A quadruped robot hierarchical path planning method and system considering power safety potential and gait switching cost

Abstract

Disclosed is a kind of four-legged robot hierarchical path planning method and system considering electric power safety potential and gait switching cost, first, build the integrated three-dimensional cost map containing ground geometric features and three-dimensional electric field safety potential field, map the risk of electric shock in space to potential cost value;Second, define the combined state containing position and gait mode, build the global search criteria integrating movement cost, risk cost and gait mode switching cost;Then, execute joint path search, automatically force the robot to switch to low posture crawl mode in high electric field potential area to increase vertical safety distance;Finally, the foot end landing suitability of the path is verified by using virtual straight corridor, and redundant nodes are removed based on terrain elevation dispersion and normal inclination. The invention solves the problems of four-legged robot easily entering high-voltage area, frequent gait switching energy consumption and smooth path easily leading to foot end slipping in substation inspection, and realizes safe, efficient and stable autonomous inspection.

Description

Technical Field

[0001] This invention relates to the technical field of robot motion planning and control, and in particular to a hierarchical path planning method and system for quadruped robots that takes into account electrical safety potential energy and gait switching costs. Background Technology

[0002] With the advancement of new power system construction, substations, as the core hubs of the power grid, are gradually shifting their operation and maintenance model from manual inspection to unmanned, intelligent inspection. Quadruped robots, with their superior adaptability to unstructured terrain and omnidirectional mobility, have become an important tool for performing inspection tasks in substations, especially in complex substation scenarios undergoing renovation and expansion. These robots are typically equipped with LiDAR, dual-spectrum cameras, and robotic arms to perform tasks such as equipment visual inspection, infrared temperature measurement, and switchgear operation.

[0003] Current path planning technology for quadruped robots mainly follows the approach used for wheeled robots, namely, constructing a two-dimensional occupancy grid map or a three-dimensional point cloud map based on SLAM (Simultaneous Localization and Mapping) technology, and utilizing AI... Global path search is performed using algorithms such as RRT (Rapid Expanding Random Tree) or Dijkstra's algorithm. To optimize path smoothness, existing techniques often employ polynomial interpolation, Bézier curves, or B-spline curves to geometrically smooth the polyline path, thereby reducing robot motion pauses.

[0004] However, in complex operational scenarios involving 500kV and above ultra-high voltage substations, especially in renovation and expansion projects, the aforementioned conventional path planning methods reveal significant safety hazards and efficiency bottlenecks, specifically in the following three aspects: First, there is a lack of ability to perceive and avoid "non-contact" electric shock risks in the vertical dimension. Existing path planning algorithms typically treat environmental obstacles as rigid entities (such as walls and enclosures), focusing primarily on obstacle avoidance at the horizontal level. However, in substation environments, there are invisible strong electric fields around high-voltage live equipment. Air breakdown discharge does not require physical contact; it can occur simply by intruding into a specific safe clearance distance. At renovation and expansion construction sites, the ground environment is extremely complex, often with temporary piles of earth, building materials, or scaffolding. These obstacles effectively "raise" the robot's motion base. When a quadruped robot climbs to the top of a pile of earth or raises its robotic arm to perform a detection task, its apex envelope height increases significantly, causing a sharp decrease in the vertical distance between the robot and the high-voltage busbar above. This can easily exceed the safety insulation limit, inducing a severe power grid flashover accident. Existing two-dimensional or purely geometric three-dimensional planning methods cannot dynamically perceive this vertical electric field risk caused by the coupling of terrain elevation and posture extension.

[0005] Second, the decoupling of path planning and gait mode leads to high energy consumption and motion oscillations. Quadruped robots possess multiple motion modes, including walking, trotting, crawling, and jumping, with significant differences in energy consumption and stability between these modes. Existing technologies typically employ a hierarchical control architecture, where the upper layer plans the path, and the lower layer passively switches gait modes based on the terrain. This decoupling strategy ignores the cost of gait switching itself. In complex construction sites, to pursue the geometrically shortest path, the robot may be planned to traverse discontinuous and complex terrain, requiring frequent switching between standing and crawling. Frequent posture adjustments not only impose huge acceleration and deceleration impact loads on the joint servo motors, significantly shortening battery life, but also cause pitch oscillations in the robot body, affecting the imaging quality of onboard precision testing instruments (such as partial discharge detectors and infrared thermal imagers).

[0006] Third, path smoothing based on pure geometry lacks physical suitability verification of the foot landing point. To eliminate A... The jagged paths generated by algorithms are widely smoothed using mathematical curves such as B-splines in existing technologies. However, the movement of quadruped robots depends on discrete foot support. Purely geometric smoothing often only considers the continuity of the curve, ignoring the microscopic terrain features beneath it. The smoothed path is highly likely to pull the robot's footing to the edge of a pile of gravel, the sloping surface of a smooth metal cover, or a soft sand pit. For wheeled robots, this might only result in slight slippage; but for quadruped robots, the foot landing in unreliable areas with high roughness or large inclination angles can easily lead to slippage, missteps, or even the entire robot tipping over. Existing technologies lack a mechanism to pre-verify the foot mechanics of the proposed straight path during the path smoothing stage.

[0007] In summary, a hierarchical path planning method for quadruped robots that can proactively sense potential risks in the spatial electric field, collaboratively optimize gait mode selection, and verify the physical suitability of foot landing points is an urgent need to resolve the current contradiction between safety and efficiency in intelligent substation inspections. Summary of the Invention

[0008] To address the aforementioned technical problems in the existing technology, this invention proposes a hierarchical path planning method and system for quadruped robots that takes into account both electrical safety potential energy and gait switching costs, thereby solving the above-mentioned technical problems.

[0009] According to a first aspect of the present invention, a hierarchical path planning method for a quadruped robot that takes into account electrical safety potential energy and gait switching costs is proposed, comprising: S1: Acquire laser point cloud data and digital twin model information of the substation operation area to construct a grid map reflecting the geometric features of the ground; at the same time, based on the voltage level and spatial location of the live equipment in the digital twin model, construct a three-dimensional electric field safety potential energy field and map this potential energy field onto the space above the grid map to form a comprehensive three-dimensional cost map that takes into account both the surface traffic impedance and the risk of electric shock in space. S2: Define the path planning state of the quadruped robot as a combination of position coordinates, heading angle and gait mode; set the cost criterion for global search, which is composed of movement distance cost, spatial safety risk cost and gait mode switching cost; S3: In the comprehensive 3D cost map, the optimal path from the starting point to the ending point is searched based on the cost criterion, and the initial planned path containing the position sequence and the corresponding gait mode sequence is output; during the search process, when the planned path enters a high electric potential energy region, the algorithm automatically constrains the gait mode to a low posture mode to increase the safe distance in the vertical direction; and S4: Extract key turning nodes in the initial planned path, and perform foot landing suitability verification on straight paths between non-adjacent nodes; only when the terrain flatness of the area covered by the straight path meets the anti-slip requirements of the quadruped robot's feet, remove intermediate redundant nodes and generate the final execution path.

[0010] In some specific embodiments, the three-dimensional electric field safety potential energy field in S1 is constructed as follows: Using the geometric center of the charged equipment as the field source, a distribution model is established to show that the potential energy intensity decreases with increasing spatial distance. A safe breakdown restricted area is defined, and the spatial area that is less than the restricted area from the energized equipment is marked as an infinite potential energy zone and is regarded as an impassable obstacle. The spatial electric shock risk cost in the comprehensive three-dimensional cost map is positively correlated with the robot's planned height. When the robot adopts different time modes at the same horizontal position, different electric field potential energy costs are triggered due to the different body heights.

[0011] In some specific embodiments, the potential energy value of the three-dimensional electric field safety potential energy field It's about spatial location. And the current planned height of the robot The function is calculated using the following formula: ,in, The number of electrically charged devices to be sensed. For the first The nominal voltage level of each live device The coordinates of the equipment center are Let the coordinates of the top envelope point of the robot be given in the planned pose. For safe insulation distance, This is an adjustment coefficient; when the distance is less than... hour, Let it be infinity.

[0012] In some specific embodiments, in S1, the construction of the integrated environment map also includes the fusion of semantic attributes: The ground grid is divided into rigid impassable areas, flexible obstacle-crossing areas, and flat passable areas; For flexible obstacle-crossable areas, the cost criterion assigns a travel cost value higher than that of flat traversable areas but lower than that of rigid impassable areas, allowing the path planning algorithm to plan a path through gravel or shallow pit areas when no flat path is available.

[0013] In some specific embodiments, the calculation logic for the gait mode switching cost in S2 is as follows: Multiple gait modalities are preset, including at least a high-posture walking mode, a low-posture crawling mode, and an obstacle-crossing mode; During the node expansion process in path search, the gait mode selected by the current node to be expanded is compared with the gait mode of the parent node; If the two gait modes are inconsistent, an additional switching penalty value is added to the cumulative cost. The magnitude of the switching penalty value is proportional to the energy consumption and time cost of joint adjustment required for the robot to perform gait transition, so as to suppress the robot from frequently switching motion postures when unnecessary.

[0014] In some specific embodiments, the triggering and deactivation of the low-pose mode in S3 adopts a hysteresis comparison strategy, the specific logic of which is as follows: Triggering logic: When the electric potential energy value at the robot's planned position exceeds the preset warning threshold, the gait mode in the planned state is immediately locked to the low-posture crawling mode. Release logic: When the electric field potential energy value at the robot's planned position drops below the warning threshold, it does not immediately resume a high posture, but maintains a low posture crawling mode to continue planning a preset safe buffer distance, and then releases the height lock, allowing the algorithm to automatically select the gait based on the movement cost.

[0015] In some specific embodiments, the foot landing suitability check in S4 specifically includes: Using the line connecting the two non-adjacent nodes to be verified as the central axis, a virtual straight corridor with a certain width is constructed. The width is adapted to the lateral standing span of the quadruped robot to cover the foot landing area when the robot actually walks. Extract all surface raster data within the coverage area of ​​the virtual straight corridor, and calculate the comprehensive terrain stability index of the path segment based on elevation dispersion and normal tilt. The specific calculation formula is as follows: ,in: This represents the total number of valid sampling points within the virtual linear corridor. For the first The relative elevation values ​​of each sampling point This represents the average elevation within the corridor. For the first Local fitting plane normal vector at each sampling point; The unit vector of gravity in the opposite direction in the world coordinate system; This characterizes the degree of inclination of the Earth's surface. These are preset weighting coefficients used to balance the effects of ruggedness and inclination on foot stability; The calculated results With the preset foot stability threshold Perform a comparison; if If the terrain features of the straight path meet the requirements for full foot contact of the quadruped robot, node removal and path straightening operations are allowed; otherwise, if there is a risk of slipping or falling out, the intermediate nodes in the original path are retained to maintain the detour trajectory.

[0016] In some specific embodiments, the method further includes a dynamic obstacle avoidance step: During the robot's movement along the final execution path, the onboard LiDAR is used to detect dynamic obstacles in the local environment in real time; When a dynamic obstacle is detected encroaching on the current path, the local path planning algorithm is activated to generate an obstacle avoidance trajectory. When generating obstacle avoidance trajectories, the three-dimensional electric field safety potential energy field constructed in S1 is introduced as a hard constraint into the evaluation function of the local programming algorithm to prevent the obstacle avoidance trajectory from rising upwards or encroaching on the vertical safety clearance area below the live equipment.

[0017] According to a second aspect of the invention, a computer-readable storage medium is provided on which one or more computer programs are stored, which, when executed by a computer processor, implement the method described above.

[0018] According to a third aspect of the present invention, a hierarchical path planning system for a quadruped robot that takes into account electrical safety potential energy and gait switching costs is proposed, comprising: The integrated environmental modeling module is configured to acquire laser point cloud data and digital twin model information of the substation operation area and construct a grid map that reflects the geometric features of the ground. At the same time, based on the voltage level and spatial location of the live equipment in the digital twin model, a three-dimensional electric field safety potential energy field is constructed and mapped onto the space above the grid map to form a comprehensive three-dimensional cost map that takes into account both the surface traffic impedance and the risk of electric shock in space. The state and cost definition module is configured to define the path planning state of the quadruped robot as a combination of position coordinates, heading angle and gait mode; and to set the cost criterion for global search, which is composed of movement distance cost, spatial safety risk cost and gait mode switching cost. The joint path search module is configured to search for the optimal path from the starting point to the ending point based on the cost criterion in a comprehensive 3D cost map, and output an initial planned path containing a position sequence and a corresponding gait mode sequence. During the search process, when the planned path enters a high electric potential energy region, the gait mode is automatically constrained to a low-posture mode to increase the safe distance in the vertical direction; and The path smoothing optimization module is configured to extract key turning nodes in the initial planned path and perform foot landing suitability verification on straight paths between non-adjacent nodes. Only when the terrain flatness of the area covered by the straight path meets the anti-slip requirements of the quadruped robot's feet, intermediate redundant nodes are removed, and the final execution path is generated.

[0019] This invention proposes a hierarchical path planning method and system for quadruped robots that takes into account both electrical safety potential energy and gait switching costs. Compared with existing technologies, it has the following advantages: This invention overcomes the limitations of traditional two-dimensional planning, which ignores the risks of the vertical electric field, by constructing a three-dimensional electric field safety potential energy field that integrates voltage level and spatial distance. This significantly improves the inherent safety of quadruped robots with arms. Addressing common scenarios in construction and renovation projects involving terrain elevation and high-arm operation, this invention can sensitively detect the risk of reduced vertical clearance due to terrain rise or extended posture. Through an algorithm, it automatically triggers a "low-posture crawling mode," actively lowering the robot's height without altering its horizontal path. This physically increases the air insulation distance between the robot and the overhead high-voltage busbar, effectively preventing grid flashover and robot damage accidents caused by insufficient safety distance.

[0020] This invention creatively proposes a gait mode switching cost function, achieving global coordination between spatial path and motion mode, and significantly reducing the robot's operational energy consumption. During the path search phase, instead of blindly pursuing the geometric shortest path, it tends to plan a coherent path that maintains a single, efficient gait. This effectively suppresses frequent standing-crawling switching by the robot under unnecessary circumstances, avoiding energy loss and mechanical wear caused by frequent acceleration and deceleration of the high-power joint motors in quadruped robots. This significantly extends the robot's single-operation endurance under complex working conditions.

[0021] This invention introduces a foot landing suitability verification mechanism based on a virtual straight corridor, solving the problem of mismatch between geometric path smoothness and foot landing physical characteristics in existing technologies, and greatly enhancing the robot's anti-slip capability. Through a specific terrain comprehensive stability index model, the elevation dispersion and normal inclination of the path's micro-regions are calculated, ensuring that each shortcut segment after algorithm straightening meets the physical and mechanical requirements of a quadruped robot's full foot contact. This fundamentally eliminates the risks of slipping, misstepping, and tipping over caused by blindly smoothing the path, guaranteeing the robot's walking stability on unstructured surfaces such as gravel and slopes.

[0022] This invention employs a hysteresis comparison strategy to control the triggering and deactivation of the low-posture mode, eliminating the posture oscillation problem when the robot is at the edge of a high-voltage electric field. When the robot leaves the high-risk area, it is controlled to maintain a low posture and crawl for a safe buffer distance before resuming a high posture. This design, which conforms to control engineering principles, effectively prevents the robot from falling into a "stand up-lie down" vicious cycle at the safety boundary due to small fluctuations in the electric field potential energy value, ensuring the smoothness of the inspection action and the robustness of the control system. Attached Figure Description

[0023] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments and, together with the description, serve to explain the principles of the invention. Other embodiments and many anticipated advantages of the embodiments will be readily recognized as they become better understood through reference to the following detailed description. Other features, objects, and advantages of this application will become more apparent from reading the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart of a hierarchical path planning method for a quadruped robot that takes into account electrical safety potential energy and gait switching costs, according to one embodiment of this application. Figure 2 This is a flowchart illustrating a specific embodiment of the method for constructing a three-dimensional electric field safety potential energy field according to this application. Figure 3 This is a flowchart illustrating the calculation logic of gait mode switching cost according to a specific embodiment of this application; Figure 4 This is a flowchart of the triggering and deactivation logic of the low-posture mode in S3 of a specific embodiment of this application; Figure 5 This is a flowchart of a specific embodiment of the foot landing suitability verification in this application; Figure 6 This is an architecture diagram of a hierarchical path planning system for a quadruped robot that takes into account the potential energy for electrical safety and the cost of gait switching, according to one embodiment of this application. Figure 7This is a schematic diagram of the structure of a computer system used to implement the electronic device of the present application. Detailed Implementation

[0024] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0025] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0026] Figure 1 A flowchart illustrating a hierarchical path planning method for a quadruped robot, considering electrical safety potential energy and gait switching costs according to an embodiment of this application, is shown. Figure 1 As shown, the method includes the following steps: S1: Acquire laser point cloud data and digital twin model information of the substation operation area to construct a grid map reflecting the geometric features of the ground; at the same time, based on the voltage level and spatial location of the live equipment in the digital twin model, construct a three-dimensional electric field safety potential energy field and map this potential energy field onto the space above the grid map to form a comprehensive three-dimensional cost map that takes into account both the surface traffic impedance and the risk of electric shock in space.

[0027] In a specific embodiment, Figure 2 A flowchart illustrating a specific embodiment of this application for constructing a three-dimensional electric field safety potential energy field is shown, as follows: Figure 2 As shown, the three-dimensional electric field safety potential energy field is constructed as follows: S11: Using the geometric center of the charged equipment as the field source, establish a distribution model for the potential energy intensity to decrease with increasing spatial distance; S12: Set a safe breakdown restricted area, and mark the space area that is less than the restricted area from the energized equipment as an infinite potential energy area, which is regarded as an impassable obstacle; S13: The spatial electric shock risk cost in the comprehensive three-dimensional cost map is positively correlated with the robot's planned height. When the robot adopts different time modes at the same horizontal position, different electric field potential energy costs are triggered due to the different body heights.

[0028] In a specific embodiment, the potential energy value of the three-dimensional electric field safety potential energy field is... It's about spatial location. And the current planned height of the robot The function is calculated using the following formula: ,in, The number of electrically charged devices to be sensed. For the first The nominal voltage level of each live device The coordinates of the equipment center are Let the coordinates of the top envelope point of the robot be given in the planned pose. For safe insulation distance, This is an adjustment coefficient; when the distance is less than... hour, Let it be infinity. Potential energy value. Voltage level of live equipment of It is directly proportional to the power of the distance from the top of the robot to the device. The power is inversely proportional. A key feature of this formula is the introduction of the "robot planning height". "As a variable, this means that the map is no longer a static impedance distribution, but a risk field that dynamically changes with the robot's posture (standing or crawling). The technical effect is that when the quadruped robot's height increases due to terrain elevation or extension of the robotic arm, the potential energy cost will increase exponentially, thereby forcing the planning algorithm to automatically avoid or lower its posture in the critical area of ​​electric shock, thus achieving active defense against non-contact breakdown risk."

[0029] In a specific embodiment, the construction of the comprehensive environment map also includes the fusion of semantic attributes: The ground grid is divided into rigid impassable areas, flexible obstacle-crossing areas, and flat passable areas; For flexible obstacle-crossable areas, the cost criterion assigns a higher cost than flat traversable areas but lower than rigid impassable areas, allowing path planning algorithms to plan paths through gravel or shallow pits when no flat path is available. This design fully leverages the superior unstructured terrain adaptability of quadruped robots compared to wheeled robots. At substation renovation and expansion sites, construction often temporarily cuts off flat roads or fills them with gravel. Traditional algorithms treat gravel as obstacles, leading to "no way to go" or planning failure. However, the strategy of this invention allows the algorithm to accept a higher cost penalty in extreme cases where "no flat path is available," deciding to plan a "suboptimal escape path" through gravel or shallow pits. This not only ensures the robot's task accessibility in complex construction sites but also automatically guides the robot back to a good path at a low cost when a flat path exists, achieving a dynamic balance between traversability and motion stability.

[0030] S2: Define the path planning state of the quadruped robot as a combination of position coordinates, heading angle and gait mode; set the cost criterion for global search, which is composed of movement distance cost, spatial safety risk cost and gait mode switching cost.

[0031] In a specific embodiment, Figure 3A flowchart illustrating the calculation logic of gait mode switching cost according to a specific embodiment of this application is shown, as follows: Figure 3 As shown, the calculation logic for gait mode switching cost is as follows: S21: Preset multiple gait mode sets, including at least high posture walking mode, low posture crawling mode and obstacle crossing mode; S22: During the node expansion process of path search, compare the gait mode selected by the current node to be expanded with the gait mode of the parent node; S23: If the gait modes of the two robots are inconsistent, an additional switching penalty value is added to the cumulative cost. The magnitude of the switching penalty value is proportional to the joint adjustment energy consumption and time cost required for the robot to perform gait transitions, in order to suppress the robot from frequently switching motion postures unnecessarily. The gait switching process of a quadruped robot is accompanied by frequent start-stop and acceleration / deceleration of the joint motors, and its instantaneous power is much higher than that of constant speed walking. By introducing a switching cost, the algorithm macroscopically plans a path that can maintain a single efficient gait. Even if the path is slightly longer geometrically, its total energy consumption is lower because the posture adjustment consumption is reduced. At the same time, it suppresses unnecessary frequent "stand-up-lie-down" switching, effectively avoiding body pitch oscillations. This not only ensures the imaging stability of the onboard sensors but also reduces the alternating stress wear of the servo motors and reducers, extending the service life of the hardware.

[0032] S3: In the comprehensive three-dimensional cost map, the optimal path from the starting point to the ending point is searched based on the cost criterion, and the initial planned path containing the position sequence and the corresponding gait mode sequence is output. During the search process, when the planned path enters the high electric potential energy region, the algorithm automatically constrains the gait mode to a low posture mode to increase the safe distance in the vertical direction.

[0033] In a specific embodiment, Figure 4 This document illustrates a flowchart of the triggering and deactivation logic for the low-pose mode in S3 of a specific embodiment of this application, as shown below. Figure 4 As shown, the triggering and deactivation of the low-pose mode adopts a hysteresis comparison strategy, and the specific logic is as follows: Trigger logic S31: When the electric field potential energy value at the robot's planned position exceeds the preset warning threshold, the gait mode in the planned state is immediately locked to a low-profile crawling mode. This setting establishes absolute priority for safety risks, ensuring that the robot can instantly respond by lowering its posture to avoid danger when approaching a hazard source, and establish a safe clearance in the vertical direction without delay.

[0034] Release logic S32: When the electric field potential energy value at the robot's planned position drops below the warning threshold, it does not immediately resume a high posture. Instead, it maintains a low-posture crawling mode to continue planning a preset safe buffer distance. Afterward, the height lock is released, allowing the algorithm to automatically select the gait based on the movement cost. In the complex electromagnetic environment of substations, sensor data or calculated potential energy values ​​often exhibit slight fluctuations. If only a single threshold is used for judgment, the robot is prone to falling into a dead loop of "standing up-lying down-standing up" at the safety boundary due to numerical fluctuations. The spatial lag strategy of this invention effectively filters out this boundary noise, ensuring the monotonicity and continuity of the robot's posture switching, avoiding repeated overshoot oscillations in the control system, thereby significantly improving the stability of inspection operations and motor lifespan.

[0035] S4: Extract key turning nodes in the initial planned path, and perform foot landing suitability verification on straight paths between non-adjacent nodes; only when the terrain flatness of the area covered by the straight path meets the anti-slip requirements of the quadruped robot's feet, remove intermediate redundant nodes and generate the final execution path.

[0036] In a specific embodiment, Figure 5 A flowchart illustrating the foot landing suitability verification process of a specific embodiment of this application is shown, as follows: Figure 5 As shown, the foot landing point suitability check specifically includes: S41: Using the line connecting the two non-adjacent nodes to be verified as the central axis, construct a virtual straight corridor with a certain width. The width is adapted to the lateral standing span of the quadruped robot to cover the landing area of ​​the robot's feet when it actually walks. S42: Extract all surface raster data within the coverage area of ​​the virtual straight corridor, and calculate the comprehensive terrain stability index of this path segment based on elevation dispersion and normal tilt. The specific calculation formula is as follows: ,in: This represents the total number of valid sampling points within the virtual linear corridor. For the first The relative elevation values ​​of each sampling point This represents the average elevation within the corridor. For the first Local fitting plane normal vector at each sampling point; This is the unit vector of gravity in the opposite direction in the world coordinate system (i.e., the vertically upward direction or the positive unit vector of the Z-axis in the world coordinate system); This characterizes the degree of inclination of the Earth's surface. These are preset weighting coefficients used to balance the effects of ruggedness and inclination on foot stability; S43: The calculated result With the preset foot stability threshold Perform a comparison; if If the terrain features of the straight path meet the requirements for full foot contact of the quadruped robot, node removal and path straightening operations are allowed; otherwise, if there is a risk of slipping or falling out, the intermediate nodes in the original path are retained to maintain the detour trajectory.

[0037] In a specific embodiment, when smoothing the initial path, instead of directly detecting whether the connection between non-adjacent nodes encounters obstacles, a virtual straight corridor with a physical width is constructed using this connection as the central axis. The corridor width W is set to fit the lateral standing span of the quadruped robot (e.g., W = 0.5 meters). This design expands path planning from a one-dimensional "geometric line" to a two-dimensional "physical surface," ensuring that the verification range fully covers the possible landing areas of the robot's four legs during actual walking, preventing awkward situations such as "no obstacles on the center line, but the left leg misses or the right leg hits a wall." Terrain Comprehensive Stability Index In the calculation formula, the first term The elevation dispersion (i.e., the standard deviation of point cloud height) is used to identify terrain that, although the average height is uniform, is locally fragmented and rugged, such as piles of gravel, speed bumps, or scattered bricks; this prevents robots from slipping on uneven ground. (Second item) The normal tilt (i.e., the deviation of the dot product between the surface normal vector and the gravity reverse vector) is used to identify terrains that, although smooth, have an excessively large overall slope, such as sloping metal covers or steep slopes; this prevents the robot from sideslipping or becoming unstable due to insufficient ground friction. Weighting coefficients are used to... Combining the two above, only when... Only when the path falls below a safety threshold is it deemed both flat and level, allowing it to be straightened. This mechanism fundamentally eliminates the risks of "building a road into a pit" or "building a road onto a landslide" that may be caused by traditional geometric smoothing algorithms (such as B-splines), significantly enhancing the walking reliability of quadruped robots on unstructured ground in substations.

[0038] In a specific embodiment, the method of this application further includes a dynamic obstacle avoidance step: During the robot's movement along the final execution path, the onboard LiDAR is used to detect dynamic obstacles in the local environment in real time; When a dynamic obstacle is detected encroaching on the current path, a local path planning algorithm (such as the conventional DWA dynamic window method or the TEB time elastic band algorithm) is activated to generate an obstacle avoidance trajectory. When generating obstacle avoidance trajectories, the three-dimensional electric field safety potential energy field constructed in S1 is introduced as a hard constraint into the evaluation function of the local programming algorithm, prohibiting the obstacle avoidance trajectory from bulging upwards or encroaching on the vertical safety clearance area below the powered equipment. This avoidance step prevents the robot from incorrectly planning an "upward-bulging" trajectory (e.g., climbing onto a nearby mound of earth to avoid it, or adjusting its body posture to avoid it) in pursuit of rapid horizontal maneuvering when urgently avoiding dynamic obstacles on the ground. By introducing the hard constraint of the electric field potential energy, it is ensured that even under dynamic obstacle avoidance conditions, the robot's top envelope height is always limited to the vertical safety clearance area below the powered equipment.

[0039] In a specific application example, the above method is illustrated using a circuit breaker inspection task at a 500kV substation renovation and expansion site: At a 500kV substation renovation and expansion project site, the 500kV high-voltage busbar is energized above, with sufficient nominal height relative to the ground. However, a temporary pile of construction waste, approximately 1.5 meters high, is piled up below the construction area. The ground consists of a flat cement road, scattered gravel (flexible obstacle crossing area), and inclined steel plates left over from construction. A quadruped robot needs to start from point A (the entrance to the maintenance room), traverse the aforementioned complex area, and reach point B (the circuit breaker mechanism box) to perform infrared temperature measurement and meter reading. The robot is equipped with a robotic arm with an extension height of 0.5 meters on its back. In a traditional solution, the navigation system of an ordinary robot only sees the pile of dirt on the ground. Its calculations assume: "The pile of dirt is 1.5 meters high; it can be climbed over." Therefore, a shortcut to climb over the pile of dirt is planned. Consequences: When the robot climbs to the top of the mound, it is elevated by 1.5 meters. Combined with its own height and the robotic arm, this encroaches on the 5.0-meter safety clearance of the 500kV busbar above, potentially causing a high-voltage arc to break down the air and resulting in a flashover accident. Solution of this invention: Identify that the voltage of the busbar above is 500kV and the safe insulation distance... =5.0m. Using the aforementioned formula Calculate the potential energy field. The calculations reveal that although the ground mound is geometrically passable, the spatial potential energy of the area increases significantly after adding the mound's height. The reading was off the charts. Decision: The top of the mound was identified as a high-risk area, but to shorten the path, it was necessary to pass through the low-lying area on the side of the mound. The algorithm searched the map for an optimal path. When the path extended to the high-potential-energy critical zone on the side of the mound, a forced locking logic was triggered, and the planning command automatically changed to: in the high-potential-energy critical zone, force a switch to low-posture crawling mode; at this time, the robot actively retracted its limbs, reducing its body height by 30cm, successfully controlling the top envelope within a safe clearance; when the robot moved away from the side of the mound and into flat ground, although the electric field potential energy value had decreased, according to the hysteresis comparison strategy, the robot was ordered to continue crawling forward for 2 meters (a safe buffer distance); this avoided the robot's "standing up-lying down" twitching oscillation at the edge of the mound due to sensor data jitter. On the remaining flat cement road section, walking and jogging modes were compared. Because the jogging mode is faster and more energy-efficient, and to avoid frequent switching, the jogging mode was locked throughout the long straight sections, reducing the additional penalty value for switching. There is a gravel area left over from construction in the middle of the path; taking a straight line is the shortest route. At the end of the path, there is a smooth steel plate leaning against the roadside; it appears flat. A virtual straight corridor was constructed, and the aforementioned formula was used to calculate... For scree beaches: Although the average slope is not large, the elevation dispersion measured by lidar (the first item) Extremely high, judgment If the force exceeds the threshold, straightening is refused, and a detour on a flat road is chosen to avoid twisting an ankle and falling into a hole. For smooth steel plates: although the surface is extremely flat (elevation dispersion is almost 0), the dot product deviation between the normal vector and the gravity reverse vector (the second term) It's very large. (Judgment) If the distance exceeds the threshold, the robot refuses to straighten and continues walking on the adjacent level surface, avoiding a side slip and fall. The quadruped robot successfully completed the inspection task. Throughout the process, it actively crawled, maintaining a safe vertical distance from the 500kV busbar. It only switched gait once in the danger zone, maintaining an efficient trot for the rest of the route, reducing power consumption by approximately 15% compared to traditional methods that frequently switch gaits. It successfully avoided areas of gravel and steel plates that appeared walkable but could actually cause slippage, and no slippage alarms occurred.

[0040] This invention addresses the challenges of vertical electric shock risk and complex terrain adaptation faced by quadruped robots in substation renovation and expansion scenarios by proposing a hierarchical path planning method. First, by constructing a three-dimensional electric field safety potential energy field integrating voltage levels, the robot can proactively sense hidden high-voltage risks and forcibly trigger a low-posture crawling mode in the critical zone, effectively avoiding non-contact breakdown accidents caused by terrain elevation or robotic arm extension. Second, a gait mode switching cost is creatively introduced into the global planning process, suppressing unnecessary frequent "standing-crawl" transitions and achieving an optimal balance between operational energy efficiency and mechanical lifespan. Finally, during the path smoothing stage, a foot landing suitability verification mechanism based on a virtual straight corridor is adopted, utilizing the terrain comprehensive stability index to eliminate "pseudo-shortcuts" with the risk of slipping or misstepping, thereby ensuring the robot's inherent safety, high efficiency, energy saving, and stable walking in high-risk unstructured environments.

[0041] Figure 6 This paper illustrates an architecture diagram of a hierarchical path planning system for a quadruped robot, considering both electrical safety potential energy and gait switching costs, according to a specific embodiment of this application. Figure 6 As shown, the system includes: a comprehensive environment modeling module 601, a state and cost definition module 602, a joint path search module 603, and a path smoothing optimization module 604. The comprehensive environment modeling module 601 is configured to acquire laser point cloud data and digital twin model information of the substation operating area to construct a grid map reflecting the geometric features of the ground. Simultaneously, based on the voltage level and spatial location of the energized equipment in the digital twin model, a three-dimensional electric field safety potential energy field is constructed and mapped onto the space above the grid map, forming a comprehensive three-dimensional cost map that considers both surface traffic impedance and spatial electric shock risk. The state and cost definition module 602 is configured to define the path planning state of the quadruped robot as a combination of position coordinates, heading angle, and gait mode; and to set the cost criteria for global search, which consists of movement distance cost, spatial safety risk cost, and gait mode switching. The system comprises several components: a joint path search module 603, configured to search for the optimal path from start to finish in the integrated 3D cost map based on cost criteria, and output an initial planned path containing a position sequence and corresponding gait mode sequence; during the search process, when the planned path enters a high electric potential energy region, the gait mode is automatically constrained to a low-posture mode to increase the safe distance in the vertical direction; and a path smoothing optimization module 604, configured to extract key turning nodes in the initial planned path and perform foot landing suitability checks on straight paths between non-adjacent nodes; only when the terrain flatness of the area covered by the straight path meets the anti-slip requirements of the quadruped robot's feet are intermediate redundant nodes removed, generating the final execution path. The system can execute the methods shown in the aforementioned figure.

[0042] The following is for reference. Figure 7It shows a schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application. Figure 7 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0043] like Figure 7 As shown, the computer system includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 702 or programs loaded from storage section 708 into random access memory (RAM) 703. RAM 703 also stores various programs and data required for the operation of system 700. CPU 701, ROM 702, and RAM 703 are interconnected via bus 704. Input / output (I / O) interface 705 is also connected to bus 704.

[0044] The following components are connected to I / O interface 705: an input section 706 including a keyboard, mouse, etc.; an output section 707 including a liquid crystal display (LCD) and speakers, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card and a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to I / O interface 705 as needed. A removable medium 711, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 710 as needed so that computer programs read from it can be installed into storage section 708 as needed.

[0045] Specifically, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 709, and / or installed from removable medium 711. When the computer program is executed by central processing unit (CPU) 701, it performs the functions defined in the methods of this application. It should be noted that the computer-readable storage medium of this application can be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. The computer-readable storage medium can 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 of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer 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. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium that can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. Program code contained on a computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0046] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0047] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0048] The modules described in the embodiments of this application can be implemented in software or in hardware.

[0049] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire laser point cloud data and digital twin model information of the substation operating area, and construct a grid map reflecting the geometric features of the ground; simultaneously, based on the voltage level and spatial location of the energized equipment in the digital twin model, construct a three-dimensional electric field safety potential energy field, and map this potential energy field onto the space above the grid map, forming a comprehensive three-dimensional cost map that takes into account both surface traffic impedance and spatial electric shock risk; define the path planning state of the quadruped robot as a combined state including position coordinates, heading angle, and gait mode; and set a cost criterion for global search. The cost is composed of movement distance cost, spatial safety risk cost, and gait mode switching cost. In the comprehensive three-dimensional cost map, the optimal path from the starting point to the ending point is searched based on the cost criterion, and the initial planned path containing the position sequence and the corresponding gait mode sequence is output. During the search process, when the planned path enters the high electric field potential energy region, the algorithm automatically constrains the gait mode to a low posture mode to increase the safety distance in the vertical direction. Key turning nodes in the initial planned path are extracted, and the suitability of foot landing points is checked for straight paths between non-adjacent nodes. Only when the terrain flatness of the area covered by the straight path meets the anti-slip requirements of the quadruped robot's feet, intermediate redundant nodes are eliminated, and the final execution path is generated.

[0050] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A hierarchical path planning method for a quadruped robot that considers electrical safety potential energy and gait switching costs, characterized in that, include: S1: Acquire laser point cloud data and digital twin model information of the substation operation area to construct a grid map reflecting the geometric features of the ground; at the same time, based on the voltage level and spatial location of the live equipment in the digital twin model, construct a three-dimensional electric field safety potential energy field and map this potential energy field onto the space above the grid map to form a comprehensive three-dimensional cost map that takes into account both the surface traffic impedance and the risk of electric shock in space. S2: Define the path planning state of the quadruped robot as a combination of position coordinates, heading angle and gait mode; set the cost criterion for global search, which is composed of movement distance cost, spatial safety risk cost and gait mode switching cost; S3: In the comprehensive three-dimensional cost map, search for the optimal path from the starting point to the ending point based on the cost criterion, and output the initial planned path containing the position sequence and the corresponding gait mode sequence; During the search process, when the planned path enters a region of high electric potential energy, the algorithm automatically constrains the gait mode to a low-pose mode to increase the safe distance in the vertical direction; and S4: Extract key turning nodes in the initial planned path, and perform foot landing suitability verification on straight paths between non-adjacent nodes; only when the terrain flatness of the area covered by the straight path meets the anti-slip requirements of the quadruped robot's feet, remove intermediate redundant nodes and generate the final execution path.

2. The hierarchical path planning method for quadruped robots that considers electrical safety potential energy and gait switching costs according to claim 1, characterized in that, The three-dimensional electric field safety potential energy field in S1 is constructed as follows: Using the geometric center of the energized equipment as the field source, a distribution model is established to show that the potential energy intensity decreases with increasing spatial distance. A safe breakdown restricted area is defined, and the spatial region that is less than the restricted area from the energized equipment is marked as an infinite potential energy region and regarded as an impassable obstacle. The spatial electric shock risk cost in the comprehensive three-dimensional cost map is positively correlated with the robot's planned height. When the robot adopts different time modes at the same horizontal position, different electric potential energy costs are triggered due to the different body heights.

3. The hierarchical path planning method for quadruped robots that considers electrical safety potential energy and gait switching costs according to claim 2, characterized in that, The potential energy value of the three-dimensional electric field safety potential energy field It's about spatial location. And the current planned height of the robot The function is calculated using the following formula: ,in, The number of electrically charged devices to be sensed. For the first The nominal voltage level of each live device The coordinates of the equipment center are Let the coordinates of the top envelope point of the robot be given in the planned pose. For safe insulation distance, This is an adjustment coefficient; when the distance is less than... hour, Let it be infinity.

4. The hierarchical path planning method for quadruped robots considering electrical safety potential energy and gait switching costs according to any one of claims 1-3, characterized in that, In step S1, the construction of the integrated environment map also includes the fusion of semantic attributes: The ground grid is divided into rigid impassable areas, flexible obstacle-crossing areas, and flat passable areas; For flexible obstacle-crossable areas, the cost criterion assigns a travel cost value higher than that of flat traversable areas but lower than that of rigid impassable areas, allowing the path planning algorithm to plan a path through gravel or shallow pit areas when no flat path is available.

5. The hierarchical path planning method for quadruped robots that considers electrical safety potential energy and gait switching costs according to claim 1, characterized in that, The calculation logic for the gait mode switching cost in S2 is as follows: Multiple gait modalities are preset, including at least a high-posture walking mode, a low-posture crawling mode, and an obstacle-crossing mode; During the node expansion process in path search, the gait mode selected by the current node to be expanded is compared with the gait mode of the parent node; If the two gait modes are inconsistent, an additional switching penalty value is added to the cumulative cost; the magnitude of the switching penalty value is proportional to the joint adjustment energy consumption and time cost required for the robot to perform gait transition, so as to suppress the robot from frequently switching motion postures when unnecessary.

6. The hierarchical path planning method for quadruped robots that considers electrical safety potential energy and gait switching costs according to claim 1, characterized in that, The triggering and deactivation of the low-pose mode in S3 adopts a hysteresis comparison strategy, the specific logic of which is as follows: Triggering logic: When the electric potential energy value at the robot's planned position exceeds the preset warning threshold, the gait mode in the planned state is immediately locked to the low-posture crawling mode. Release logic: When the electric field potential energy value at the robot's planned position drops below the warning threshold, it does not immediately resume a high posture, but maintains a low posture crawling mode to continue planning a preset safe buffer distance, and then releases the height lock, allowing the algorithm to automatically select the gait based on the movement cost.

7. The hierarchical path planning method for quadruped robots that considers electrical safety potential energy and gait switching costs according to claim 1, characterized in that, The foot landing suitability check in S4 specifically includes: Using the line connecting two non-adjacent nodes to be verified as the central axis, a virtual straight corridor with a certain width is constructed. The width is adapted to the lateral standing span of the quadruped robot to cover the landing area of ​​the robot's feet when it actually walks. Extract all surface raster data within the coverage area of ​​the virtual straight corridor, and calculate the comprehensive terrain stability index of the path segment based on elevation dispersion and normal tilt. The specific calculation formula is as follows: ,in: The total number of valid sampling points within the virtual linear corridor; For the first The relative elevation values ​​of each sampling point This represents the average elevation within the corridor. For the first Local fitting plane normal vector at each sampling point; The unit vector of gravity in the opposite direction in the world coordinate system; This characterizes the degree of inclination of the Earth's surface. These are preset weighting coefficients used to balance the effects of ruggedness and inclination on foot stability; The calculated results With the preset foot stability threshold Perform a comparison; if If the terrain features of the straight path meet the requirements for full foot contact of the quadruped robot, node removal and path straightening operations are allowed; otherwise, if there is a risk of slipping or falling out, the intermediate nodes in the original path are retained to maintain the detour trajectory.

8. The hierarchical path planning method for quadruped robots that considers electrical safety potential energy and gait switching costs according to claim 1, characterized in that, The method also includes a dynamic obstacle avoidance step: During the robot's movement along the final execution path, the onboard LiDAR is used to detect dynamic obstacles in the local environment in real time; When a dynamic obstacle is detected encroaching on the current path, the local path planning algorithm is activated to generate an obstacle avoidance trajectory. When generating the obstacle avoidance trajectory, the three-dimensional electric field safety potential energy field constructed in S1 is introduced as a hard constraint into the evaluation function of the local programming algorithm to prevent the obstacle avoidance trajectory from rising upwards or encroaching on the vertical safety clearance area below the live equipment.

9. A computer-readable storage medium having one or more computer programs stored thereon, characterized in that, When the one or more computer programs are executed by a computer processor, they perform the method according to any one of claims 1-8.

10. A hierarchical path planning system for a quadruped robot that considers electrical safety potential energy and gait switching costs, characterized in that, include: The integrated environmental modeling module is configured to acquire laser point cloud data and digital twin model information of the substation operation area and construct a grid map that reflects the geometric features of the ground. At the same time, based on the voltage level and spatial location of the live equipment in the digital twin model, a three-dimensional electric field safety potential energy field is constructed and mapped onto the space above the grid map to form a comprehensive three-dimensional cost map that takes into account both the surface traffic impedance and the risk of electric shock in space. The state and cost definition module is configured to define the path planning state of the quadruped robot as a combination of position coordinates, heading angle and gait mode; and to set the cost criterion for global search, which is composed of movement distance cost, spatial safety risk cost and gait mode switching cost. The joint path search module is configured to search for the optimal path from the starting point to the ending point in the integrated three-dimensional cost map based on the cost criterion, and output an initial planned path containing a position sequence and a corresponding gait mode sequence. During the search process, when the planned path enters a region with high electric potential energy, the gait mode is automatically constrained to a low-posture mode to increase the safe distance in the vertical direction. as well as The path smoothing optimization module is configured to extract key turning nodes in the initial planned path and perform foot landing suitability verification on straight paths between non-adjacent nodes. Only when the terrain flatness of the area covered by the straight path meets the anti-slip requirements of the quadruped robot's feet, intermediate redundant nodes are removed, and the final execution path is generated.

Images