A method and system for controlling the gripping of an explosive ordnance disposal robot

Abstract

The application provides a kind of EOD robot gripping control method and system, it is related to robot control technical field, method includes: obtaining the environmental data of EOD area;According to the environmental data, calculate joint damage probability;According to the joint damage probability, determine target contact area;According to the target contact area and the state of EOD robot, judge whether EOD target is accessible;If yes, according to the target contact area, generate target gripping posture;According to the target gripping posture, generate gripping strategy through reinforcement learning model;According to the gripping strategy, generate low-disturbance gripping trajectory;Control the EOD robot according to the gripping strategy, execute the low-disturbance gripping trajectory, reach the EOD target;Through bistable mechanism and coating locking mechanism, the EOD target is gripped.The application can improve the safety, stability, reliability and task execution success rate of gripping operation under complex EOD environment.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a grasping control method and system for a bomb disposal robot. Background Technology

[0002] With the increasing demand for public safety and the growing complexity of explosive ordnance disposal tasks, bomb disposal robots are being widely used because they can replace personnel in high-risk areas to perform tasks such as reconnaissance, identification, transfer, and disposal.

[0003] In actual bomb disposal operations, bomb disposal robots typically need to use robotic arms and end effectors to contact, grasp, and transfer explosives, suspected hazardous materials, or explosive devices. However, explosives are usually characterized by complex structures, uncertain distribution of sensitive components, and numerous environmental hazards, leading to high safety risks and technical challenges for bomb disposal robots during the grasping process. Bomb disposal robot grasping control technology is a crucial foundation for the safe disposal of hazardous targets, and its grasping effectiveness directly affects the success rate of bomb disposal missions and the safety of on-site personnel and equipment.

[0004] However, existing bomb disposal robot grasping methods are mainly based on target geometric features, robotic arm kinematic constraints, and grasping stability. They do not adequately consider explosion risks, fragmentation risks, and hazardous sensitive areas. They lack quantitative assessment of the degree of danger and risk co-optimization mechanisms, making them susceptible to environmental disturbances and collision risks. This results in low grasping safety and reliability, making it difficult to meet the application requirements of high-risk bomb disposal operations. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a grasping control method and system for a bomb disposal robot, which can solve the technical problems of existing bomb disposal robot grasping methods that are mainly based on target geometric features, robotic arm kinematic constraints and grasping stability, and do not adequately consider explosion risk, fragmentation risk and dangerous sensitive areas. They also lack quantitative assessment and risk co-optimization mechanisms for the degree of danger, are easily affected by environmental disturbances and collision risks, resulting in low grasping safety and reliability, and are difficult to meet the application requirements of high-risk bomb disposal operations.

[0006] A first aspect of this invention provides a grasping control method for an explosive ordnance disposal robot, comprising: S1: Obtain environmental data of the bomb disposal area; S2: Calculate the combined damage probability based on the environmental data; S3: Determine the target contact area based on the combined damage probability; S4: Based on the target contact area and the status of the bomb disposal robot, determine whether the bomb disposal target is reachable; if yes, proceed to step S5; otherwise, return to step S1. S5: Generate the target grasping posture based on the target contact area; S6: Based on the target grasping posture, generate a grasping strategy through a reinforcement learning model; S7: Generate a low-disturbance grasping trajectory according to the grasping strategy; S8: Control the bomb disposal robot to execute the low-disturbance grasping trajectory according to the grasping strategy and reach the bomb disposal target; S9: Capture the bomb disposal target through a bistable mechanism and an encapsulation locking mechanism.

[0007] A second aspect of the present invention provides a grasping control system for an explosive ordnance disposal robot, comprising: a processor and a memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the grasping control method for the bomb disposal robot as described in the first aspect.

[0008] A third aspect of the present invention provides a readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the grasping control method for a bomb disposal robot as described in the first aspect.

[0009] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, by acquiring environmental data of the bomb disposal area, calculating the joint damage probability, and determining the target contact area based on the joint damage probability, combined with bomb disposal target reachability analysis, target grasping posture generation, reinforcement learning grasping strategy generation, and low-disturbance grasping trajectory planning, intelligent and safe grasping control of the bomb disposal robot for dangerous targets is realized. At the same time, by completing the stable grasping of dangerous targets through a bistable mechanism and an envelopment locking mechanism, the disturbance and false triggering risk to dangerous targets during the grasping process can be effectively reduced, thereby improving the safety, stability, reliability, and task execution success rate of grasping operations in complex bomb disposal environments. Attached Figure Description

[0010] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0011] Figure 1This is a flowchart illustrating a grasping and control method for a bomb disposal robot provided in an embodiment of the present invention.

[0012] Figure 2 This is a schematic diagram of a risk-constrained low-disturbance grasping trajectory provided in an embodiment of the present invention.

[0013] Figure 3 This is a schematic diagram of a bistable triggering and encapsulation locking mechanical state provided by an embodiment of the present invention.

[0014] Figure 4 This is a schematic diagram of the grasping control system of an explosive ordnance disposal robot provided in an embodiment of the present invention. Detailed Implementation

[0015] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0016] The grasping and control method of the bomb disposal robot provided in this invention will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0017] Reference manual attached Figure 1 The diagram shows a flowchart of a grasping control method for a bomb disposal robot provided in an embodiment of the present invention.

[0018] This invention provides a grasping control method for an explosive ordnance disposal robot, which may include the following steps: S1: Obtain environmental data of the bomb disposal area.

[0019] The environmental data specifically includes: three-dimensional spatial data of the bomb disposal area, obstacle data, bomb disposal target data, explosion hazard source data, and environmental disturbance data.

[0020] Specifically, the bomb disposal robot utilizes LiDAR, depth cameras, visible light cameras, infrared thermal imagers, millimeter-wave radar, inertial measurement units, and environmental monitoring sensors to perform real-time perception and data acquisition of the bomb disposal area, thereby obtaining environmental data for the area. The 3D spatial data of the bomb disposal area is used to characterize the terrain, building structures, and spatial layout information within the area, specifically including point cloud data, depth data, and 3D map data. Obstacle data characterizes information about obstacles within the bomb disposal area that may affect the robot's movement and grasping operations, specifically including the obstacle's location, size, shape, and motion state. Bomb disposal target data characterizes the characteristics of the explosive target to be handled, specifically including the target's location, attitude, size, outline, mass distribution, and surface contact characteristics. Explosion hazard source data characterizes the hazard level of the explosive and the distribution of sensitive areas, specifically including the explosive type, explosive yield, fuse location, detonator location, location of highly sensitive areas, and hazard impact range. Environmental disturbance data is used to characterize the impact of the external environment on the movement and grasping process of the bomb disposal robot, specifically including wind speed, wind direction, ground vibration, temperature, humidity, air pressure, and electromagnetic interference intensity.

[0021] S2: Calculate the combined damage probability based on environmental data.

[0022] It should be noted that, during the grasping process, the choice of grasping position by the bomb disposal robot not only affects the stability of the grasp but also directly impacts the explosion risk after the hazardous target is touched, as well as the safety of the robot and surrounding personnel. Therefore, before determining the target contact area, it is necessary to first establish an explosion risk distribution model of the hazardous target. By calculating the joint damage probability corresponding to each spatial location within the bomb disposal area, the traditional grasping planning based on geometric features and mechanical stability is transformed into a safe grasping planning based on risk field constraints. This allows the determination of the target contact area to simultaneously consider explosion risk, fragmentation risk, and grasping feasibility, thereby achieving low-disturbance, high-safety bomb disposal grasping control.

[0023] In one possible implementation, S2 specifically includes: S201: Based on environmental data, construct a three-dimensional geometric model of the environment, a geometric model of hazardous targets, and a characteristic model of hazardous sources.

[0024] Specifically, spatial registration and 3D reconstruction are performed on the acquired environmental images, depth point clouds, LiDAR data, and scene map data to obtain a 3D geometric model of the environment corresponding to the bomb disposal area. Simultaneously, the outline, size parameters, attitude parameters, and surface structure of the hazardous target are modeled to obtain a geometric model of the hazardous target. Furthermore, based on the type of explosive, mass, TNT equivalent, number of fragments, fragment orientation distribution, and explosion energy parameters corresponding to the hazardous target, a hazard source characteristic model is established, thereby obtaining basic scene data for explosion risk analysis.

[0025] S202: By combining the three-dimensional geometric model of the environment, the geometric model of the hazardous target, and the characteristic model of the hazardous source, a three-dimensional explosion risk scenario model is obtained.

[0026] Specifically, the building structure, wall structure, obstacle structure, and spatial boundary information in the three-dimensional geometric model of the environment are fused with the location, attitude, and size parameters of the hazardous target in the geometric model of the hazardous target. Combined with the explosive energy parameters, fragment parameters, and explosion propagation parameters in the hazard source feature model, a three-dimensional explosion risk scenario model is established, which includes the location of the explosion source, the spatial obstacle structure, the geometric structure of the hazardous target, and the energy propagation relationship. This provides spatial scene constraints for subsequent shock wave propagation analysis and fragment dispersion analysis.

[0027] S203: Based on the three-dimensional explosion risk scenario model, determine the shock wave hazard field and the fragmentation hazard field.

[0028] It should be noted that the shock wave hazard field refers to the distribution of hazard intensity generated by the shock wave formed by the release of detonation energy after an explosion of a hazard source, propagating in a three-dimensional scene space. It is used to characterize the levels of overpressure load and impulse load experienced by various spatial locations within the scene. Its construction process is as follows: based on the location of the hazard source, the equivalent amount of TNT explosive, the environmental geometry, and the distribution of obstacles in the three-dimensional explosion risk scene model, the effective scaled distance from any point in the scene space to the hazard source is calculated, and the corresponding shock wave physical quantities are determined based on the scaled distance, thereby constructing the shock wave hazard field.

[0029] The specific physical quantities of a shock wave are:

[0030]

[0031] in, Indicates proportional distance Z The corresponding physical quantity of the shock wave can be at least one of the following: peak overpressure, duration of barotropic pressure, impulse, or time to arrival of the shock wave. Represents an exponential function. Indicates the first jThe empirical fitting coefficients corresponding to the order-1 fitting polynomial, where ln represents the logarithmic function. Indicates proportional distance. J This represents the highest order of the empirically fitted polynomial. Indicates the effective transmission distance. This indicates the TNT equivalent of the explosive. Indicates the mass of the explosive. This indicates that the explosive is overheating. This indicates that TNT is extremely hot.

[0032] A fragmentation hazard field refers to the spatial hazard density distribution formed in a three-dimensional scene space after the outer shell, internal metal components, or pre-formed fragments of a hazardous target are ejected under the action of an explosion. It is used to characterize the risk level of fragment impact, penetration, and kinetic energy damage at different spatial locations. Its construction process is as follows: Based on the hazard source characteristics, fragment quantity distribution, launch direction distribution, and scene obstacle structure information in the three-dimensional explosion risk scene model, a spatial assessment grid is established. The coverage of representative fragment propagation trajectories in each spatial grid is statistically analyzed, and the fragment hazard density per unit volume is calculated, thereby constructing the fragmentation hazard field.

[0033] The specific hazardous density of fragments is as follows:

[0034] in, Indicates the danger density of fragments. Indicates the trajectory of the fragment. Indicates the trajectory of the fragment. The corresponding probability of occurrence, Indicates the trajectory of the fragment. The corresponding number of fragments, Indicates the trajectory of the fragment. The length of travel within a grid cell. This indicates the volume of a grid cell.

[0035] S204: Calculate the probability of explosion damage and the probability of fragment damage based on the shock wave hazard field and the fragment hazard field.

[0036] It should be noted that the probability of blast damage is used to characterize the degree of tympanic membrane and lung damage caused by the shock wave to the human body, while the probability of fragment damage is used to characterize the degree of impact and injury caused by blast fragments to the human body or target area, thereby obtaining the comprehensive blast hazard level corresponding to each spatial location within the bomb disposal area.

[0037] Specifically, based on the peak overpressure, effective overpressure, effective impulse, and shock wave duration corresponding to each spatial location in the shock wave hazard field, the tympanic membrane injury Probit value and lung injury Probit value at the corresponding spatial location are calculated, and then converted into the corresponding blast injury probability. Simultaneously, based on the fragment trajectory probability, fragment quantity, fragment hit probability, and fragment injury probability corresponding to each spatial location in the fragmentation hazard field, the fragment injury probability at the corresponding spatial location is calculated.

[0038] The specific probability of explosion damage is as follows:

[0039]

[0040]

[0041]

[0042] The specific probability of fragment damage is as follows:

[0043] in, Indicates the first in the bomb disposal area The spatial region or the first The probability of explosion damage at each spatial grid point This represents the cumulative distribution function of the standard normal distribution. Indicates the bomb disposal area r The Middle The first spatial region Probit values ​​for tympanic membrane damage at each spatial grid location Pascal represents the unit of pressure. Indicates the bomb disposal area r The Middle The first spatial region Probit values ​​for lung injury at each spatial grid point Indicates standard atmospheric pressure. Indicates effective overpressure. Indicates effective impulse. Indicates body mass. Indicates the unit of mass. Pascal second is the unit of measurement. Indicates peak overpressure. Indicates the duration of the shock wave baropressure. Indicates the probability of fragment damage. This indicates that the total fragment trajectory and spatial region are multiplied together. Indicates the trajectory of the fragment. The corresponding fragment hit the first The first spatial region The probability of occurrence at each spatial grid point Indicates the trajectory of the fragment. The corresponding fragment hit the first The first spatial region The probability of damage occurring after a spatial grid.

[0044] S205: Calculate the combined damage probability by combining the probability of explosion damage and the probability of fragment damage.

[0045] Specifically, in this embodiment of the invention, for each spatial region or spatial grid within the bomb disposal area, the probability of explosion damage and the probability of fragmentation damage corresponding to that location are obtained. The probability of explosion damage characterizes the likelihood of damage to personnel, robots, or surrounding targets caused by the overpressure and impulse of the shock wave, while the probability of fragmentation damage characterizes the likelihood of explosion fragments hitting and causing damage. Subsequently, the explosion damage event and the fragmentation damage event are treated as two independent or nearly independent hazardous damage events and fused together for calculation. First, the safety probability of the spatial location without explosion shock wave damage and the safety probability of the spatial location without fragmentation damage are calculated separately. Then, the two are combined to obtain the comprehensive safety probability that the spatial location is simultaneously free from both explosion shock wave damage and fragmentation damage. Finally, the result of subtracting the comprehensive safety probability is taken as the joint damage probability corresponding to the spatial location. In other words, the joint damage probability characterizes the comprehensive probability of occurrence of at least one of the following damage effects at the same spatial location: explosion shock wave damage, fragmentation damage, or damage from both. This approach avoids the problem of underestimating risk when evaluating bomb disposal risks using only a single damage type, enabling subsequent decisions on target contact area selection, grasping posture screening, and low-disturbance trajectory planning to be based on a more comprehensive and integrated risk level. When different types of explosives, scene enclosure levels, fragment quantities, or mission safety levels result in varying degrees of importance for blast shockwave damage and fragment damage, corresponding influence weights can be assigned to the blast damage probability and fragment damage probability respectively. These two probabilities can then be weighted and corrected before calculating the overall safety probability, thus obtaining a combined damage probability applicable to different bomb disposal scenarios.

[0046] In this embodiment of the invention, by constructing a shock wave hazard field and a fragmentation hazard field, and further calculating the joint damage probability, a quantitative assessment of the explosion risk in the bomb disposal area is achieved. This enables the grasping plan to no longer rely solely on the target's geometric features and mechanical stability, but to comprehensively consider explosion risk, fragmentation risk, and environmental safety factors. This provides a risk constraint basis for subsequent target contact area selection, grasping posture generation, and low-disturbance trajectory planning, effectively improving the safety, reliability, and decision-making accuracy of the bomb disposal robot's grasping operation.

[0047] S3: Determine the target contact area based on the combined damage probability.

[0048] It should be noted that, in this embodiment of the invention, after obtaining the joint damage probability corresponding to the spatial grid of the bomb disposal area, the joint damage probability is mapped to the candidate contact area on the surface of the hazardous target. Specifically, candidate contact points on the target surface are extracted based on the geometric model of the hazardous target, and a contact influence spatial domain is constructed with each candidate contact point as the center, along its surface normal direction, the robot end effector approach direction, and the gripper envelope range. The joint damage probability corresponding to each spatial grid within the contact influence spatial domain is statistically analyzed, and the local contact risk value of the candidate contact area is calculated using a weighted average, maximum value constraint, or distance decay integral method. The local contact risk value is used to characterize the explosion risk, fragmentation risk, and degree of danger to surrounding personnel and the robot itself that may be caused when the bomb disposal robot performs contact, gripping, or enveloping actions in the candidate contact area.

[0049] In one possible implementation, S3 specifically includes: S301: Based on the combined damage probability, the risk level of the bomb disposal area is classified.

[0050] Specifically, the combined damage probability is used as a risk assessment indicator, and the bomb disposal area is classified according to a preset risk grading rule. When the combined damage probability exceeds a first threshold, the corresponding area is determined to be a high-risk area. When the combined damage probability is between the first and second thresholds, it is determined to be a medium-risk area. When the combined damage probability is below the second threshold, it is determined to be a low-risk area.

[0051] S302: Identify low-disturbance contact areas within the bomb disposal zone based on the risk level classification results.

[0052] Specifically, based on the risk level classification results, low-risk areas are prioritized as candidate contact areas, and an analysis is conducted considering the external surface geometry, curvature distribution, structural stability, and contact feasibility of the hazardous target. Areas with sharp edges, suspended structural parts, and areas prone to deformation under stress are eliminated, retaining areas with surface flatness meeting preset requirements and sufficient contact area. Simultaneously, considering the current posture of the hazardous target and the contact constraints of the robot's end effector, areas capable of stable contact with minimal disturbance to the hazardous target are identified as low-disturbance contact areas.

[0053] S303: Avoid and screen hazardous and sensitive areas within low-disturbance contact zones.

[0054] Specifically, based on the geometric model of the hazardous target and the characteristic model of the hazard source, hazardous sensitive areas on the hazardous target are identified. Hazardous sensitive areas include fuse areas, detonator installation areas, pressure triggering areas, circuit control areas, sensor areas, and other critical structural areas that may be triggered by external forces. The spatial distance and contact impact degree between each low-disturbance contact area and the hazardous sensitive area are calculated. If the distance between the low-disturbance contact area and the hazardous sensitive area is less than a preset safety distance, or if contact with the area may cause a change in the state of the hazardous target, the corresponding area is removed from the candidate areas. Areas that meet the safety distance requirements and will not trigger the sensitive structures of the hazardous target are retained, completing the hazardous sensitive area avoidance screening.

[0055] S304: Generate a set of contact priority areas based on the avoidance screening results.

[0056] Specifically, a comprehensive evaluation is conducted on candidate contact areas after the hazardous and sensitive area avoidance screening. Evaluation indicators include joint damage probability, area stability, contact area, robot operability, and grasping reliability. Priority scores are calculated for each area based on these evaluation indicators, and the areas are ranked according to the scores. Areas with scores higher than a preset threshold are selected as priority contact areas. Multiple priority contact areas that meet these criteria form a priority contact area set, providing candidate contact positions for subsequent grasping posture planning.

[0057] S305: Determine the target contact area based on the set of priority contact areas.

[0058] Specifically, based on the current pose of the bomb disposal robot, the range of motion of the robotic arm, the structural parameters of the end effector, and the target reachability analysis results, the reachability of each candidate region in the set of priority contact regions is verified. Furthermore, a comprehensive evaluation is conducted, incorporating the potential disturbance levels, collision risks, and grasping stability during the robot's movement. The candidate region with the best comprehensive evaluation result is selected as the target contact region, and the corresponding contact position, contact direction, and contact posture parameters are determined, thus providing target constraints for subsequent grasping posture generation and low-disturbance grasping trajectory planning.

[0059] In this embodiment of the invention, by mapping the joint damage probability to candidate contact areas on the surface of a dangerous target, and combining risk level classification, avoidance of dangerous and sensitive areas, structural stability analysis, and robot accessibility assessment, a target contact area that balances safety, stability, and operability can be selected from multiple candidate contact areas, thereby reducing the risk of triggering explosions, contacting sensitive structures, and generating excessive disturbances during the grasping process from the source.

[0060] S4: Based on the target contact area and the status of the bomb disposal robot, determine whether the bomb disposal target is reachable. If yes, proceed to step S5. Otherwise, return to step S1.

[0061] Specifically, the system first acquires the location, normal direction, and risk level of the target contact area, and simultaneously acquires the current position, posture, joint angles of the robotic arm, end effector pose, and motion capability parameters of the bomb disposal robot. Then, based on the location and normal direction of the target contact area, the system performs inverse kinematics calculations on the robotic arm to determine if there is a feasible joint configuration that meets the requirements of joint angle range, end effector posture, and gripper approach direction. Simultaneously, the system calculates the cumulative risk of the approach path from the current position to the target contact area using a spatial risk distribution map, and determines whether this cumulative risk is below a preset risk threshold. Further, based on obstacle data and robot geometry, the system detects whether the robot body, robotic arm, and gripper collide with obstacles or get too close to high-risk sensitive areas during the approach process. If all three conditions are met simultaneously—robotic arm kinematic reachability, path risk below the threshold, and no collision during the approach process—the bomb disposal target is determined to be reachable, and the process proceeds to step S5. If any condition is not met, the system determines that the current environmental state or the target contact area is unreachable, returns to step S1 to reacquire environmental data for the bomb disposal area, and updates subsequent judgments.

[0062] It should be noted that those skilled in the art can set the size of the preset risk threshold according to actual needs, and this invention does not limit this.

[0063] In this embodiment of the invention, by verifying the reachability of the target contact area and the current state of the bomb disposal robot before generating the grasping posture, and by comprehensively considering the kinematic constraints of the robotic arm, the path risk level, and the environmental collision risk, it is possible to identify unexecutable or high-risk grasping schemes in advance and return to re-perceive and plan in a timely manner, thereby avoiding subsequent invalid planning and dangerous operations, and improving the success rate, safety, and overall planning efficiency of the grasping task.

[0064] S5: Generate the target grasping posture based on the target contact area.

[0065] In one possible implementation, S5 specifically includes: S501: Extract the geometric feature parameters of the target contact area.

[0066] Specifically, the geometric feature parameters include the surface normal vector, local curvature, contact area, boundary contour, and spatial position parameters corresponding to the target contact area.

[0067] Specifically, based on the geometric model of the hazardous target, the three-dimensional point cloud data and surface mesh data corresponding to the target contact area are obtained. The surface normal vector and local curvature of the target contact area are calculated through surface fitting. At the same time, the boundary contour information, effective contact area, and spatial coordinate information of the target contact area are extracted, thereby obtaining geometric feature parameters that characterize the spatial morphology and surface properties of the target contact area.

[0068] S502: Determine the contact direction based on geometric characteristic parameters.

[0069] Specifically, the initial approach direction of the end effector is determined based on the surface normal vector corresponding to the target contact area. The initial approach direction is then corrected by considering the current posture of the dangerous target, the position of its center of gravity, and the distribution of surrounding obstacles. At the same time, the constraints of the robotic arm's motion space and the structural constraints of the end effector are taken into account to avoid interference with the dangerous target or environmental obstacles during the approach process, thereby obtaining a contact direction that meets the safety approach requirements.

[0070] S503: Generate multiple candidate grasping postures based on the contact direction.

[0071] Specifically, a local grasping coordinate system is established along the contact direction, using the center of the target contact area as a reference point. Within this coordinate system, the position offset, attitude angle parameters, and gripper opening parameters of the end effector are sampled in combination to generate multiple candidate grasping postures. Each candidate grasping posture corresponds to a spatial pose and gripping configuration of the end effector, thus forming a set of candidate grasping postures.

[0072] S504: Perform collision filtering on each candidate grasping posture to generate the target grasping posture.

[0073] In one possible implementation, S504 specifically includes: S5041: Calculate the minimum safe distance and spatial interference index between the bomb disposal robot and environmental obstacles under each candidate grasping posture.

[0074] Specifically, since the distribution of the robotic arm links, joints, and end effector of the bomb disposal robot varies in space under different candidate grasping postures, it is necessary to assess the collision risk between the robot and environmental obstacles based on the robot's spatial occupancy area corresponding to the candidate grasping posture. To this end, the Euclidean distance between any two points between the robot's spatial occupancy area and each environmental obstacle is calculated, and the minimum value among all distances is determined as the minimum safe distance for the corresponding candidate grasping posture. To avoid the problem that relying solely on geometric distances may fail to reflect the risk level of the hazardous area and the degree of posture adaptation, a spatial interference index is calculated by combining the minimum safe distance, joint damage probability, posture deviation between the end effector and the target contact area, and motion intrusion sensitivity. The minimum safe distance characterizes the collision approach risk, the joint damage probability characterizes the explosion hazard level of the area where the candidate grasping posture is located, the posture deviation characterizes the matching degree between the end effector and the target contact surface, and the motion intrusion sensitivity characterizes the disturbance trend of the robot to the space surrounding the hazardous target during its movement. By coupling the above factors, the spatial interference index is obtained, which can not only reflect whether there is a collision risk in the candidate grasping posture, but also comprehensively reflect the degree of intrusion of the posture into the dangerous area and the potential danger in the execution process.

[0075] The minimum safe distance and spatial interference index are as follows:

[0076]

[0077]

[0078] in, Indicates the first k The minimum safe distance corresponding to each candidate pose, where min represents minimizing. Indicates the first An environmental obstacle, Represents a collection of environmental obstacles. Indicates the first The spatial area occupied by the bomb disposal robot corresponding to each candidate grasping posture, i.e., the robot's spatial envelope model. This represents the sampling points in the robot's spatial envelope model. Indicates sampling points within environmental obstacles. Represents Euclidean distance. Indicates the first k Spatial interferometry index corresponding to each candidate pose This represents the Sigmoid normalization function. Indicates the weight of the safety distance. Indicates the preset safety distance. Represents the combined damage probability weights. Indicates the first The joint damage probability corresponding to the spatial region where each candidate grasping posture is located. Indicates the attitude deviation weight. Indicates the first The attitude deviation between each candidate grasping posture and the target contact area Indicates the sensitivity weight of motion intrusion. Indicates the first Motion intrusion sensitivity corresponding to each candidate grasping posture n k Indicates the first The approach direction unit vector of the end effector under each candidate grasping posture, with the subscript T denoteing transpose. This represents the surface normal vector of the target contact area. This indicates that the bomb disposal robot has moved from its current posture to the [missing information]. Change in end-effector position during each candidate grasping posture This indicates that the bomb disposal robot has moved from its current posture to the [missing information]. The change in end-effector pose during each candidate grasping pose. This represents the stability constant to prevent the denominator from being zero.

[0079] In an embodiment of the present invention, This value is used to measure the importance of the minimum safety clearance, and the preferred value is: . This value is used to measure the importance of the combined damage probability, and the preferred value is: . This value is used to measure the importance of attitude matching, and the preferred value is: . This value is used to measure the importance of motion intrusion sensitivity, and the preferred value is: And satisfy: This is to ensure that the relative weights of the various influencing factors remain consistent.

[0080] According to the The end effector pose parameters, robotic arm joint angle parameters, and gripper opening parameters corresponding to each candidate grasping posture are used to solve the actual spatial positions of each robotic arm link and end effector using a bomb disposal robot kinematic model. Then, the geometric models of each link, joint, and end effector are spatially mapped, and their outer contours are enveloped to obtain the area occupied by the bomb disposal robot in three-dimensional space under the corresponding candidate grasping posture. .

[0081] Attitude deviation Unit vector of approach direction of end effector Surface normal vector of the contact area with the target The angle relationship is calculated from the angle between the two directions. The more consistent the directions of the two directions are, the better. The closer the value is to 0, the better the fit between the grasping posture and the target contact surface. (End-effector position change) and end attitude change Based on the current robot state and the first The position difference and attitude difference between each candidate grasping posture are calculated and further used to calculate the motion intrusion sensitivity. This is used to characterize the degree of intrusion into the surrounding space and the potential disturbance trend during robot movement.

[0082] S5042: Eliminate candidate grasping postures whose minimum safety distance is less than the preset safety distance threshold and whose spatial interference index is greater than the preset interference threshold, and generate the target grasping posture.

[0083] It should be noted that those skilled in the art can set the size of the preset safety distance threshold and the preset interference threshold according to actual needs, and the present invention does not limit them.

[0084] In this embodiment of the invention, by analyzing the geometric features of the target contact area and combining parameters such as the target surface normal vector, local curvature, contact area, and spatial position to generate the target grasping posture, the end effector can approach the bomb disposal target in a more matched and stable manner, effectively reducing impact disturbances and posture deviations during the contact process, improving grasping stability, target coverage effect, and subsequent bistable gripper triggering success rate, thus providing favorable conditions for safely and reliably completing the bomb disposal grasping task.

[0085] S6: Generate a grasping strategy based on the target grasping posture using a reinforcement learning model.

[0086] In one possible implementation, S6 specifically includes: S601: Construct the bomb disposal grasping state space based on the target contact area, target grasping posture, and joint damage probability.

[0087] Specifically, the positional parameters, normal direction parameters, and risk level parameters of the target contact area are used as target state information, while the position, attitude, and gripper configuration of the end effector corresponding to the target grasping posture are used as grasping state information. Simultaneously, spatial risk distribution information corresponding to the joint damage probability is combined to construct the bomb disposal grasping state space. Furthermore, the current pose of the bomb disposal robot, the joint state of the robotic arm, the state of the end effector, and the distribution information of surrounding obstacles are all incorporated into the state space, enabling the reinforcement learning model to comprehensively represent the current bomb disposal grasping environment state and provide state input for subsequent strategy decisions.

[0088] S602: Construct the motion space based on the approach direction, approach speed, attitude correction amount, and gripper control parameters of the bomb disposal robot's end effector.

[0089] Specifically, the adjustment amount of the end effector along the approach direction of the target contact area, the adjustment amount of the approach speed, and the attitude angle correction amount are used as continuous action variables. At the same time, the gripper opening and closing amount, the gripper trigger control parameters, and the contact force adjustment parameters are used as execution action variables, which together constitute the reinforcement learning action space. By defining the controllable behavior of the bomb disposal robot during the grasping process through the action space, the reinforcement learning model can autonomously select the optimal grasping action according to the environmental state.

[0090] S603: Construct the reward function.

[0091] Specifically, a reward function is constructed using joint damage probability, grasping stability, contact impact level, collision risk, and avoidance of hazardous and sensitive areas as evaluation indicators. Positive rewards are given when the bomb disposal robot approaches and grasps areas with low joint damage probability. Positive rewards are also given when grasping stability improves and the gripper's coverage effect is enhanced. Negative penalties are applied when contact impact increases, collision risk increases, or when the robot enters hazardous and sensitive areas. Furthermore, terminal rewards are given based on whether the grasping task is successfully completed, thereby guiding the reinforcement learning model to learn grasping strategies that meet safety and stability requirements.

[0092] S604: Based on the bomb disposal grasping state space, action space, and reward function, the reinforcement learning policy network is iteratively trained to generate a grasping decision model.

[0093] Specifically, the constructed state space is used as the input to the policy network, and the action space is used as the output of the policy network. The parameters of the policy network are updated according to the reward function. During the training process, the policy network is iteratively optimized through multiple rounds by continuously acquiring the environmental state, executing the corresponding actions, and calculating the reward value. This allows the policy network to gradually learn the optimal grasping behavior under different risk environments, ultimately generating a grasping decision model that meets the requirements of low risk, low disturbance, and high stability.

[0094] S605: Based on the grasping decision model, perform strategy reasoning on the current environmental state and generate a grasping strategy.

[0095] Specifically, the current bomb disposal environment state is input into the trained grasping decision model. Through strategy reasoning, the corresponding action output is obtained, and based on the action output, the target approach direction, target approach velocity, attitude correction parameters, and gripper control parameters of the end effector are determined. Furthermore, these parameters are combined to form a grasping strategy, which guides the bomb disposal robot in subsequent low-disturbance grasping trajectory planning and bomb disposal target grasping operations.

[0096] In this embodiment of the invention, by constructing a reinforcement learning state space based on information such as the target contact area, the target grasping posture, and the joint damage probability, and by constructing a reward function based on grasping stability, collision risk, and avoidance of dangerous and sensitive areas, the bomb disposal robot can autonomously learn the optimal grasping decision strategy under different risk environments. This enables the robot to achieve synergistic optimization of grasping safety, stability, and success rate, thereby improving its environmental adaptability and intelligent decision-making ability in complex bomb disposal scenarios.

[0097] Reference manual attached Figure 2 The diagram illustrates a risk-constrained low-disturbance grasping trajectory provided by an embodiment of the present invention.

[0098] like Figure 2 As shown in the figure, the initial pose and the target contact pose are used as trajectory endpoints. P1, P2, and P3 represent the control points used in Bezier trajectory optimization. The spatial risk distribution of the bomb disposal area is represented by a gray-scale risk field and contour lines, and environmental obstacles are also marked. The short path shown by the dashed line has spatial conflicts with obstacles and is considered an unexecutable path. The low-disturbance grasping trajectory shown by the solid line, while meeting the obstacle avoidance requirements, detours over the high-risk area and reaches the target contact pose, demonstrating the relationship between path risk integral, collision risk, and trajectory smoothness in trajectory optimization.

[0099] S7: Generate a low-disturbance grasping trajectory based on the grasping strategy.

[0100] In one possible implementation, S7 specifically includes: S701: Construct a low-disturbance trajectory objective function with the goal of minimizing the path risk integral, contact impact energy, and higher-order motion changes of the trajectory.

[0101] The objective function for a low-disturbance trajectory can be set as follows:

[0102] in, J This represents the objective function for a low-disturbance trajectory. min Indicates minimization. T This represents the total time for trajectory planning. This represents the risk weighting coefficient. This represents the contact impact weighting coefficient. This represents the trajectory smoothing weighting coefficient. This represents the attitude stability weighting coefficient. Indicates time Spatial position of the end effector of the bomb disposal robot. Q ( ) indicates the explosion risk value. This represents the contact force when the end effector comes into contact with the target. This represents the fourth derivative of the terminal trajectory position with respect to time. This represents the rate of change of end-effector attitude. This represents the differential operator.

[0103] It should be noted that, ∈[0.30,0.50], used to measure the importance of path risk, with the preferred value being: =0.40. ∈[0.20,0.35], used to measure the importance of contact impact, with preferred values: =0.25. ∈[0.15,0.30], used to measure the importance of trajectory smoothness, with preferred values: =0.20. ∈[0.10,0.25], used to measure the importance of attitude stability, with preferred values: =0.15.

[0104] For example, when the risk level of a bomb disposal target is high, the risk weighting coefficient can be appropriately increased. The value to be set, for example, setting =0.50、 =0.20、 =0.15、 =0.15, to enhance the trajectory's ability to avoid high-risk areas. When the bomb disposal target is highly sensitive to contact impact, the contact impact weighting coefficient can be appropriately increased. The value to be set, for example, setting =0.35、 =0.35、 =0.15、 =0.15, to reduce the impact disturbance generated when the end effector contacts the target. When the bomb disposal environment is complex and it is necessary to improve trajectory continuity, the trajectory smoothing weight coefficient can be appropriately increased. The value of can be adjusted. When high precision is required in maintaining the end-effector's posture during the grasping process, the posture stability weight coefficient can be appropriately increased. The value of is determined so that the low-disturbance grasping trajectory can adapt to the needs of different bomb disposal missions.

[0105] S702: Set the trajectory terminal contact constraint.

[0106] It should be noted that the trajectory terminal contact constraints include trigger force constraints, safe contact force constraints, contact offset constraints, terminal contact velocity constraints, and attitude matching constraints. Among them, the trigger force constraint is used to ensure that the bistable gripper undergoes a bistable transition; the safe contact force constraint is used to limit contact impact; the contact offset constraint is used to ensure that the bomb disposal target is within the effective coverage triggering area; the terminal contact velocity constraint is used to reduce the disturbance energy during the grasping process; and the attitude matching constraint is used to improve the gripper's ability to cover and lock onto the bomb disposal target.

[0107] S703: Under the constraints of trajectory terminal contact constraints, the trajectory control points are generated using the risk-constrained Bezier optimization algorithm based on the low-disturbance trajectory objective function.

[0108] Specifically, the current end-effector pose of the bomb disposal robot is taken as the trajectory starting point, and the grasping pose corresponding to the target contact area is taken as the trajectory ending point. The grasping trajectory is then parameterized using Bezier curves. Subsequently, using a low-disturbance trajectory objective function as the optimization objective and trajectory end-effector contact constraints as the constraint conditions, the risk field information of the bomb disposal area is mapped to the Bezier control point optimization space. During the control point solution process, higher risk penalty weights are applied to control points located near high-risk areas, causing control points to preferentially distribute towards low-risk areas, thereby generating an initial set of trajectory control points that simultaneously satisfies risk constraints, bistable triggering constraints, and kinematic constraints.

[0109] S704: With the goal of minimizing disturbance energy, the trajectory control points are iteratively optimized to generate a low-disturbance grasping trajectory.

[0110] Specifically, the Bezier trajectory control points are used as optimization variables. An iterative optimization algorithm continuously adjusts the positions of these control points, with the goal of minimizing the disturbance energy during the grasping process. During optimization, factors such as the rate of change of contact force, the higher-order derivative of the trajectory, and the rate of change of attitude are comprehensively considered to continuously reduce the impact and vibration disturbances generated during trajectory execution. Simultaneously, the trajectory is verified in real-time to ensure it meets end-effector contact constraints, risk area avoidance constraints, and robot kinematic constraints. When the objective function converges or reaches the preset number of iterations, the corresponding optimal Bezier trajectory curve is output and used as the low-disturbance grasping trajectory for the bomb disposal robot to perform the grasping task. This ensures that the end effector can approach the bomb disposal target smoothly and with low impact, triggering the bistable gripper to complete the grasping.

[0111] S8: Control the bomb disposal robot to follow the grasping strategy, execute a low-disturbance grasping trajectory, and reach the bomb disposal target.

[0112] Specifically, based on the grasping strategy, the target approach direction, target contact posture, approach speed, and target contact area of ​​the bomb disposal robot's end effector are acquired. Simultaneously, based on the low-disturbance grasping trajectory, the position, posture, and velocity control commands for the end effector at each moment are obtained. Subsequently, the grasping strategy is used as a high-level decision constraint, and the low-disturbance grasping trajectory is used as a motion control reference trajectory to control the coordinated movement of the bomb disposal robot's chassis, robotic arm, and end effector gripper, enabling the end effector to gradually approach the bomb disposal target along the low-disturbance grasping trajectory. During the movement, the current position, posture, and end effector state of the bomb disposal robot are acquired in real time and compared with the trajectory reference state. Motion control commands are dynamically adjusted based on state deviations to ensure the end effector always operates stably along the planned trajectory. When the end effector reaches the target contact area and meets the preset contact posture requirements, precise approach to the bomb disposal target is completed, providing initial contact conditions for subsequent bistable gripper triggering, enveloping contact, and enveloping locking.

[0113] Reference manual attached Figure 3 The diagram illustrates a bistable triggering and encapsulation locking mechanical state according to an embodiment of the present invention.

[0114] like Figure 3 As shown, the horizontal axis represents the grasping process time, and the vertical axis represents the normalized force. The background area corresponds to four stages in sequence: gripper approach, contact establishment, bistable transition, and envelopment locking. In the approach stage, the gripper has not yet formed effective contact with the bomb disposal target, and both the real-time contact force and the holding force are close to zero. After entering the contact establishment stage, the real-time contact force, shown by the solid line, gradually increases as the gripper continues to approach and load. When it reaches the trigger force threshold shown by the dotted line, the gripper satisfies the bistable transition condition and switches from a flat stable state to a coiled stable state. During the transition, the real-time contact force exhibits short-term overshoot and damped oscillations, and then stabilizes near the trigger force threshold, while always remaining below the upper limit of the safe contact force shown by the dotted line, indicating that the gripper has received sufficient external excitation to complete the state transition without generating excessive contact impact. After the bistable transition is completed, the holding force, shown by the dashed line, begins to establish and gradually increases with the increase in the degree of envelopment of the bomb disposal target by the gripper, eventually stabilizing to maintain the target's envelopment locking state. The figure specifically illustrates the continuous mechanical process in which "real-time contact force is used to trigger bistable transitions, and holding force is used to maintain the encapsulation and locking after the transition".

[0115] S9: Captures bomb disposal targets through a bistable mechanism and an encapsulation locking mechanism.

[0116] In one possible implementation, S9 specifically includes: S901: Determine whether the gripper of the bomb disposal robot satisfies the bistable transition condition. If yes, proceed to step S902. Otherwise, adjust the target gripping posture and return to step S6.

[0117] Specifically, this embodiment of the invention first acquires the contact state information between the bomb disposal target and the gripper of the bomb disposal robot, and determines whether the bomb disposal target is located within the effective coverage triggering area of ​​the bistable gripper based on the contact position. Then, it determines the gripper triggering force threshold based on the gripper's structural parameters and compares it with the real-time contact force between the gripper and the bomb disposal target. When the real-time contact force is greater than or equal to the triggering force threshold, it is determined that the bistable gripper satisfies the bistable transition condition. Otherwise, the end effector pose of the bomb disposal robot is adjusted, and the process returns to step S6 to regenerate the gripping strategy.

[0118] In one possible implementation, S901 specifically includes: S9011: Obtain the contact status between the bomb disposal target and the gripper of the bomb disposal robot.

[0119] Specifically, the bomb disposal robot is controlled to gradually approach the bomb disposal target along a low-disturbance grasping trajectory. Force sensors, tactile sensors, and a pose detection module located at the end of the grippers acquire real-time contact state information between the grippers and the target. This contact state information includes the contact position, contact normal direction, contact offset, contact force magnitude, and the relative attitude relationship between the grippers and the target. Subsequently, this contact state information is mapped to the gripper coordinate system to determine the actual contact position of the bomb disposal target within the gripper's workspace.

[0120] S9012: Based on the contact status, determine whether the bomb disposal target is within the effective coverage trigger area of ​​the bomb disposal robot's gripper. If yes, proceed to step S9013. Otherwise, adjust the target gripping posture and return to step S6.

[0121] Specifically, based on the contact position and contact offset, the distance between the contact point of the bomb disposal target and the triggering area at the center of the gripper is calculated and compared with the preset effective coverage triggering area. When the contact point is within the effective coverage triggering area, it indicates that the gripper can form a stable coverage structure after completing the bistable transition, and the process proceeds to step S9013. Otherwise, it is determined that the current contact position cannot guarantee the subsequent coverage locking effect. At this time, the target grasping attitude is adjusted according to the contact offset direction and offset, including replanning the gripper approach direction, contact angle, and target contact point position, and the process returns to step S6 to regenerate the grasping strategy to improve the success rate of subsequent bistable triggering and coverage stability.

[0122] S9013: Determine the trigger force threshold of the explosive ordnance disposal robot gripper based on the structural parameters of the gripper.

[0123] Specifically, the bistable triggering force threshold is determined based on the structural parameters of the bistable gripper. These structural parameters include the curvature of the prestressed elastic band, the spacing of the internal support beams, the spring stiffness, and the preload. The potential energy barrier required for the gripper to transition from a flat stable state to a coiled stable state is calculated using a mechanical model of the bistable gripper, and the corresponding minimum triggering contact force is determined as the triggering force threshold. The triggering force threshold characterizes the minimum external excitation intensity required for the gripper to undergo a bistable transition.

[0124] S9014: Based on the trigger force threshold and the real-time contact force of the bomb disposal robot's gripper, determine whether the bomb disposal robot's gripper satisfies the bistable transition condition. If yes, proceed to step S902. Otherwise, adjust the target gripping posture and return to step S6.

[0125] Specifically, a force sensor at the end of the gripper is used to acquire the contact force between the gripper and the bomb disposal target in real time, and this real-time contact force is compared with a triggering force threshold. When the real-time contact force is greater than or equal to the triggering force threshold, it is determined that the gripper has obtained sufficient external excitation energy to drive the bistable structure to cross the potential energy barrier and undergo a state transition, thus proceeding to step S902. Otherwise, it indicates that the current contact state is insufficient to trigger the bistable gripper to complete the state switch. At this time, based on the difference between the real-time contact force and the triggering force threshold, the approach direction, approach speed, and contact position in the target grasping posture are adjusted, and the process returns to step S6 to regenerate the grasping strategy, so that the gripper can meet the bistable triggering condition in subsequent contact processes.

[0126] S902: Switch the gripper of the bomb disposal robot from a flat stable state to a curled stable state.

[0127] S903: To envelop and contact the bomb disposal target.

[0128] Specifically, after the gripper enters the curled stable state, the gripper automatically bends along the outer contour of the explosive ordnance disposal target and forms a continuous contact area with the target, so that a covering structure is formed between the gripper and the target, improving the gripping stability and reducing local contact stress.

[0129] S904: Based on the holding force of the bomb disposal robot's grippers in the coiled stable state, determine whether the bomb disposal target is covered and locked. If yes, proceed to step S905. Otherwise, adjust the end effector pose of the bomb disposal robot and return to step S6.

[0130] Specifically, the holding force generated by the grippers in their coiled stable state is compared with the disturbance force acting on the bomb disposal target. If the holding force is greater than the external disturbance force, the target is determined to be stably locked inside the grippers. Otherwise, the gripping posture is adjusted and the process returns to step S6.

[0131] S905: Establish a dynamic model for the bomb disposal robot.

[0132] Specifically, the bomb disposal robot chassis, robotic arm, and bistable gripper are treated as a unified dynamic system. The Euler-Lagrange method is used to establish a system dynamic model to describe the relationship between the system state and control input during the gripping process.

[0133] S906: Construct tracking error variables based on the low-disturbance capture trajectory.

[0134] Specifically, the low-disturbance grasping trajectory is taken as the desired trajectory, and the deviation between the current state and the desired state of the bomb disposal robot is calculated to construct a composite error variable.

[0135] S907: Combining dynamic models and tracking error variables, construct an uncertainty compensation model for the grasping process.

[0136] Specifically, the dynamic model is as follows:

[0137]

[0138] The tracking error variables are as follows:

[0139] After establishing the overall dynamic model of the bomb disposal robot and the tracking error variables, the composite error variables are differentiated and substituted into the dynamic model of the bomb disposal robot for transformation. The system mass matrix, Coriolis force term, gravity term, contact impact disturbance, and external environmental disturbance are unified and reduced to the total uncertainty term, thus obtaining the error dynamic equation with the composite error variables as the state:

[0140] Furthermore, utilizing the bounded properties of the mass matrix, Coriolis matrix, gravity term, and external disturbance in the dynamic model, the upper bound of the norm of the total uncertainty term is estimated and transformed into a quadratic polynomial form with respect to the norm of the error vector. Finally, an uncertainty compensation model for the grasping process is constructed.

[0141] in, Represents the mass matrix, Represents the state acceleration vector. Represents the Coriolis force and centrifugal force matrices. Represents the state velocity vector. Represents the gravity term. This represents the external disturbance term. Indicates the control quantity. Indicates position control quantity. Indicates attitude control variables. This indicates the control parameters of the robotic arm joints. This represents the rotation matrix from the bomb disposal robot's coordinate system to the world coordinate system. This indicates the control thrust generated by the propulsion unit. Indicates state tracking error. Represents the state vector. Indicates the desired state. Represents a composite error variable. Indicates the rate of change of state tracking error. Represents the rate of change of the composite error variable. This represents the positive definite gain matrix. Indicates an uncertain term. This represents the upper bound of the constant-valued uncertain term. This represents the coefficient of the first-order error term. Represents the coefficient of the quadratic error term. This represents the error vector.

[0142] S908: Solve the uncertainty compensation model and generate control variables.

[0143] Specifically, the uncertainty compensation model shows that there are unknown uncertainties during the grasping process caused by target contact impact, load changes, and external environmental disturbances. Furthermore, the upper bound of the uncertain term satisfies a quadratic polynomial form. Since the true value of the uncertain term cannot be directly obtained, an uncertainty compensation gain is first constructed using the error vector, where the uncertainty compensation gain characterizes the disturbance compensation intensity required by the system at the current moment. Subsequently, based on the composite error variable... The direction of control action is determined, and a robust compensation term is constructed using the opposite direction of the composite error variable as the control direction, ensuring that the control input always acts in the direction that reduces system error. Simultaneously, to improve system tracking performance, a linear error feedback term is introduced based on the robust compensation term. Using the control gain matrix The error convergence speed is adjusted. Finally, the linear error feedback term and the robust compensation term are superimposed to obtain the control quantity:

[0144]

[0145] in, This represents the gain to compensate for uncertainty. Indicates the upper bound of the constant uncertain term. The estimated value, Represents the coefficient of the first-order error term The estimated value, Represents the coefficient of the quadratic error term The estimated value, This represents the positive definite control gain matrix.

[0146] S909: Based on the control quantity, control the bomb disposal robot to grasp the bomb disposal target while it is covered and locked.

[0147] Specifically, control quantity The components are distributed to the bomb disposal robot's motion platform, robotic arm driver, and end effector to compensate in real time for contact impacts, target load changes, and environmental disturbances generated during the grasping process, ensuring that the grippers maintain a continuous covering and locking state and stably grasp the bomb disposal target.

[0148] In this embodiment of the invention, an adaptive envelopment and grasping mechanism for explosive ordnance disposal targets is achieved through a bistable mechanism and an envelopment and locking mechanism. Combined with a dynamic model, trajectory tracking control, and uncertainty compensation control, the contact impact, load changes, and environmental disturbances generated during the grasping process are compensated in real time, so that the gripper can maintain a stable locked state. This effectively improves the stability, safety, and reliability of the grasping process for dangerous targets and reduces the risk of target detachment, false triggering, and grasping failure.

[0149] Reference manual attached Figure 4 The diagram shows a structural schematic of a gripping control system for a bomb disposal robot provided in an embodiment of the present invention.

[0150] This invention provides a grasping control system 20 for an explosive ordnance disposal robot, comprising: a processor 201 and a memory 202; The memory 202 stores programs or instructions that can run on the processor 201. When the program or instructions are executed by the processor 201, they implement the steps of the above-described grasping control method for the bomb disposal robot and achieve the same technical effect. To avoid repetition, the present invention will not elaborate further.

[0151] It should be understood that the processor 201 in this embodiment of the invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0152] It should also be understood that the memory 202 in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DR RAM).

[0153] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0154] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0155] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0156] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0157] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0158] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0159] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0160] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0161] This invention provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the steps of the above-described explosive ordnance disposal robot grasping and control method, and achieve the same technical effect. To avoid repetition, this invention will not elaborate further.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A grasping control method for a bomb disposal robot, characterized in that, include: S1: Obtain environmental data of the bomb disposal area; S2: Calculate the combined damage probability based on the environmental data; S3: Determine the target contact area based on the combined damage probability; S4: Based on the target contact area and the status of the bomb disposal robot, determine whether the bomb disposal target is reachable; if yes, proceed to step S5; otherwise, return to step S1. S5: Generate the target grasping posture based on the target contact area; S6: Based on the target grasping posture, generate a grasping strategy through a reinforcement learning model; S7: Generate a low-disturbance grasping trajectory according to the grasping strategy; S8: Control the bomb disposal robot to execute the low-disturbance grasping trajectory according to the grasping strategy and reach the bomb disposal target; S9: Capture the bomb disposal target through a bistable mechanism and an encapsulation locking mechanism.

2. The grasping control method for the bomb disposal robot according to claim 1, characterized in that, S2 specifically includes: S201: Based on the environmental data, construct a three-dimensional geometric model of the environment, a geometric model of hazardous targets, and a characteristic model of hazardous sources; S202: Combining the three-dimensional geometric model of the environment, the geometric model of the hazardous target, and the characteristic model of the hazardous source, a three-dimensional explosion risk scenario model is obtained; S203: Based on the aforementioned three-dimensional explosion risk scenario model, determine the shock wave hazard field and the fragmentation hazard field; S204: Calculate the probability of explosion damage and the probability of fragment damage based on the shock wave hazard field and the fragment hazard field; S205: Calculate the combined damage probability by combining the explosion damage probability and the fragment damage probability.

3. The grasping control method for the bomb disposal robot according to claim 1, characterized in that, S3 specifically includes: S301: Based on the combined damage probability, classify the risk level of the bomb disposal area; S302: Based on the risk level classification results, identify the low-disturbance contact areas in the bomb disposal area; S303: Avoid and screen dangerous and sensitive areas within the low-disturbance contact area; S304: Generate a set of contact priority areas based on the avoidance screening results; S305: Determine the target contact area based on the set of contact priority areas.

4. The grasping control method for the bomb disposal robot according to claim 1, characterized in that, S5 specifically includes: S501: Extract the geometric feature parameters of the target contact area; S502: Determine the contact direction based on the geometric feature parameters; S503: Generate multiple candidate grasping postures based on the contact direction; S504: Perform collision screening on each of the candidate grasping postures to generate the target grasping posture.

5. The grasping control method for the bomb disposal robot according to claim 4, characterized in that, Specifically, S504 includes: S5041: Calculate the minimum safe distance and spatial interference index between the bomb disposal robot and environmental obstacles under each of the candidate grasping postures; S5042: Eliminate candidate grasping postures whose minimum safety distance is less than a preset safety distance threshold and whose spatial interference index is greater than a preset interference threshold, and generate the target grasping posture.

6. The grasping control method for the bomb disposal robot according to claim 1, characterized in that, S6 specifically includes: S601: Construct a bomb disposal grasping state space based on the target contact area, the target grasping posture, and the joint damage probability; S602: Construct the motion space based on the approach direction, approach speed, attitude correction amount, and gripper control parameters of the bomb disposal robot's end effector; S603: Construct the reward function; S604: Based on the bomb disposal grasping state space, the action space and the reward function, the reinforcement learning policy network is iteratively trained to generate a grasping decision model; S605: Based on the grasping decision model, perform strategy reasoning on the current environmental state to generate the grasping strategy.

7. The grasping control method for the bomb disposal robot according to claim 1, characterized in that, Specifically, S7 includes: S701: Construct a low-disturbance trajectory objective function with the goal of minimizing the path risk integral, contact impact energy, and higher-order motion changes of the trajectory; S702: Set trajectory terminal contact constraints; S703: Under the constraint of the trajectory terminal contact constraint, the trajectory control points are generated by using the risk constraint Bezier optimization algorithm according to the low-disturbance trajectory objective function; S704: With the goal of minimizing disturbance energy, the trajectory control points are iteratively optimized to generate the low-disturbance grasping trajectory.

8. The grasping control method for the bomb disposal robot according to claim 1, characterized in that, S9 specifically includes: S901: Determine whether the gripper of the bomb disposal robot satisfies the bistable transition condition; if yes, proceed to step S902; otherwise, adjust the target gripping posture and return to step S6. S902: Switch the bomb disposal robot's gripper from a straight, stable state to a curled, stable state; S903: Cover and contact the bomb disposal target; S904: Based on the holding force of the bomb disposal robot's gripper in the curled stable state, determine whether the bomb disposal target is covered and locked; if yes, proceed to step S905; otherwise, adjust the end effector pose of the bomb disposal robot and return to step S6. S905: Establish the dynamic model of the bomb disposal robot; S906: Construct tracking error variables based on the low-disturbance grasping trajectory; S907: Combining the dynamic model and the tracking error variables, construct an uncertainty compensation model for the grasping process; S908: Solve the uncertainty compensation model to generate control variables; S909: According to the control quantity, control the bomb disposal robot to grab the bomb disposal target in the covered and locked state.

9. The grasping control method for the bomb disposal robot according to claim 8, characterized in that, Specifically, S901 includes: S9011: Obtain the contact state between the bomb disposal target and the gripper of the bomb disposal robot; S9012: Based on the contact state, determine whether the bomb disposal target is located within the effective coverage trigger area of ​​the bomb disposal robot's gripper; if yes, proceed to step S9013; otherwise, adjust the target gripping posture and return to step S6. S9013: Determine the trigger force threshold of the explosive ordnance disposal robot gripper based on the structural parameters of the gripper. S9014: Based on the trigger force threshold and the real-time contact force of the bomb disposal robot gripper, determine whether the bomb disposal robot gripper satisfies the bistable transition condition; if yes, proceed to step S902; otherwise, adjust the target grasping posture and return to step S6.

10. A gripping control system for a bomb disposal robot, characterized in that, include: Processor and memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the grasping control method for the bomb disposal robot as described in any one of claims 1 to 9.

Images