A hybrid control method of a six-bar tensile robot for mars surface exploration

Abstract

The application provides a hybrid control method for a six-bar tensegrity robot for Mars surface exploration, which firstly models the six-bar tensegrity robot used, defines the coordinates and adjacency relationship of each face, and pre-establishes an energy landscape graph. Then in the process of controlling the six-bar tensegrity robot, the environment data is synchronously collected, and the rolling or jumping mode is switched according to the robot state and the environment data. In the rolling mode, the A algorithm is used to search for an energy-optimal face conversion sequence, and based on the adaptive thrust of the quadratic programming, the optimal motor thrust distribution is obtained by solving a constrained optimization problem for each face conversion in the face conversion sequence. In the jumping mode, a safe margin height is reserved, the size and angle of the motor thrust are adjusted to overcome the Mars gravity, and the terrain adaptability of the six-bar tensegrity robot is improved. The combination of the two solves the problems of low marching efficiency, high energy loss and weak maneuverability of the prior art on the complex surface of Mars.

Description

Technical Field

[0001] This invention relates to the field of robot motion control technology, and in particular to a hybrid control method for a six-bar tensioning robot designed for Mars surface exploration. Background Technology

[0002] Exploring the surface of Mars is one of the most challenging frontiers in the field of space robotics. The Martian environment, characterized by low gravity (approximately 3.72 m / s²), abrasive dust, significant temperature variations, and highly irregular terrain, places stringent demands on the mobility of exploration robots. While traditional wheeled Mars rovers have achieved success in several missions, they exhibit significant limitations in areas such as obstacle encounters, soft soil subsidence, and the risk of malfunctions due to mechanical complexity.

[0003] Tensorized monolithic structures, as a novel structural form, consist of discontinuous compressive elements (rods) and continuous tension elements (cables), possessing excellent strength-to-weight ratio, inherent flexibility, and the ability to withstand large deformations without structural failure. NASA's Super Ball Bot project has demonstrated the potential of six-bar tensioned robots for planetary landing and motion.

[0004] In recent years, propeller-based drive systems have been introduced into tensioned monolithic structures, enabling active control of rolling dynamics and expanding motion capabilities. However, existing methods only consider control under simple environmental conditions, and have the following limitations on the rugged and complex surfaces of planets:

[0005] 1. Low thrust distribution efficiency: Existing methods use heuristic rules to distribute thrust among multiple propellers, resulting in low energy efficiency after thrust distribution and limited robot mobility.

[0006] 2. Gait planning ignores energy: The selection of rolling gait sequences is usually based solely on geometric considerations, ignoring the significant energy differences between different surface transitions, which limits motion control in complex environments.

[0007] 3. Most designs focus on a single rolling motion pattern, limiting their ability to cross obstacles and adapt to terrain, making them unsuitable for the complex environments of planetary surfaces.

[0008] 4. Simple trajectory tracking control: Path tracking relies on basic control strategies, such as open-loop or simple feedback configurations, which cannot guarantee performance under the uncertainty of the Martian environment.

[0009] Therefore, there is an urgent need to develop a control method for a six-bar tensioning robot that can enable the six-bar tensioning robot to have high energy efficiency, high mobility and strong robustness in complex planetary environments, so as to meet the needs of Mars surface exploration. Summary of the Invention

[0010] To address the above issues, this invention proposes a hybrid control method for a six-bar tensioning robot designed for Mars surface exploration. By employing adaptive thrust allocation optimization based on quadratic programming, energy-sensing gait planning, and a hybrid rolling-jumping motion strategy, it achieves efficient and energy-saving movement on the Martian surface.

[0011] This invention is achieved through the following technical solution. On one hand, this invention provides a hybrid control method for a six-bar tensioning robot for Mars surface exploration, which includes: S10: initializing the six-bar tensioning robot model and acquiring six-bar tensioning robot data to update the state information of the six-bar tensioning robot; wherein, the information of the six-bar tensioning robot model includes the coordinates of each node of the six-bar tensioning robot, closed triangle faces, open triangle faces, and the adjacency relationship diagram between faces; S20: Based on whether the state information and environmental information of the six-bar tensioning robot meet the first condition, if they do, the motion decision mode is rolling; otherwise, it is jumping. When the motion decision mode is scrolling, steps S31A and S31B are executed, including: S31A: Based on the current ground contact surface, current position coordinates, target position coordinates, and an energy landscape map of the six-bar tensioning robot, perform energy-sensing-based gait planning, generate the energy-optimal surface transformation sequence, and obtain a reference trajectory that includes rolling surface transformation.

[0012] S31B: Based on the reference trajectory and the current ground contact surface, the adaptive thrust allocation optimization is performed to obtain the optimal thrust vector, and the six-bar tensioning robot is controlled to roll according to the optimal thrust vector; When the motion decision mode is jumping, steps S32A and S32B are executed, including: S32A: Calculate the launch speed and angle based on the current position coordinates of the six-bar tensioning robot and the obstacle height in the environmental information; S32B: Thrust is applied using a trapezoidal thrust profile based on the launch velocity and angle.

[0013] Furthermore, the optimal thrust vector in step S31B is obtained by the following formula:

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020]

[0021] in, The unit direction vector of the roll axis. Let be the position vector of the first node of the shared edge in the world coordinate system. The position vector of the second node of the shared edge in the world coordinate system; Using the world coordinate system, This is the gravitational torque; Let be the vector from the endpoint of the rolling edge to the robot's center of mass; The rolling edge vector; For the first i The torque coefficient of each motor; This is the rotation matrix from the body coordinate system to the world coordinate system; For the first i The position vector of each motor in the machine body coordinate system; The world coordinates of the reference point on the roll axis; is the unit z-axis vector in the body coordinate system; Let W be the thrust vector of the four motors; W be the energy weighting matrix; and c be the linear energy loss vector. This is the torque margin coefficient. This is the total thrust limiting factor; For Martian gravitational acceleration, This represents the maximum thrust of the motor.

[0022] Furthermore, after step S31B or step S32B, the method further includes: S40: Obtain the reference trajectory and the actual trajectory of the six-bar tensioning robot in the preset control time domain, construct the MPC optimization problem by combining the preset state weight matrix and control weight matrix, and output the optimal face selection decision for the next step according to the constraints.

[0023] Furthermore, the MPC optimization problem is constructed using the following formula:

[0024]

[0025]

[0026]

[0027] in, For prediction in the time domain; for The actual state under the serial number includes coordinate information and direction; For reference trajectory; To make a decision based on the surface; To control the time domain; This is the state weight matrix; To control the weight matrix; This is the set of adjacent faces that are reachable in the current state; This is the set of state constraints.

[0028] Furthermore, if convergence fails after the MPC optimization solution in step S40, the state constraint set is updated using the following formula:

[0029] in, For constraint coefficients, This is the set of state constraints.

[0030] Furthermore, the energy landscape map is constructed through the following steps: Calculate the thrust energy loss, potential energy loss, and friction loss when each surface rolls to the corresponding adjacent surface to obtain the corresponding surface transition energy loss. An energy landscape map is constructed based on the surface transformation relationship and the corresponding surface transformation energy loss.

[0031] Furthermore, the formula for calculating the emission velocity in step S32A is as follows:

[0032] in, For Martian gravitational acceleration, The height of the obstacle. For safety margin height, For horizontal crossing distance.

[0033] Furthermore, after applying thrust using a trapezoidal thrust profile in step S32B, the method further includes: The robot's acceleration before touching the ground is obtained. If the robot's acceleration exceeds a preset threshold, the motor is controlled to apply a damping thrust.

[0034] Furthermore, the optimal surface transformation sequence is improved by A The algorithm searches within the energy landscape map.

[0035] On the other hand, the present invention also provides a hybrid control system for a six-bar tensioning robot, comprising: Motion decision unit: Used to determine whether the state information and environmental information of the six-bar tensioning robot meet the first condition. If the condition is met, the motion decision mode is rolling; otherwise, it is jumping. The rolling control unit includes a rolling path optimization subunit: used to perform energy-sensing-based gait planning based on the current ground contact surface, current position coordinates, target position coordinates, and an energy landscape map of the six-bar tensioning robot, generate an energy-optimal surface transformation sequence, and obtain a reference trajectory including rolling surface transformation.

[0036] Rolling thrust optimization subunit: used to perform adaptive thrust allocation optimization based on the reference trajectory and the current ground contact surface to obtain the optimal thrust vector, and control the six-bar tensioning robot to roll according to the optimal thrust vector; The jump control unit includes a launch parameter calculation subunit: used to calculate the launch speed and angle based on the current position coordinates of the six-bar tensioning robot and the obstacle height in the environmental information; Skip thrust optimization subunit: used to apply thrust using a trapezoidal thrust profile based on the launch velocity and angle.

[0037] This invention proposes a hybrid control method for a six-bar tensioning robot designed for Martian surface exploration. First, the six-bar tensioning robot is modeled, defining the coordinates and adjacency relationships of each face, and an energy landscape map is pre-established. Then, during the control of the six-bar tensioning robot, environmental data is collected synchronously, and rolling or jumping modes are switched based on the robot's state and the environmental data. In rolling mode, A... The algorithm searches for the energy-optimal surface transition sequence and, based on adaptive thrust using quadratic programming, obtains the optimal motor thrust allocation for each surface transition in the sequence by solving a constraint optimization problem. In jump mode, a safe height margin is maintained, and the magnitude and angle of the motor thrust are adjusted to overcome Martian gravity, improving the terrain adaptability of the six-bar tensioning robot. This combination solves the problems of low travel efficiency, high energy loss, and weak maneuverability of existing technologies on the complex Martian surface. Furthermore, this invention introduces MPC trajectory tracking, which re-optimizes the reference trajectory based on the error between the reference trajectory and the actual trajectory, as well as the updated robot state, achieving error control during the control process. Constraint coefficients are introduced to correct the trajectory using the state constraint set, thereby overcoming the uncertainties brought by the Martian surface environment and enhancing the robustness of the hybrid control system of the six-bar tensioning robot.

[0038] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0039] Figure 1 This is a mechanical structure diagram of an exemplary six-bar tensioning robot according to the present invention; Figure 2 This is a structural block diagram of the hybrid control system of the six-bar tensioning robot of the present invention; Figure 3 For the implementation of this invention Figure 1 A flowchart of the hybrid control method for the hybrid control system shown; Figure 4 This is a structural block diagram of a preferred hybrid control system for a six-bar tensioning robot according to the present invention; Figure 5 For the implementation of this invention Figure 4 The flowchart shows the hybrid control method of the hybrid control system. Detailed Implementation

[0040] The six-bar tensioning robot used in this embodiment of the invention, such as Figure 1 As shown, the mechanical structure of this six-bar tensioning robot is a six-bar integrated tensioning structure, comprising 12 nodes, 6 pressure bars (0.5 m in length), and 24 cables, with a total mass of 1.27 kg. The drive unit includes four brushless DC motors, each with a maximum thrust of 5.0 N, symmetrically arranged inside the mechanical structure. The sensing unit includes various sensing components: an inertial measurement unit (IMU) for attitude detection, an accelerometer for contact detection, and a terrain sensor for obstacle detection. These components form the overall structural framework of the six-bar tensioning robot.

[0041] Based on the aforementioned six-bar tensioning robot, this invention provides a hybrid control system for a six-bar tensioning robot designed for Mars surface exploration. Please refer to... Figure 2 The hybrid system includes a robot state acquisition unit 10, a motion decision unit 20, a rolling control unit 31, and a jumping control unit 32. The robot state acquisition unit 10 initializes the model of the six-bar tensioning robot before task execution. Subsequently, the motion decision unit selects a motion strategy based on the robot's state and environmental information, choosing either the rolling control unit 31 to perform the rolling task or the jumping control unit 32 to perform the jumping task, thus improving the robot's travel efficiency. The rolling control unit 31 performs rolling path planning based on an optimal energy loss strategy and combines it with adaptive thrust allocation through quadratic planning, significantly reducing energy loss in the Martian environment.

[0042] Please see Figure 3 The operation of each component of the hybrid control system for a six-bar tensioning robot designed for Mars surface exploration includes: The robot state acquisition unit 10 is used to execute step S10: initialize the six-bar tensioning robot model and acquire the six-bar tensioning robot data to update the state information of the six-bar tensioning robot; wherein, the information of the six-bar tensioning robot model includes the coordinates of each node of the six-bar tensioning robot, the closed triangle face, the open triangle face, and the adjacency relationship diagram between the faces; The initialization process requires establishing a structural model of the six-bar tensioning robot, including defining the coordinates of the 12 nodes as follows:

[0043]

[0044]

[0045]

[0046]

[0047]

[0048] The six compression strut connection nodes are defined as: (1, 2), (3, 4), (5, 6), (7, 8), (9, 10), (11, 12). Twenty-four cables form a tension network to maintain structural integrity. Four brushless DC motors and a propeller are symmetrically installed inside the structure. The coordinates of each brushless DC motor are:

[0049]

[0050] Eight closed triangular faces C1-C8 and twelve open triangular faces O1-O12 are defined as possible ground contact surfaces. Face transitions can be categorized into four gait types: CO gait (from closed to open triangle), OC gait (from open to closed triangle), CC gait (transition between closed triangles), and OO gait (transition between open triangles). After defining face transitions, an adjacency graph is established between faces. Two faces are considered adjacent if they share exactly two nodes. The final adjacency graph contains 20 vertices and 30 edges.

[0051] Constructing a six-bar tensioning robot structural model is equivalent to system initialization, providing a foundation for subsequent robot motion control. An energy landscape diagram is constructed based on the surface transformation relationship and the corresponding surface transformation energy loss. It is constructed through the following steps: Calculate the thrust energy loss, potential energy loss, and friction loss when each surface rolls to the corresponding adjacent surface to obtain the corresponding surface transition energy loss.

[0052] The formula for calculating energy loss during surface-to-surface conversion is as follows:

[0053]

[0054]

[0055]

[0056]

[0057] in, For noodles Convert to surface Total energy consumption For thrust energy consumption, This is the change in gravitational potential energy. For frictional loss, The energy conversion coefficient of the motor. Let be the optimal thrust value for the k-th motor. The duration of a single scroll. For the total mass of the robot, For Martian gravitational acceleration, For noodles The height of the center of mass.

[0058] The above initialization process is executed only once during a single task. Subsequently, the robot state acquisition unit corrects the coordinates of each node based on the initialized model to update the robot's state information.

[0059] The motion decision unit 20 is used to execute step S20: whether the state information and environmental information of the six-bar tensioning robot meet the first condition. If they do, the motion decision mode is rolling; otherwise, it is jumping.

[0060] The selection logic for the motion decision-making model is as follows:

[0061] Where mode is the selected motion mode. The height of the obstacle. The obstacle threshold; Battery energy; Minimum reserves.

[0062] That is, motion mode decisions are made based on the terrain detection results detected by the sensing unit and the energy state of the six-bar tensioning robot, when the obstacle height... Less than the threshold And battery energy Greater than minimum reserves It uses a rolling mode when the obstacle height exceeds the threshold or the terrain is impassable. It switches to jump mode when the obstacle height exceeds the threshold or the terrain is impassable.

[0063] When the motion mode is rolling, the rolling control unit 31 is activated, wherein the rolling path optimization subunit 31A is used to execute step S31A: based on the current ground contact surface, current position coordinates, target position coordinates and an energy landscape map of the six-bar tensioning robot, perform energy-sensing-based gait planning, generate the energy-optimal surface transformation sequence, and obtain a reference trajectory including rolling surface transformation.

[0064] The optimal surface transformation sequence is achieved through an improved A... The algorithm searches within the energy landscape map. The calculation formula is as follows:

[0065] Where S is the candidate gait sequence, To start from the initial surface The set of all feasible paths to the target region G. This represents the energy loss during surface transformation. A reference trajectory can be fitted based on the surface transformation sequence. The gait planning output contains a complete surface transformation sequence from the current surface to the target region. = { → →···→ },in ∈G.

[0066] Rolling thrust optimization subunit 31B: used to execute step S31B: to perform adaptive thrust allocation optimization based on the reference trajectory and the current ground contact surface to obtain the optimal thrust vector, and to control the six-bar tensioning robot to roll according to the optimal thrust vector.

[0067] Based on the surface transition sequence generated by gait planning, thrust allocation optimization is performed for each transition step. This is done from the current surface... Scroll to the adjacent face of the target For example, its optimal thrust vector is obtained by the following formula:

[0068]

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075] in, The unit direction vector of the roll axis. Let be the position vector of the first node of the shared edge in the world coordinate system. The position vector of the second node of the shared edge in the world coordinate system; Using the world coordinate system, This is the gravitational torque; Let be the vector from the endpoint of the rolling edge to the robot's center of mass; The rolling edge vector; For the first i The torque coefficient of each motor; This is the rotation matrix from the body coordinate system to the world coordinate system; For the first i The position vector of each motor in the machine body coordinate system; The world coordinates of the reference point on the roll axis; is the unit z-axis vector in the body coordinate system; Let W be the thrust vector of the four motors; W be the energy weighting matrix; and c be the linear energy loss vector. This is the torque margin coefficient. This is the total thrust limiting factor; For Martian gravitational acceleration, This represents the maximum thrust of the motor.

[0076] After solving for the reference trajectory, environmental parameters can be combined to fit the interaction between the robot and the environment during the flipping process, such as the coordinate transformation of the roll axis nodes and the transformation of the robot's center of mass coordinates after the flip. This invention transforms environmental influences into torque coefficients for each motor, enabling a synergistic effect between the motor thrusts during each flip, thus avoiding the problem of low energy efficiency caused by the distribution of thrust among multiple propellers in existing heuristic rules.

[0077] Furthermore, when quadratic programming is not feasible, a relaxation strategy is adopted to solve for thrust allocation, specifically including:

[0078] st ≥0 in, For slack variables ，A certain degree of violation of the moment constraint is allowed. ρ is the relaxation penalty coefficient, which controls the degree of penalty for constraint violation. A larger value is chosen to minimize the penalty. This strategy, which enables the robot to leave the environment under extreme surface conditions, enhances the robustness of the six-bar tensioning robot.

[0079] When the motion mode is jump mode, the jump control unit 32 is activated, wherein the launch parameter calculation subunit 32A is used to execute step S32A: calculate the launch speed and angle based on the current position coordinates of the six-bar tensioning robot and the obstacle height in the environmental information.

[0080] The launch velocity is calculated using the following formula:

[0081] in, For Martian gravitational acceleration, The height of the obstacle. For safety margin height, For horizontal crossing distance.

[0082] The jump thrust optimization subunit 32B is used to perform step S32B: applying thrust using a trapezoidal thrust profile according to the launch velocity and angle.

[0083] The jump execution comprises three phases: launch phase (duration Tlaunch = 0.3 s, using a trapezoidal thrust profile); flight phase (free parabolic motion, with real-time trajectory monitoring); and landing phase (accelerometer detection of ground contact). After the jump, it returns to rolling mode and continues executing subsequent gait sequences. Furthermore, to prevent damage to the six-bar tensioning robot from the impact force after the jump, preferably, after applying thrust using the trapezoidal thrust profile in step S32B, the following steps are also included: The robot's acceleration before touching the ground is obtained. If the robot's acceleration exceeds a preset threshold, the motor is controlled to apply a damping thrust.

[0084] In this embodiment, the accelerometer detects acceleration. a The value is greater than If the direction of acceleration is the same as the direction of gravity, then a damping thrust is applied. This makes the landing stable.

[0085] Furthermore, in another preferred embodiment, to further improve the robustness of path tracking for the six-bar tensioning robot, the system introduces trajectory tracking based on model predictive control. This allows for the re-optimization of the reference trajectory based on the error between the reference trajectory and the actual trajectory, as well as the updated robot state. This enables error control during the control process, thereby overcoming the uncertainties introduced by the Martian surface environment and enhancing the robustness of the hybrid control system for the six-bar tensioning robot. Please refer to [link to relevant documentation]. Figure 4 and Figure 5 At this time, the hybrid control system also includes an MPC trajectory optimization unit 40, which is used to execute step S40: obtain the reference trajectory and the real trajectory of the six-bar tensioning robot in the preset control time domain, construct the MPC optimization problem by combining the preset state weight matrix and the control weight matrix, and output the optimal face selection decision for the next step according to the constraints.

[0086] In this embodiment, the prediction time domain is set. =10, control time domain =5, sampling time 0.5s; the preset state weight matrix Q focuses on position tracking, and the control weight matrix R balances control energy consumption; in each control cycle, the MPC optimization problem is solved, and the optimal surface for the next step is selected from the candidate surfaces provided by the gait planning sequence; firstly, a discrete-time dynamic model is established based on the state of the six-bar tensioning robot:

[0087] Among them, state Includes x-axis coordinates, y-axis coordinates, and angles; control input. Choose a strategy for the surface.

[0088] Then, the MPC optimization problem is constructed:

[0089]

[0090]

[0091]

[0092] in, For prediction in the time domain; for The actual state under the serial number includes coordinate information and direction; For reference trajectory; To make a decision based on the surface; To control the time domain; This is the state weight matrix; To control the weight matrix; This is the set of adjacent faces that are reachable in the current state; This is the set of state constraints.

[0093] Thus, MPC and gait planning form a coordination mechanism, with gait planning generating an initial surface transition sequence at the start of the task. In each control cycle (sampling time 0.5 s), MPC adjusts the current actual state. and reference trajectory From the candidate face set provided by the gait sequence The system optimizes the selection of the next optimal surface; when the actual execution deviates from the gait planning path, the MPC can request to replan the local gait sequence; this coordination mechanism enables the system to utilize the global energy optimization of gait planning, and also to achieve local trajectory tracking and deviation correction through the MPC.

[0094] Preferably, if convergence fails after the MPC optimization solution in step S40, the state constraint set is updated using the following formula:

[0095] in, For constraint coefficients, This is the set of state constraints. Uncertainties related to gravity, friction, and wind disturbances are mitigated by tightening the constraints by 10%, providing a margin to enhance the system's robustness.

[0096] Simulation verification was conducted under Martian environmental parameters (gravity 3.72 m / s², ground stiffness 20000 N / m, friction coefficient 0.9), and the results are as follows: (1) Straight path (50 m): Tracking error <0.3 m, energy consumption reduced by 23.5% compared to heuristic method; (2) Zigzag path (6 segments, lateral amplitude 5m): Path tracking successfully completed; (3) Curved path (sine trajectory, amplitude 4m): Tracking accuracy meets requirements; (4) Obstacle crossing test: Successfully crossed an obstacle with a height of 0.4 m.

[0097] In summary, the hybrid control method for a six-bar tensioning robot for Mars surface exploration proposed in this invention first models the six-bar tensioning robot, defining the coordinates and adjacency relationships of each face, and pre-establishing an energy landscape map. Then, during the control of the six-bar tensioning robot, environmental data is collected synchronously, and the rolling or jumping mode is switched based on the robot's state and the environmental data. In rolling mode, A... The algorithm searches for the energy-optimal surface transition sequence and, based on adaptive thrust using quadratic programming, obtains the optimal motor thrust allocation for each surface transition in the sequence by solving a constraint optimization problem. In jump mode, a safe height margin is maintained, and the magnitude and angle of the motor thrust are adjusted to overcome Martian gravity, improving the terrain adaptability of the six-bar tensioning robot. This combination solves the problems of low travel efficiency, high energy loss, and weak maneuverability of existing technologies on the complex Martian surface. Furthermore, this invention introduces MPC trajectory tracking, which re-optimizes the reference trajectory based on the error between the reference trajectory and the actual trajectory, as well as the updated robot state, achieving error control during the control process. Constraint coefficients are introduced to correct the trajectory using the state constraint set, thereby overcoming the uncertainties brought by the Martian surface environment and enhancing the robustness of the hybrid control system of the six-bar tensioning robot.

[0098] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and the present invention also intends to include these modifications and variations.

Claims

1. A hybrid control method for a six-bar tensioning robot for Mars surface exploration, characterized in that, include: S10: Initialize the six-bar tensioning robot model and acquire the six-bar tensioning robot data to update the state information of the six-bar tensioning robot; wherein, the information of the six-bar tensioning robot model includes the coordinates of each node of the six-bar tensioning robot, the closed triangle face, the open triangle face, and the adjacency relationship diagram between the faces; S20: Based on whether the state information and environmental information of the six-bar tensioning robot meet the first condition, if they do, the motion decision mode is rolling; otherwise, it is jumping. When the motion decision mode is scrolling, steps S31A and S31B are executed, including: S31A: Based on the current ground contact surface, current position coordinates, target position coordinates, and an energy landscape map of the six-bar tensioning robot, perform energy-sensing-based gait planning, generate the energy-optimal surface transformation sequence, and obtain a reference trajectory that includes rolling surface transformation. S31B: Based on the reference trajectory and the current ground contact surface, the adaptive thrust allocation optimization is performed to obtain the optimal thrust vector, and the six-bar tensioning robot is controlled to roll according to the optimal thrust vector; When the motion decision mode is jumping, steps S32A and S32B are executed, including: S32A: Calculate the launch speed and angle based on the current position coordinates of the six-bar tensioning robot and the obstacle height in the environmental information; S32B: Thrust is applied using a trapezoidal thrust profile based on the launch velocity and angle.

2. The hybrid control method for a six-bar tensioning robot for Mars surface exploration according to claim 1, characterized in that, The optimal thrust vector in step S31B is obtained by the following formula: in, The unit direction vector of the roll axis. Let be the position vector of the first node of the shared edge in the world coordinate system. The position vector of the second node of the shared edge in the world coordinate system; Using the world coordinate system, This is the gravitational torque; Let be the vector from the endpoint of the rolling edge to the robot's center of mass; The rolling edge vector; For the first i The torque coefficient of each motor; This is the rotation matrix from the body coordinate system to the world coordinate system; For the first i The position vector of each motor in the machine body coordinate system; The world coordinates of the reference point on the roll axis; is the unit z-axis vector in the body coordinate system; Let W be the thrust vector of the four motors; W be the energy weighting matrix; and c be the linear energy loss vector. This is the torque margin coefficient. This is the total thrust limiting factor; For Martian gravitational acceleration, This represents the maximum thrust of the motor.

3. The hybrid control method for a six-bar tensioning robot for Mars surface exploration according to claim 2, characterized in that, Following step S31B or step S32B, the method further includes: S40: Obtain the reference trajectory and the actual trajectory of the six-bar tensioning robot in the preset control time domain, construct the MPC optimization problem by combining the preset state weight matrix and control weight matrix, and output the optimal face selection decision for the next step according to the constraints.

4. The hybrid control method for a six-bar tensioning robot for Mars surface exploration according to claim 3, characterized in that, The MPC optimization problem is constructed using the following formula: in, For prediction in the time domain; for The actual state under the serial number includes coordinate information and direction; For reference trajectory; To make a decision based on the surface; To control the time domain; This is the state weight matrix; To control the weight matrix; This is the set of adjacent faces that are reachable in the current state; This is the set of state constraints.

5. The hybrid control method for a six-bar tensioning robot for Mars surface exploration according to claim 4, characterized in that, If convergence fails after the MPC optimization solution in step S40, the state constraint set is updated using the following formula: in, For constraint coefficients, This is the set of state constraints.

6. The hybrid control method for a six-bar tensioning robot for Mars surface exploration according to claim 5, characterized in that, The energy landscape map is constructed through the following steps: Calculate the thrust energy loss, potential energy loss, and friction loss when each surface rolls to the corresponding adjacent surface to obtain the corresponding surface transition energy loss. An energy landscape map is constructed based on the surface transformation relationship and the corresponding surface transformation energy loss.

7. The hybrid control method for a six-bar tensioning robot for Mars surface exploration according to any one of claims 1-6, characterized in that, The formula for calculating the launch velocity in step S32A is as follows: in, For Martian gravitational acceleration, The height of the obstacle. For safety margin height, For horizontal crossing distance.

8. The hybrid control method for a six-bar tensioning robot for Mars surface exploration according to claim 7, characterized in that, After applying thrust using a trapezoidal thrust profile in step S32B, the method further includes: The robot's acceleration before touching the ground is obtained. If the robot's acceleration exceeds a preset threshold, the motor is controlled to apply a damping thrust.

9. The hybrid control method for a six-bar tensioning robot for Mars surface exploration according to claim 8, characterized in that, The optimal surface transformation sequence is achieved through an improved A... The algorithm searches within the energy landscape map.

10. A hybrid control system for a six-bar tensioning robot, characterized in that, include: Robot state acquisition unit: used to initialize the six-bar tensioning robot model and acquire six-bar tensioning robot data to update the state information of the six-bar tensioning robot; wherein, the information of the six-bar tensioning robot model includes the coordinates of each node of the six-bar tensioning robot, closed triangle faces, open triangle faces, and the adjacency relationship diagram between faces; Motion decision unit: Used to determine whether the state information and environmental information of the six-bar tensioning robot meet the first condition. If the condition is met, the motion decision mode is rolling; otherwise, it is jumping. The rolling control unit includes a rolling path optimization subunit: used to perform energy-sensing-based gait planning based on the current ground contact surface, current position coordinates, target position coordinates, and an energy landscape map of the six-bar tensioning robot, generate an energy-optimal surface transformation sequence, and obtain a reference trajectory including rolling surface transformation. Rolling thrust optimization subunit: used to perform adaptive thrust allocation optimization based on the reference trajectory and the current ground contact surface to obtain the optimal thrust vector, and control the six-bar tensioning robot to roll according to the optimal thrust vector; The jump control unit includes a launch parameter calculation subunit: used to calculate the launch speed and angle based on the current position coordinates of the six-bar tensioning robot and the obstacle height in the environmental information; Skip thrust optimization subunit: used to apply thrust using a trapezoidal thrust profile based on the launch velocity and angle.

Images