A method for lateral flip self-resetting of a tracked serpentine robot

Abstract

The application relates to the technical field of robots, and more particularly to a side-overturning self-resetting method for a tracked snake-shaped robot, which takes the core criterion that the gravity center projection exceeds the support boundary, and clearly defines the side-overturning self-resetting process as four continuous stages of "accumulating momentum, exceeding the boundary, rolling and falling stably". In the early stage of the self-resetting, the body shape is changed through joint bending, part of the mass is offset to the target overturning direction, the support relationship between the body and the ground is changed, one side is gradually unloaded, the support area is gradually reduced, and the system approaches a critical unstable state from a stable state; after the gravity center projection exceeds the support boundary, the robot no longer depends on the continuous output of large torque of the driver to forcibly turn over, but uses the overturning torque formed by gravity to complete subsequent rolling. The application can still maintain high self-resetting capability in a narrow and harsh environment where both sides are limited, and can be adapted to various types of robots.

Description

Technical Field

[0001] This application relates to the field of robotics, and more specifically, to a method for self-resetting a tracked snake robot after it has overturned. Background Technology

[0002] Currently, research on rollover self-reset technology for snake robots mainly revolves around two major directions: "active joint torsion" and "active center of gravity adjustment." In the field of active joint torsion technology, the core of existing technology lies in utilizing the multi-joint, high-degree-of-freedom structural characteristics of snake robots. Through the coordinated movement of the yaw and pitch joints, the robot is driven to generate rotational torque around its own roll axis. The robot uses its onboard inertial measurement unit (IMU) to monitor the roll angle in real time and drives the joints to generate reverse torsional motion through proportional control laws, gradually restoring the rolled-over robot to its right position. However, this method only works in spaces where the robot body can bend sufficiently; it completely fails in scenarios where bending is impossible, such as straight lines or narrow ditches. Furthermore, relying entirely on the IMU to measure the roll angle results in significant sensor noise and vibration, leading to unstable righting. Another typical approach is the automatic reset technology for tracked snake robots. This method is designed for three-section tracked snake robots (snake head track + joints + snake tail track). It achieves self-recovery from tipping over through posture detection and joint center of gravity adjustment. Reset is achieved through a sequence of operations including posture detection, posture judgment, directional righting, and repositioning. However, it currently relies mainly on preset angles and fixed steps, and cannot automatically adjust according to ground hardness, slope, or jamming conditions, nor can it adapt to snake robots with different parameters.

[0003] In the above solutions, the active joint torsion scheme relies on the full bending of the fuselage to generate the return torque, which poses a risk of failure in confined spaces such as straight narrow ditches, vertical pipes, and clamping from both sides. Active joint torsion uses only simple proportional control based on IMU, relying on a single attitude feedback mechanism. When faced with sensor noise, fuselage sway, or ground bumps, it is prone to overcorrection, unstable return, or even secondary rollover. Active joint torsion is only suitable for multi-segment continuously bending serpentine bodies and cannot be adapted to tracked or wheeled composite structures. Active joint torsion lacks the ability to autonomously determine the scenario, cannot determine whether the current conditions for resetting are met, and has no retry or strategy switching logic after failure. While the center of gravity adjustment scheme is designed for tracked structures, the resetting action requires a certain amount of swing space, which can easily lead to joint jamming and inability to shift the center of gravity in narrow passages. The center of gravity adjustment scheme relies on preset angles and fixed step movements, and cannot adjust torque and attitude in real time according to ground hardness, slope, friction, etc. The success rate of resetting drops significantly in muddy, slippery, or rugged terrain, lacking online correction and dynamic decision-making capabilities. The center-of-gravity adjustment scheme is specifically designed for three-section tracked snake robots. Its control parameters, motion flow, and mechanical structure are deeply integrated, making it difficult to directly apply to snake robots with different degrees of freedom, sizes, and drive types. The center-of-gravity adjustment scheme is cumbersome and has a long execution chain; any joint failure or motion deviation will interrupt the entire reset process. The core reason is that existing technologies focus on achieving the "reset capability" function, without considering fault-tolerant design under non-ideal working conditions.

[0004] In summary, the main shortcomings of existing snake robot rollover self-reset technology can be attributed to: weak adaptability to confined spaces, insufficient adaptive control capability, poor versatility across multiple models, and low fault tolerance in the reset process. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing snake robot side-tipping self-reset technology, such as weak adaptability to confined space, insufficient adaptive control capability, and poor versatility, and to provide a side-tipping self-reset method for tracked snake robots, which effectively improves the spatial adaptability of snake robots for side-tipping self-reset, as well as the adaptive control capability and versatility.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for self-reset of a tracked snake robot after tipping over is provided, comprising the following steps: S1. Side-rolling state recognition: Real-time acquisition of robot parameter information, including posture parameters and the angles of each joint, and determination of the robot's side-rolling direction based on the parameter information; S2. Support Boundary Recognition: Based on the robot's geometric parameters and real-time collected parameter information, identify the effective contact points between the robot and the ground; identify the support boundaries based on the effective contact points; calculate the lateral distance between the robot's center of gravity projection and the support boundaries based on the support boundaries. S3. Active bending and accumulating momentum: Control at least one of the robot's movable joints to bend in a preset direction and amplitude, so that the robot gradually changes from its original stable ground-hugging configuration to an asymmetrical arch or offset configuration, so that the center of gravity projection gradually approaches the support boundary. S4. Critical Boundary Control: Based on lateral distance, control the robot to make its center of gravity projection cross the support boundary; S5. Gravity-assisted roll: Controls the robot to reduce the output of the driving torque, so that the robot enters a rolling phase dominated by gravity; S6. Landing and posture recovery: After the robot flips to the other side and forms a new stable contact with the ground, control each moving joint to recover to the traveling posture or standby posture; if stable support is not formed within the predetermined time, repeat steps S1 to S6.

[0007] This invention provides a self-resetting method for a tracked snake robot that overturns. It introduces the concept of energy barrier crossing: in the first half of the reset process, only one potential energy barrier needs to be overcome, converting some electrical energy into gravitational potential energy, raising the center of gravity to near the top of the support edge; after crossing the critical point, the center of gravity begins to descend, the potential energy is released in the opposite direction, and gravity itself forms a flipping torque. By fine-tuning the internal configuration to approach the instability boundary, the required lateral extension space is much smaller than the space required for overall torsional reset, enabling it to maintain efficient self-resetting capability even in harsh environments with narrow terrain and limited lateral movement. This invention achieves decoupling between the control logic that changes the relative relationship between the center of gravity and the support boundary and the robot's specific mechanical degrees of freedom, number of tracks, and drive type, possessing strong generalization ability and adaptable to various types of robots.

[0008] Furthermore, in step S1, the robot control system presets a rollover threshold. When the preset rollover threshold is exceeded and the angular velocity approaches zero within a set time, the robot is determined to be in a static rollover state. Then, based on the parameter information, the three-dimensional spatial linkage configuration of the current robot body is calculated to confirm the rollover direction of the robot.

[0009] Furthermore, step S2 specifically includes: Based on the robot's geometric parameters and real-time collected parameter information, the three-dimensional spatial position of each segment in the absolute gravity coordinate system is calculated. Select the set of lowest points based on their height coordinates in three-dimensional space. Based on the set of lowest points, effective contact points are selected according to the contact / stress state. Determine the support boundary based on the effective contact points; Calculate the signed lateral distance d between the robot's center of gravity projection and the support boundary.

[0010] Furthermore, the criteria for determining whether a robot has tipped over include: when This indicates that the center of gravity projection is still located inside the support boundary, and the robot is in a stable state. when This indicates that the center of gravity projection has reached the support boundary, and the robot is in a critical state of instability. when This indicates that the center of gravity projection has crossed the support boundary, and the gravitational torque will promote the robot to continue to rotate.

[0011] Furthermore, in step S3, during the process of controlling the robot to transform into an asymmetric arch or offset configuration, the mass of some segments is laterally transferred towards the target flipping direction, and the contact distribution between the segments and the ground is changed so that one side of the segments is gradually unloaded, while the contact edge on the other side becomes the main support edge; at the same time, the failed contact points that have been lifted off the ground are eliminated.

[0012] Furthermore, in step S4, when d approaches zero or enters a critical instability state, the target joint is further adjusted in small increments until the center of gravity projection crosses the support boundary, i.e., from... Transform into .

[0013] Furthermore, in step S5, when it is detected that the center of gravity projection has crossed the support boundary, or the rolling angular velocity is consistent with the target flipping direction, the robot is controlled to reduce the continuous hard-top driving torque output, so that the robot enters the gravity-dominated flipping stage.

[0014] The present invention also provides a self-resetting system for a tracked snake robot that overturns, comprising: Side-rolling state recognition module: used to collect robot parameter information in real time, including posture parameters and the angles of each joint, and determine the robot's side-rolling direction based on the parameter information; Support boundary recognition module: used to identify the effective contact points between the robot and the ground based on the robot's geometric parameters and real-time collected parameter information; identify the support boundary based on the effective contact points; and calculate the lateral distance between the robot's center of gravity projection and the support boundary based on the support boundary. Active bending and accumulating module: used to control at least one of the robot's movable joints to bend in a preset direction and amplitude, so that the robot gradually changes from its original stable ground-hugging configuration to an asymmetrical arch or offset configuration, so that the center of gravity projection gradually approaches the support boundary. Critical boundary crossing control module: used to control the robot to make its center of gravity projection cross the support boundary based on the lateral distance; Gravity-assisted roll module: Used to control the robot to reduce the output of driving torque, so that the robot enters the flipping phase dominated by gravity; Landing and posture recovery module: Used to control each moving joint to recover to the traveling posture or standby posture after the robot flips to the other side and forms a new stable contact with the ground.

[0015] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described above.

[0016] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method described above.

[0017] Compared with the prior art, the beneficial effects of the present invention are: This invention discloses a self-resetting method for a tracked snake robot that overturns. By fine-tuning the internal configuration to approach the instability boundary, the required lateral extension space is much smaller than the space required for overall torsional reset, enabling it to maintain efficient self-resetting capability even in harsh environments with narrow terrain and limited lateral space. This invention achieves decoupling between the control logic that changes the relative relationship between the center of gravity and the support boundary and the robot's specific mechanical degrees of freedom, number of tracks, drive type, etc., giving it strong generalization ability and adaptability to various types of robots. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall process of a self-resetting method for a tracked snake robot that has rolled over according to the present invention. Figure 2 This is a schematic diagram illustrating the self-resetting principle of the robot's side-tilt mechanism in this invention. Figure 3 This is a flowchart illustrating a self-resetting method for a tracked snake robot that overturns, as described in one embodiment. Figure 4 This is a schematic diagram illustrating the robot's stable contact with the ground in another embodiment; Figure 5 This is a schematic diagram of a robot actively bending and accumulating energy in one embodiment; Figure 6 This is a schematic diagram of a robot reaching a critical instability state in one embodiment. Detailed Implementation

[0019] The present invention will be further described below with reference to specific embodiments. The accompanying drawings are for illustrative purposes only, representing schematic diagrams rather than actual physical objects, and should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some components in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0020] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0021] Example 1 This embodiment is the first embodiment of a self-resetting method for a tracked snake robot that overturns. The self-resetting method provided in this embodiment uses the projection of the center of gravity crossing the support boundary as the core criterion, and defines the self-resetting process as four consecutive stages: "accumulation of momentum, boundary crossing, rolling, and stabilization." In the early stage of resetting, the robot's shape is changed by bending the joints, causing some mass to shift towards the target flipping direction. Simultaneously, the support relationship between the robot and the ground is altered, gradually unloading one side and gradually reducing the support area, causing the system to approach a critical instability state from a stable state. Once the projection of the center of gravity crosses the support boundary, the robot no longer relies on the continuous output of a large torque from the actuator to forcibly flip; instead, it utilizes the flipping torque formed by gravity to complete the subsequent rolling. This method is essentially a resetting approach that "first creates the conditions for flipping, then uses gravity to complete the flipping," which can reduce the drive load and mechanical impact, and improve the stability of the resetting process.

[0022] In existing technologies, the normal return torque is usually generated directly through large joint twisting, or the posture is adjusted according to preset angles and fixed steps. However, in this embodiment, instead of "directly straightening," the robot's body configuration is first adjusted by moving joints, changing the spatial position and ground contact state of each segment, so that the robot's overall center of gravity projection gradually approaches and eventually crosses the current support boundary. When the center of gravity projection is still inside the support boundary, the system remains stable; when the center of gravity projection crosses the boundary, gravity is converted into a driving force that causes the robot to roll over, thus completing the subsequent roll reset. Therefore, this embodiment grasps the fundamental condition for roll reset to be established, changing the reset control from "action-driven" to "state-driven."

[0023] This embodiment provides a self-resetting method for a tracked snake robot that overturns. It is applicable to a tracked snake robot system consisting of a head module, a tail module, several intermediate segments, multiple movable joints, an attitude sensing unit, a joint angle sensing network, a main control computing unit, and a drive execution unit. The head and tail modules can serve as end-capsulation and motion units, respectively. The intermediate segments are connected in series via multiple movable joints to form a flexible body. The attitude sensing unit is preferably located within the head module and is used to acquire the robot's roll angle, pitch angle, yaw angle, and angular velocity information. Angle sensors or encoders are installed at each movable joint to collect the relative rotation angle between adjacent segments in real time. The main control computing unit is electrically connected to the attitude sensing unit, the joint angle sensing network, and each joint actuator, and is used to complete state recognition, mechanical criterion calculation, reset trajectory generation, and execution control. In one embodiment, the system can consist of a head attitude sensing module, a distributed joint angle sensing network, and a main control computing unit, and information transmission and control closed-loop are achieved through a bus.

[0024] This embodiment provides a method for the self-resetting of a tracked snake robot after tipping over. Figures 1 to 3 As shown, the specific steps include: Step S1. Side rollover state recognition: Collect robot parameter information in real time, including posture parameters and the angles of each joint, and determine the robot's side rollover direction based on the parameter information.

[0025] The main control computing unit reads the roll angle, pitch angle, yaw angle, and corresponding angular velocity information output by the attitude sensing unit (IMU) in real time. Simultaneously, it combines this with the relative rotation angle data between adjacent segments collected by angle sensors or encoders distributed at each moving joint. The system internally presets a rollover threshold (e.g., a roll angle deviating from the normal upright posture by nearly 90°). When the roll angle detected by the IMU exceeds this threshold, and the angular velocity approaches zero within a certain time window, the main control unit determines that the robot is in a static rollover state. This filters out brief bumps or violent shaking that occur when the robot is normally moving on rough terrain, avoiding accidental triggering of the reset procedure. After confirming a rollover, the main control unit uses data from the joint angle sensor network to calculate the current three-dimensional spatial linkage configuration of the robot. This step is not only used to determine the specific direction of the rollover (left or right rollover), but also to confirm whether the robot is in a completely straight state or has already exhibited a bending configuration due to a fall / jamming, thus providing an accurate initial state reference for subsequent active bending and momentum building.

[0026] Step S2. Support Boundary Identification: Based on the robot's geometric parameters and real-time collected parameter information, calculate the three-dimensional spatial position of each segment in the absolute gravity coordinate system; select the set of lowest points based on the height coordinates in the three-dimensional spatial position; based on the set of lowest points, select effective contact points according to the contact / force state; determine the support boundary based on the effective contact points; calculate the signed lateral distance d between the robot's center of gravity projection and the support boundary.

[0027] In this step, using the head module containing the IMU or a specific segment as the reference coordinate system, the main control computing unit calculates the three-dimensional spatial position of each segment's shell in the absolute gravity coordinate system in real time using forward kinematics based on the known robot geometric parameters (length, width, and height of each segment) and the real-time acquired relative angles of each joint. Contact candidate point screening and support surface fitting: In the absolute gravity coordinate system, the set of lowest points on the outer contour of each segment is calculated. These surfaces or line segments with the lowest distance from the direction of gravity are identified as effective contact points that physically interfere with the ground and obstacle edges. Based on this, combined with the robot's geometric parameters and mass distribution model, the signed lateral distance d of the total center of gravity projection relative to the support boundary is calculated, or the equivalent overturning moment Tg is calculated. At this point, in the initial reset phase (when the robot body is relatively straight), the system identifies multiple segments simultaneously touching the ground; the support domain at this time is a relatively broad support boundary formed by these widely distributed contact points.

[0028] Specifically, it includes the following steps: S21. Let the three-dimensional spatial position of the i-th candidate contact point on the robot in the gravity coordinate system be:

[0029] in, This represents the height coordinate of the point in the gravity coordinate system. If The smaller the value, the closer the point is to the ground, and the more likely it is to become a contact point.

[0030] S22. Calculate the minimum height; among all candidate contact points, take the minimum height coordinate:

[0031] in, This indicates the minimum height of the robot in its current configuration.

[0032] S23. Obtain the set of minimum points; considering sensor error, ground unevenness, and model error, introduce a small tolerance. The set of lowest points is then defined as:

[0033] in, This represents all points whose altitude is close to the lowest point. These points are candidate points that may touch the ground.

[0034] S24. Screening effective contact points; not all points in the set of lowest points are actually bearing weight. Therefore, it is necessary to combine the end contact state, ground contact relationship, and segment stress state to screen out the points that actually participate in the support from the set of lowest points.

[0035]

[0036] in, Indicates the effective contact area.

[0037] S25. Determine the support boundary; the actual support area can be obtained from the projection of the effective contact points onto the ground. The outer edge of this support area is the support boundary:

[0038] in, Indicates the support area. Indicates the supporting boundary.

[0039] S26. Calculate the relationship between the centroid projection and the support boundary; Let the robot's center of gravity be:

[0040] After projecting the overall center of gravity onto the supporting plane, calculate its signed lateral distance relative to the supporting boundary:

[0041] in, This represents the coordinates of the center of gravity projected onto the target's flipping direction. This indicates the coordinates of the support boundary in that direction.

[0042] The criteria for determining whether a robot will tip over include: when This indicates that the center of gravity projection is still located inside the support boundary, and the robot is in a stable state. when This indicates that the center of gravity projection has reached the support boundary, and the robot is in a critical state of instability. when This indicates that the center of gravity projection has crossed the support boundary, and the gravitational torque will promote the robot to continue to rotate.

[0043] The corresponding gravitational torque can be expressed as: ;in, For the total mass of the robot, This is the acceleration due to gravity.

[0044] Step S3. Active bending and accumulating momentum: Control at least one movable joint of the robot to bend in a preset direction and amplitude, so that the robot gradually changes from the original stable ground-hugging configuration to an asymmetrical arch or offset configuration, so that the center of gravity projection gradually approaches the support boundary.

[0045] During this process, one or more movable joints are controlled to bend in a preset direction and amplitude, gradually transforming the robot from its original stable, ground-hugging configuration to an asymmetrical arched or offset configuration. This achieves two effects: first, it causes a portion of the segment's mass to be transferred laterally towards the target flipping direction; second, it alters the contact distribution between the segments and the ground, gradually unloading one side of the segments while the contact edge on the other side becomes the primary support boundary. At this point, the main control unit continuously removes failed contact points that have lifted off the ground, updating the edge or a few contact segments that truly bear the weight of the entire robot to the current "narrow support band." Through this step, the system's stability margin gradually decreases, and the projected center of gravity gradually approaches the support boundary.

[0046] Step S4. Critical Boundary Control: Based on the lateral distance, control the robot to make the center of gravity projection cross the support boundary.

[0047] When the main control computing unit determines that d is close to zero or the system enters a critical instability state, it continues to make small incremental adjustments to the target joint; if necessary, it can combine the braking, differential speed, short-time pulse drive, or joint compensation torque of the end track to further promote the lateral displacement of mass and the reduction of the support area, so that the total center of gravity projection crosses the support boundary, that is, from Transform into .

[0048] Step S5. Gravity-assisted roll: Control the robot to reduce the output of the driving torque, so that the robot enters the flipping stage dominated by gravity.

[0049] When the system detects that the center of gravity projection has crossed the support boundary, or that the roll velocity is aligned with the target flipping direction, the controller reduces the continuous hard-drive torque output, allowing the robot to enter a gravity-driven flipping phase. At this point, the energy input from the motors has already pushed the system to the vicinity of the potential energy barrier. After crossing the barrier, the system's center of gravity descends, releasing gravitational potential energy, which can automatically assist in completing most of the flipping process, thereby significantly reducing peak motor load and mechanical impact.

[0050] Step S6. Landing and posture recovery: After the robot flips to the other side and forms a new stable contact with the ground, control each moving joint to recover to the traveling posture or standby posture; if stable support is not formed within the predetermined time, repeat steps S1 to S6.

[0051] Once the robot flips to the other side and forms a new stable contact with the ground, the main control computing unit confirms that the system has re-entered the stable zone based on feedback from the IMU and encoder, and controls each moving joint to return to the traveling posture or standby posture. If stable support is not formed within the predetermined time, the main control computing unit readjusts the joint angles based on the real-time posture error and performs a second charging or retry action.

[0052] This embodiment provides a self-resetting method for a tracked snake robot that overturns. This method is not limited to resetting on flat ground; it can also be applied to conditions such as obstacle edges, trench edges, partial suspension, and discontinuous support surfaces. In obstacle edge conditions, the obstacle edge can be considered a temporary support boundary. By coordinating the configuration of the preceding and following segments, the moment distribution of the segment's center of gravity relative to the edge is changed, thereby achieving overturning. This method focuses on the relative relationship between the "center of gravity and the support boundary," thus exhibiting weak coupling with the specific number of segments, dimensions, and drive type, resulting in good platform mobility.

[0053] The working principle of the self-resetting method for side-tipping of a tracked snake robot provided in this embodiment is as follows: Let the robot be Composed of segments, the first The mass of each segment is Its center of gravity in the direction of target flipping is (coordinates omitted) The total mass of the robot is:

[0054] The robot's overall center of gravity is located in this direction as follows:

[0055] The above formula shows that the position of the robot's overall center of gravity is determined by the mass and spatial position of each segment. When joints bend or the positions of the segments change, Changes have occurred, and the overall center of gravity has shifted. And so it changes.

[0056] Let the current effective contact edge between the robot and the ground be the support boundary, and its position be denoted as . Then the signed lateral distance of the overall center of gravity relative to the supporting boundary can be expressed as:

[0057] in, It reflects the relative positional relationship between the center of gravity projection and the support boundary, and is a key criterion for determining whether a robot can flip over.

[0058] Based on this, the torque generated by gravity relative to the supporting boundary can be approximated as:

[0059] in, This is the acceleration due to gravity.

[0060] Therefore, the robot flipping criterion can be derived: when When the center of gravity is projected inside the support boundary, gravity generates a restoring torque, and the robot is in a stable state and is not easy to flip over. when When the center of gravity projection reaches the support boundary, the system is in a critical instability state. when When the center of gravity projection has crossed the support boundary, the gravitational torque turns into a force that promotes rotation, and the robot will continue to roll under the action of gravity.

[0061] If we write this process in time-varying form, then we have:

[0062] The flipping process can be represented as:

[0063] This indicates that the essence of robot flipping is to change the distance of the center of gravity projection relative to the supporting boundary through configurational changes. It gradually increases from a negative value and eventually becomes a positive value.

[0064] From an energy perspective, let the height of the robot's overall center of gravity be... The gravitational potential energy of the system can then be expressed as:

[0065] During the active bending and accumulating phase, the robot's center of gravity rises, increasing the system's potential energy. Once the projected center of gravity crosses the support boundary, the overall center of gravity lowers, releasing the potential energy and converting it into a force conducive to flipping. Therefore, the robot's flipping process can be summarized as "active accumulating—crossing the boundary—gravity-assisted flipping".

[0066] If we combine the analysis with key segments, let the first segment be... The horizontal distance of each key segment relative to the support boundary is: Then the gravitational moment generated by this segment about the supporting boundary is:

[0067] The overall tilting trend of the system can be seen as the result of the combined gravitational torque of each key segment. As the joints bend, the center of gravity of some segments gradually moves towards the outside of the support boundary, corresponding to... This increases the net torque in the direction of the target flip, ultimately driving the system from a stable state to a critical state and completing the flip.

[0068] In summary, the self-reset method for a tracked snake robot that has rolled over, as provided in this embodiment, has the core advantage of transforming "kinematic rigid adjustment" into "dynamic and energy state control" compared to existing technologies that rely on active joint twisting or fixed center of gravity adjustment sequences. This is manifested in the following four significant positive effects: 1. Low power consumption and low hardware damage from an energy perspective: Existing active torsion technology requires the motor to output extremely high torque to "hard-push" against gravity throughout the reset process. This invention introduces the concept of energy barrier crossing: In the first half of the reset, the motor only needs to do work to overcome a potential energy barrier, converting some electrical energy into gravitational potential energy, raising the center of gravity to near the top of the support edge. After crossing the critical point, the center of gravity begins to descend, the potential energy is released in the opposite direction, and gravity itself forms a flipping torque. This "first build up potential, then use gravity" strategy significantly reduces the peak power consumption of the motor and the local wear of the mechanical structure, extending the robot's service life.

[0069] 2. High Robustness and High Fault Tolerance from a Dynamic Perspective (More Precise Closed-Loop Feedback): Existing center of gravity adjustment schemes rely excessively on preset time-series actions. If a track segment jams, the entire reset process fails. This invention uses the core mechanical formula Tg=mgd for state evaluation. The control system only needs to determine the sign of d: when d<0, the system is in the recovery stability zone; when d=0, the system is in the critical state; when d>0, the system enters the gravity-driven flipping zone. This judgment standard based on absolute physical state (center of gravity relative to the support boundary position) allows the robot to dynamically fine-tune the joint bending degree in real time according to the current ground inclination and friction, completely eliminating the drawbacks of "blind execution" and greatly improving the fault tolerance on complex surfaces such as gravel and slopes.

[0070] 3. High adaptability to confined spaces from a kinematic perspective: Existing active joint torsion schemes require a large swing space for the body to generate a restoring torque, which is prone to failure in narrow pipes or ditches. The core purpose of the "active bending and accumulating" stage in this invention is not to generate a direct rolling torque, but to "laterally shift a portion of the mass" and "change the force distribution between each segment and the ground," so that one side of the body becomes the fulcrum while the other side is gradually unloaded. This method of approaching the instability boundary by fine-tuning the internal configuration requires far less lateral extension space than the space required for overall torsional recovery, enabling it to maintain efficient self-recovery capability even in harsh environments with narrow terrain and confined sides.

[0071] 4. Strong platform versatility due to structural decoupling (easy to port and promote): Most existing technologies are customized for pure snake-shaped continuous joints or specific "head-body-tail" three-segment tracked structures. The method of this invention grasps the essential physical law of rollover recovery—"whether the center of gravity projection still falls within the support area." This control logic of "changing the relative relationship between the center of gravity and the support boundary" is decoupled from the specific number of mechanical degrees of freedom, the number of track segments, and the type of actuator. Therefore, this method has extremely strong generalization ability. It is not only applicable to three-segment tracked snake robots, but can also be smoothly transferred to other multi-jointed, heterogeneous size search and rescue or inspection robot platforms with slight configuration, and has broad industrial application prospects.

[0072] Example 2 This embodiment is a second embodiment of a self-reset method for a tracked snake robot that has overturned. This embodiment is similar to the first embodiment, except that, in this embodiment, it is combined with the attached... Figure 4 To be continued Figure 6 The robot reset process was demonstrated by dividing it into four stages.

[0073] Phase 1: Stable ground contact; such as... Figure 4 As shown, the body is relatively straight at this time, with many contact surfaces and a large support area. The center of gravity projection is steadily placed inside, and gravity provides a restoring effect, so it will not flip over on its own.

[0074] Phase Two: Actively bending and accumulating momentum; such as... Figure 5 As shown, the joints bend segment by segment, transforming the body into an arc shape. This action shifts some mass laterally and alters the force distribution between the segments and the ground, making one side more like a fulcrum while the other gradually unloads. Although it hasn't flipped over yet, the stability margin has already decreased rapidly. Through shape change, it transforms the originally wider support state into a narrower critical support state. Initially, when it's relatively straight, many segments are in contact with the ground simultaneously, resulting in a large support area, making it difficult for the center of gravity to cross the boundary. However, as it bends, the contact between different segments and the ground becomes asymmetrical: some segments are under heavier pressure, some are gradually unloaded, and some contact points may even slightly lift off the ground. As a result, the actual support is no longer the entire bottom surface of the body, but rather a narrow support band closer to one side edge or a few contact segments. With the support area shrinking, it becomes much easier for the center of gravity to cross the boundary.

[0075] Third stage: Reaching critical instability; such as... Figure 6As shown, the system is in its most dangerous and critical state when the center of gravity projection approaches the support boundary. At this point, even a slight lateral shift of the center of gravity, a slight narrowing of the support, or a slight increase in rolling speed will cause the center of gravity projection to cross the boundary. Once it crosses, the direction of gravity changes from "helping it stabilize" to "helping it flip over." The flip is not completely static before and after the critical point, but involves a certain amount of dynamic rolling. The flip occurs relatively quickly, indicating that the motors / joints not only "placed it there" but also provided an angular velocity for continued rolling. In other words, the robot has accumulated a certain amount of kinetic energy near the critical position; once the center of gravity just crosses the support boundary, gravity and this rolling speed are superimposed in the same direction, and the flipping action is completed quickly. If only extremely slow quasi-static movement is used, the system will often get stuck near the critical point; but now it slowly accumulates momentum and then quickly crosses the boundary, resulting in a more decisive action.

[0076] Phase Four: Gravity-Assisted Rolling; After crossing the boundary, gravity itself generates a rolling torque, eliminating the need for complete motor-assisted rotation. The machine then quickly rolls to the other side. Once the new bottom surface contacts the ground, a new support area is formed, and the center of gravity projection returns to this new support area, stabilizing the machine again.

[0077] Example 3 This embodiment is a third embodiment of a self-resetting system for a tracked snake robot that overturns, including: Side-rolling state recognition module: used to collect robot parameter information in real time, including posture parameters and the angles of each joint, and determine the robot's side-rolling direction based on the parameter information; Support boundary recognition module: used to identify the effective contact points between the robot and the ground based on the robot's geometric parameters and real-time collected parameter information; identify the support boundary based on the effective contact points; and calculate the lateral distance between the robot's center of gravity projection and the support boundary based on the support boundary. Active bending and accumulating module: used to control at least one of the robot's movable joints to bend in a preset direction and amplitude, so that the robot gradually changes from its original stable ground-hugging configuration to an asymmetrical arch or offset configuration, so that the center of gravity projection gradually approaches the support boundary. Critical boundary crossing control module: used to control the robot to make its center of gravity projection cross the support boundary based on the lateral distance; Gravity-assisted roll module: Used to control the robot to reduce the output of driving torque, so that the robot enters the flipping phase dominated by gravity; Landing and posture recovery module: Used to control each moving joint to recover to the traveling posture or standby posture after the robot flips to the other side and forms a new stable contact with the ground.

[0078] The self-resetting system for a tracked snake robot that overturns, as provided in this embodiment, is similar to the method described in Embodiment 1 when each module is executed, and will not be repeated here.

[0079] In the specific implementation of the above embodiments, the technical features can be combined in any non-contradictory way. For the sake of brevity, not all possible combinations of the above technical features are described. However, as long as the combination of these technical features is not contradictory, it should be considered to be within the scope of this specification.

[0080] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for self-resetting a tracked snake robot after it has overturned, characterized in that, Includes the following steps: S1. Side-rolling state recognition: Real-time acquisition of robot parameter information, including posture parameters and the angles of each joint, and determination of the robot's side-rolling direction based on the parameter information; S2. Support Boundary Recognition: Based on the robot's geometric parameters and real-time collected parameter information, identify the effective contact points between the robot and the ground; identify the support boundaries based on the effective contact points; The distance between the robot's center of gravity projection and the lateral distance of the support boundary is calculated based on the support boundary. S3. Active bending and accumulating momentum: Control at least one of the robot's movable joints to bend in a preset direction and amplitude, so that the robot gradually changes from its original stable ground-hugging configuration to an asymmetrical arch or offset configuration, so that the center of gravity projection gradually approaches the support boundary. S4. Critical Boundary Control: Based on lateral distance, control the robot to make its center of gravity projection cross the support boundary; S5. Gravity-assisted roll: Controls the robot to reduce the output of the driving torque, so that the robot enters a rolling phase dominated by gravity; S6. Landing and posture recovery: After the robot flips to the other side and forms a new stable contact with the ground, control each moving joint to recover to the traveling posture or standby posture; if stable support is not formed within the predetermined time, repeat steps S1 to S6.

2. The self-resetting method for a tracked snake robot that overturns according to claim 1, characterized in that, In step S1, the robot control system presets a rollover threshold. When the preset rollover threshold is exceeded and the angular velocity approaches zero within a set time, the robot is determined to be in a static rollover state. Then, based on the parameter information, the three-dimensional spatial linkage configuration of the current robot body is calculated to confirm the rollover direction of the robot.

3. The self-resetting method for a tracked snake robot that overturns according to claim 1, characterized in that, Step S2 specifically includes: Based on the robot's geometric parameters and real-time collected parameter information, the three-dimensional spatial position of each segment in the absolute gravity coordinate system is calculated. Select the set of lowest points based on their height coordinates in three-dimensional space. Based on the set of lowest points, effective contact points are selected according to the contact / stress state. Determine the support boundary based on the effective contact points; Calculate the signed lateral distance d between the robot's center of gravity projection and the support boundary.

4. The self-resetting method for a tracked snake robot that overturns according to claim 3, characterized in that, The criteria for determining whether a robot has tipped over include: when This indicates that the center of gravity projection is still located inside the support boundary, and the robot is in a stable state. when This indicates that the center of gravity projection has reached the support boundary, and the robot is in a critical state of instability. when This indicates that the center of gravity projection has crossed the support boundary, and the gravitational torque will promote the robot to continue to rotate.

5. The self-resetting method for a tracked snake robot that overturns according to claim 4, characterized in that, In step S3, during the process of controlling the robot to transform into an asymmetric arch or offset configuration, the mass of some segments is laterally transferred towards the target flipping direction, and the contact distribution between the segments and the ground is changed so that one side of the segments is gradually unloaded, while the contact edge on the other side becomes the main support edge; at the same time, the failed contact points that have been lifted off the ground are eliminated.

6. The self-resetting method for a tracked snake robot after tipping over according to claim 4, characterized in that, In step S4, when d approaches zero or enters a critical instability state, the target joint is further adjusted in small increments until the center of gravity projection crosses the support boundary, i.e., from... Transform into .

7. The self-resetting method for a tracked snake robot after tipping over according to claim 6, characterized in that, In step S5, when it is detected that the center of gravity projection has crossed the support boundary, or the rolling angular velocity is consistent with the target flipping direction, the robot is controlled to reduce the continuous hard-top driving torque output, so that the robot enters the gravity-dominated flipping stage.

8. A self-resetting system for a tracked snake-like robot that overturns, characterized in that, include: Side-rolling state recognition module: used to collect robot parameter information in real time, including posture parameters and the angles of each joint, and determine the robot's side-rolling direction based on the parameter information; Support boundary recognition module: used to identify the effective contact points between the robot and the ground based on the robot's geometric parameters and real-time collected parameter information; and to identify the support boundary based on the effective contact points. The distance between the robot's center of gravity projection and the lateral distance of the support boundary is calculated based on the support boundary. Active bending and energy storage module: used to control at least one of the robot's movable joints to bend in a preset direction and amplitude, so that the robot gradually changes from its original stable ground-hugging configuration to an asymmetrical arch or offset configuration, so that the center of gravity projection gradually approaches the support boundary. Critical boundary crossing control module: used to control the robot to make its center of gravity projection cross the support boundary based on the lateral distance; Gravity-assisted roll module: Used to control the robot to reduce the output of driving torque, so that the robot enters the flipping phase dominated by gravity; Landing and posture recovery module: Used to control each moving joint to recover to the traveling posture or standby posture after the robot flips to the other side and forms a new stable contact with the ground.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images