 may also include a program / utility 5204 having a set (at least one) program module 5205, such program module 5205 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0115] Bus 530 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0116] Electronic device 500 can also communicate with one or more external devices 570 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 500, and / or with any device that enables electronic device 500 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 550. Furthermore, electronic device 500 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 560. Network adapter 560 can communicate with other modules of electronic device 500 via bus 530. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 500, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0117] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0118] The robot autonomous navigation method, device, and medium provided in this application acquire spatial obstacle features, water flow velocity features, and robot deviation features in the travel space during the robot's journey along a preset navigation path. Based on these features, the robot performs obstacle avoidance control and deviation correction control, enabling it to be constrained by preset conditions. This allows the robot to effectively cope with high-speed water flow and complex spatial obstacles, improving navigation accuracy and operational safety in complex environments such as tailings dam drainage tunnels, and reducing the risk of collisions during autonomous navigation, thus adapting to navigation scenarios in tailings dam drainage tunnels.

[0119] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, or network device, etc.) to execute the methods described above according to the embodiments of this disclosure.

[0120] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0121] Computer-readable storage media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable storage medium may also be any readable medium other than a readable storage medium that can transmit, propagate, or transfer a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0122] Those skilled in the art will understand that the above modules can be distributed in the device as described in the embodiments, or they can be modified accordingly and placed in one or more devices that are unique to this embodiment. The modules in the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0123] Exemplary embodiments of this disclosure have been specifically shown and described above. It should be understood that this disclosure is not limited to the detailed structures, arrangements, or implementations described herein; rather, this disclosure is intended to cover various modifications and equivalent arrangements contained within the spirit and scope of the appended claims.

Claims

1. A method for autonomous navigation of a robot, characterized in that, include: While the robot is traveling along the preset navigation path, the spatial obstacle characteristics, water flow velocity characteristics, and travel deviation characteristics of the robot's location are acquired. Based on the preset navigation path and the spatial obstacle features, the robot is controlled to bypass obstacles, so that the robot passes through the current position with a corresponding preset obstacle bypass strategy. Based on the water flow velocity characteristics and the travel deviation characteristics, the robot is subjected to correction control so that the robot's travel deviation characteristics are constrained by preset correction constraints.

2. The robot autonomous navigation method according to claim 1, characterized in that, The obstacle avoidance control of the robot based on the preset navigation path and the spatial obstacle features includes: Based on the preset navigation path, the first access risk characteristic of the robot's location is determined; Based on the spatial obstacle characteristics, a second passage risk characteristic of the robot's location is determined; When the first passage risk feature meets the first obstacle avoidance condition, the robot is controlled to execute the first preset obstacle avoidance strategy; when the second passage risk feature meets the second obstacle avoidance condition, the robot is controlled to execute the second preset obstacle avoidance strategy.

3. The robot autonomous navigation method according to claim 2, characterized in that, The first obstacle avoidance condition includes at least one of the following: The turning angle of the robot's position exceeds the turning angle threshold; The radius of curvature of the path at the robot's location exceeds the radius of curvature threshold. The first preset obstacle avoidance strategy is to control the robot to retract its robotic arm and remain in the middle area of ​​its current position, traveling along the preset navigation path.

4. The robot autonomous navigation method according to claim 2, characterized in that, The second obstacle avoidance condition includes at least one of the following: The width of the robot's location exceeds the first width threshold but does not exceed the second width threshold; The width of the robot's location does not exceed the first width threshold; The second preset obstacle avoidance strategy is as follows: when the width exceeds the first width threshold but does not exceed the second width threshold, the robot performs obstacle avoidance path planning based on the first safe width threshold and the width value of the robot's position; when the width does not exceed the first width threshold, the robot performs obstacle avoidance path planning based on the second safe width threshold and the width value of the robot's position, and controls the robot to move along the planned target obstacle avoidance path; the first safe width threshold is determined based on the width value of the robot when its robotic arm is extended, and the second safe width threshold is determined based on the width value of the robot when its robotic arm is retracted.

5. The robot autonomous navigation method according to claim 4, characterized in that, The obstacle avoidance path planning for the robot includes: Based on a target safe width threshold, a passable area is determined in the robot's location; the target safe width threshold is either the first safe width threshold or the second safe width threshold. The A* path search method is used to search for the path with the minimum travel cost in each of the passable areas to obtain the initial obstacle bypass path; The initial obstacle avoidance path is smoothed to obtain the target obstacle avoidance path.

6. The robot autonomous navigation method according to claim 1, characterized in that, The travel deviation characteristics include lateral sideslip deviation and yaw angle change rate. The corrective control of the robot based on the water flow velocity characteristics and the travel deviation characteristics includes: When the lateral sideslip offset exceeds the lateral sideslip offset threshold and / or the yaw angle change rate exceeds the yaw angle change threshold, the lateral wheel speed difference of the robot is determined based on the water flow velocity characteristics and the travel offset characteristics. The robot is controlled to travel at the lateral wheel speed difference.

7. The robot autonomous navigation method according to claim 6, characterized in that, Determining the lateral wheel speed difference of the robot based on the water flow velocity characteristics and the travel offset characteristics includes: Based on the lateral sideslip offset and the yaw angle change rate, a comprehensive deviation coefficient is determined; Based on the comprehensive deviation coefficient and the water flow velocity characteristics, the corresponding proportional coefficient, integral coefficient and differential coefficient are determined; Based on the proportional coefficient, the integral coefficient, and the differential coefficient, the lateral sideslip offset is subjected to proportional-integral-differential calculations to obtain the lateral wheel speed difference.

8. The robot autonomous navigation method according to claim 1, characterized in that, Also includes: A bidirectional path matching method is used to reverse-parse the preset navigation path to obtain the corresponding return navigation path, so that the robot returns to its home position according to the return navigation path. When the robot travels a first preset mileage along the preset navigation path and / or a second preset mileage along the return navigation path, the robot is controlled to reduce its speed to a preset speed threshold.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the robot autonomous navigation method according to any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the robot autonomous navigation method according to any one of claims 1 to 8.

Images