Multi-mode motion robot and motion method thereof

Abstract

This invention provides a multi-mode motion robot and its motion method, including a ducted thruster, a hip module, a left leg module, a right leg module, and a control module; the ducted thruster is connected to the hip module; the hip module is symmetrically equipped with the left leg module and the right leg module; the hip module includes a thruster yaw unit and a hip joint unit; the left leg module and the right leg module are symmetrically equipped and respectively include a knee joint unit, a tibia unit, an ankle unit, and a heel unit; to achieve a unique multi-mode motion capability, thereby enabling a robot that can simultaneously perform search and rescue and reconnaissance missions in both ground and air environments.

Description

Technical Field

[0001] This invention relates to the field of robotics, and more specifically, to a multi-mode motion robot and its motion method. Background Technology

[0002] Current robotic platforms can be categorized into land-based and aerial platforms based on their primary application areas. Aerial platforms are well-suited for rapid maneuverability, making them ideal for applications such as automated package delivery or monitoring and inspection from advantageous positions. They can also overcome obstacles by flying over extremely bumpy terrain that legged robots might struggle to traverse, such as navigating partially collapsed buildings after an earthquake to detect life. However, due to the need to overcome viscous air to generate lift, their energy efficiency is generally low. Furthermore, the rapidly rotating wings of rotary-wing aerial robots can cause severe lacerations and other catastrophic consequences if they collide with a human, making them unsuitable for tasks involving human interaction. Unlike aerial robots, legged systems can operate safely near humans and animals, making them ideal for tasks requiring close human interaction and collaboration, such as search and rescue within buildings, digital agriculture, monitoring livestock disease symptoms, and assisting construction workers. Moreover, the bipedal structure, while possessing a certain obstacle-crossing capability, maintains excellent kinetic energy efficiency, making it ideal for long-duration operations. However, existing bipedal robots can only walk on the ground and have poor adaptability to terrain, making them unable to perform tasks in terrain conditions such as the ruins of collapsed buildings after an earthquake.

[0003] In view of this, the present invention proposes a multi-mode motion robot and its motion method, which integrates the motion capabilities of an aerial propulsion system and a leg propulsion system onto a single platform to achieve a unique multi-mode motion capability, thereby enabling a robot that can perform search and rescue and reconnaissance missions simultaneously in both ground and air environments. Summary of the Invention

[0004] The present invention aims to provide a multi-mode motion robot, comprising a ducted thruster, a hip module, a left leg module, a right leg module, and a control module. The ducted thruster is connected to the hip module and provides thrust to the motion robot to ensure operational stability. The hip module symmetrically includes the left and right leg modules. The hip module includes a thruster yaw unit and a hip joint unit. The thruster yaw unit controls the robot's steering. The hip joint unit controls the movement of the left and right leg modules. The left and right leg modules are symmetrically arranged and each includes a knee joint unit, a tibia unit, an ankle unit, and a heel unit. The knee joint unit, tibia unit, and ankle unit control the stride length of the left and right leg modules. The heel unit controls the landing angle of the feet in the left and right leg modules. The control module controls the movement of the ducted thruster, the hip module, the left leg module, and the right leg module.

[0005] Furthermore, the hip joint unit includes a first hip joint unit and a second hip joint unit, which can respectively control the stride and direction of the left leg module and the right leg module.

[0006] Furthermore, a first link is provided between the first hip joint unit and the second hip joint unit; a second link and a third link are provided between the tibia unit and the ankle unit; and a fourth link and a fifth link are provided between the ankle unit and the foot.

[0007] The present invention aims to provide a motion method for a multi-mode motion robot, comprising: determining a reference trajectory based on a target motion state; the reference trajectory refers to a reference gait sequence of motors at multiple time points required for the robot to reach the target motion state; determining an actual trajectory based on the robot's state; the actual trajectory refers to the actual gait sequence of the robot during actual movement and the predicted actual gait sequence at predicted subsequent time points; optimizing a control trajectory by reducing the difference between the reference trajectory and the actual trajectory; the control trajectory refers to the control gait sequence for controlling the robot's movement; and iteratively optimizing the control trajectory during the robot's movement until the robot reaches the target motion state.

[0008] Further, determining the reference trajectory includes: determining a simplified model of the robot; the simplified model is a link model; performing gait planning on the simplified model based on the target motion state; the gait planning includes the torque of each link; determining a full-order model of the robot; the full-order model refers to the physical model of the robot, and the full-order model includes multiple motors acting as joints; applying the gait planning to the full-order model using a parameterization method to determine the torque of the multiple motors; optimizing the full-order model based on a cost function and constraints to obtain the reference trajectory.

[0009] Furthermore, the expression for the cost function is:

[0010]

[0011] Where Cost represents the total cost function; γ represents the walking progress per step; μ represents the vertical displacement of the robot's upper body; c represents the sub-displacement calculation coefficient; q pitch q represents the amount of upper body tilt displacement; roll q represents the amount of upper body roll displacement; yaw q represents the upper body yaw displacement; hyL Indicates the yaw displacement of the left hip joint; q hpL Indicates the tilt displacement of the left hip joint; q hrL q represents the rolling displacement of the left hip joint; hyR Indicates the yaw displacement of the left hip joint; q hpR Indicates the tilt displacement of the right hip joint; q hrL This indicates the amount of rolling displacement of the left hip joint.

[0012] Furthermore, the constraints include:

[0013] During walking, the swinging foot does not contact the ground;

[0014]

[0015] in, This indicates the contact force between the swing foot and the ground during the switching process; q represents the contact force between the supporting foot and the ground during the switching process; q represents the instantaneous process of switching between the two feet; - q represents the instant before the switch. + Indicates the instant after the switch;

[0016] The supporting foot should not slip or leave the ground during walking;

[0017]

[0018] in, This represents the contact force between the supporting foot and the ground in the z-direction.

[0019]

[0020] Where μ represents the static friction coefficient of the ground. This represents the contact force between the supporting foot and the ground in the x-direction. This represents the contact force between the supporting foot and the ground in the y direction.

[0021] The swinging foot does not slip during contact with the ground;

[0022]

[0023] Where μ represents the static friction coefficient of the ground. This represents the contact force between the swinging foot and the ground in the z-direction. This represents the contact force between the swinging foot and the ground in the x-direction. This represents the contact force between the swinging foot and the ground in the y-direction.

[0024] During the switching of feet, the supporting foot leaves the ground momentarily:

[0025]

[0026] in, This indicates the speed at which the leg swings immediately after the switch.

[0027] Furthermore, determining the actual trajectory includes: acquiring the robot's pose information; performing extended Kalman filtering on the pose information to obtain the actual gait sequence and the predicted actual gait sequence.

[0028] Furthermore, the optimized control trajectory includes: acquiring the robot's foot placement; using a PD controller to input the actual trajectory obtained by the floating base parameter state estimator and the foot placement as feedforward unit signals into the control network; and using the PD controller to input the error between the reference trajectory and the actual trajectory into the control network, so that the actual gait sequence tracks the reference gait sequence.

[0029] Furthermore, it also includes adjusting the robot's pose, including: determining whether the robot's pose change is greater than a preset pose change threshold based on the robot's pose information; if so, reducing the pose change by using a ducted thruster.

[0030] The technical solutions of the embodiments of the present invention have at least the following advantages and beneficial effects:

[0031] The multi-mode motion robot provided by this invention is equipped with a dual-degree-of-freedom ducted thruster unit, which is combined with the traditional bipedal robot dynamic walking system to form a dual-modal cooperative motion system, solving the defects of traditional bipedal robots in poor adaptability to complex terrain and insufficient traversal ability.

[0032] The multi-mode motion robot proposed in this invention can achieve 3D dynamic walking on the ground using a bipedal drive system by means of movable joints in the waist, and has broad application prospects. Attached Figure Description

[0033] Figure 1 A schematic diagram illustrating the degree-of-freedom configuration of the multi-mode motion robot provided by this invention;

[0034] Figure 2 Hardware signal flow diagram of the controller for the multi-mode motion robot provided by the present invention;

[0035] Figure 3 This is a block diagram of the control method for a multi-mode motion robot provided by the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0037] The multi-mode motion robot provided by the present invention includes a ducted thruster, a hip module, a left leg module, a right leg module, and a control module.

[0038] The ducted thruster is connected to the hip module and is used to provide thrust to the robot, ensuring stability during operation. The hip module is symmetrically equipped with the left leg module and the right leg module.

[0039] The hip module includes a thruster yaw unit and a hip joint unit; the thruster yaw unit is used to control the steering of the robot; the hip joint unit is used to control the movement of the left leg module and the right leg module.

[0040] The hip joint unit includes a first hip joint unit and a second hip joint unit, which can respectively control the stride and direction of the left leg module and the right leg module. The first and second hip joint units may include a lumbar yaw joint motor, a lumbar roll joint motor, and a lumbar yaw joint motor. The left leg module and the right leg module are symmetrically arranged and each includes a knee joint unit, a tibia unit, an ankle unit, and a heel unit. The knee joint unit, the tibia unit, and the ankle unit control the stride of the left leg module and the right leg module; the heel unit controls the landing angle of the foot in the left leg module and the right leg module. The knee joint unit includes a knee joint motor; the ankle unit may include a foot end motor.

[0041] The control module is used to control the movement of the ducted propeller, the hip module, the left leg module, and the right leg module.

[0042] A first link is provided between the first hip joint unit and the second hip joint unit; a second link and a third link are provided between the tibia unit and the ankle unit; a fourth link and a fifth link are provided between the ankle unit and the foot. The first link can be a link connecting the lateral yaw joint motors on both sides; the second and third links include tibial link A and tibial link B; the fourth and fifth links include a foot motor link and a lower leg link.

[0043] Figure 1 This is a schematic diagram of the degree-of-freedom configuration of the multi-mode motion robot provided by the present invention.

[0044] like Figure 1 As shown, each motor has one degree of freedom, and the arrows in the diagram indicate the rotation axes of the corresponding motors. The 16 motors provide the robot with 16 degrees of freedom, while the ducted thrusters provide two degrees of freedom in two directions (i.e., forward and backward, and turning). This makes the robot a 16-degree-of-freedom dynamic bipedal robot equipped with dual-degree-of-freedom ducted thrusters, thus enabling at least air-land dual-modal operation.

[0045] The mechanical structure of the multi-mode motion robot includes a ducted thruster, a thruster roll joint, a waist yaw joint motor, a controller carrier plate, a waist roll joint motor, a waist yaw joint motor, a knee joint motor, a knee joint link, a tibia link A, a tibia link B, a foot motor, a foot motor link, a lower leg link, and a foot surface.

[0046] The motion method for a multi-mode motion robot provided by this invention includes:

[0047] Based on the target motion state, a reference trajectory is determined; the reference trajectory refers to the reference gait sequence of the motors at multiple time points required for the robot to achieve the target motion state.

[0048] The target motion state refers to the motion state that the robot needs to achieve. This can include the robot's foot placement and movement speed, such as walking forward at 1 m / s. The reference gait sequence can refer to the torque of multiple motors as a function of time. For example, how much force each motor needs to output to walk forward at 1 m / s; and generating a set of motor torques as a function of time. The gait sequence refers to the joint's torque as a function of time, i.e., the trajectory; and the torque values ​​corresponding to the time values ​​of the motor numbers.

[0049] The process of determining the reference trajectory includes: determining a simplified model of the robot; the simplified model is a link model; performing gait planning on the simplified model based on the target motion state; the gait planning includes the torque of each link; determining a full-order model of the robot; the full-order model refers to the physical model of the robot, which includes multiple motors acting as joints; applying the gait planning to the full-order model using a parameterization method to determine the torques of the multiple motors; and optimizing the full-order model based on a cost function and constraints to obtain the reference trajectory.

[0050] A simplified model could have a first link between the first and second hip joint units; a second and third link between the tibia and ankle units; and a fourth and fifth link between the ankle unit and the foot. A full-order model could be... Figure 1 Similarly, parametric methods can refer to describing the length of links and the forces acting on them as the robot moves, using mathematical representations; this involves mathematically describing the relationship between the length of the link and the coordinate system.

[0051] The expression for the cost function is as follows:

[0052]

[0053] Where Cost represents the total cost function; γ represents the walking progress per step; μ represents the vertical displacement of the robot's upper body; c represents the sub-displacement calculation coefficient; q pitch q represents the amount of upper body tilt displacement; roll q represents the amount of upper body roll displacement; yaw q represents the upper body yaw displacement; hyL Indicates the yaw displacement of the left hip joint; q hpL Indicates the tilt displacement of the left hip joint; q hrL q represents the rolling displacement of the left hip joint; hyRIndicates the yaw displacement of the left hip joint; q hpR Indicates the tilt displacement of the right hip joint; q hrL This indicates the amount of rolling displacement of the left hip joint.

[0054] By using this cost function, we can obtain the joint torque-time series (i.e., the gait library function) that minimizes the displacement of the robot's upper body and requires the least amount of energy.

[0055] The constraints include:

[0056] During walking, the swinging foot does not contact the ground;

[0057]

[0058] in, This indicates the contact force between the swing foot and the ground during the switching process; q represents the contact force between the supporting foot and the ground during the switching process; q represents the instantaneous process of switching between the two feet; - q represents the instant before the switch. + Indicates the instant after the switch;

[0059] The supporting foot should not slip or leave the ground during walking;

[0060]

[0061] in, This represents the contact force between the supporting foot and the ground in the z-direction.

[0062]

[0063] Where μ represents the static friction coefficient of the ground. This represents the contact force between the supporting foot and the ground in the x-direction. This represents the contact force between the supporting foot and the ground in the y direction.

[0064] The swinging foot does not slip during contact with the ground;

[0065]

[0066] Where μ represents the static friction coefficient of the ground. This represents the contact force between the swinging foot and the ground in the z-direction. This represents the contact force between the swinging foot and the ground in the x-direction. This represents the contact force between the swinging foot and the ground in the y-direction.

[0067] During the switching of feet, the supporting foot leaves the ground momentarily:

[0068]

[0069] in, This indicates the speed at which the leg swings immediately after the switch.

[0070] Gait generally exhibits periodicity, which can be described in the form of a limit cycle. The periodicity condition is then set as a constraint function. Here, the right leg can be defined as the standing foot, and the left leg as the swinging foot (they can be interchanged), and the following system output is defined as the controlled variable:

[0071]

[0072] Based on the robot's state, the actual trajectory is determined; the actual trajectory refers to the actual gait sequence of the robot during actual movement and the predicted actual gait sequence at subsequent time points.

[0073] The robot's state can refer to its position and orientation. The actual gait sequence can refer to the motor torque of the robot in its current state; the predicted actual gait sequence refers to the predicted motor torque of the robot in subsequent time periods.

[0074] In some embodiments, determining the actual trajectory includes: acquiring the robot's pose information; performing an extended Kalman filter on the pose information to obtain the actual gait sequence and a predicted actual gait sequence. The pose information may include the robot's acceleration information, angular velocity information, and angle information (magnetic field information), etc.

[0075] The control trajectory is optimized by reducing the difference between the reference trajectory and the actual trajectory; the control trajectory refers to the control gait sequence that controls the robot to move.

[0076] A control gait sequence can refer to the time-varying torque sequence of the motor as a result of the actual control signal. The output information can be adjusted based on the deviation between the reference trajectory and the actual trajectory. For example, if the deviation is large, the PD controller will output more control information to reduce the error more quickly.

[0077] During the robot's movement, the control trajectory is continuously optimized until the robot reaches the target motion state.

[0078] The optimized control trajectory includes: acquiring the robot's foot placement; using a PD controller to input the actual trajectory obtained by the floating base parameter state estimator and the foot placement as feedforward unit signals into the control network; and using the PD controller to input the error between the reference trajectory and the actual trajectory into the control network, so that the actual gait sequence tracks the reference gait sequence.

[0079] In some embodiments, the method further includes adjusting the robot's pose, including: determining, based on the robot's pose information, whether the change in the robot's pose is greater than a preset pose change threshold; if so, reducing the pose change via a ducted thruster; if not, no further processing is performed.

[0080] Figure 2 The hardware signal flow diagram of the controller for the multi-mode motion robot provided by this invention.

[0081] A low-cost, high-computing-power robot master control platform was designed, featuring collaboration between the upper and lower level computers. The upper level computer runs a real-time controller based on a Linux system. Its real-time performance is primarily achieved through EtherCAT and Xenomai. EtherCAT enables the controller to transmit real-time short-frame data, achieving real-time communication. Xenomai ensures that real-time commands are executed with the highest system priority, guaranteeing the system's real-time performance. The combination of these two technologies achieves millisecond-level motion control, increasing the robot's execution efficiency and enabling the control system to meet the stable walking requirements of a bipedal robot. The lower level computer, an Arduino microcontroller, and the Odrive motor driver board control the bipedal motors and thrust vectoring motors via RS485 and CAN buses, respectively, thereby controlling the robot's movement.

[0082] The floating parameter state estimator can include a Kalman filter. Using the floating parameter state estimator and an IMU sensor, the robot's actual gait parameters are estimated. In other words, the actual gait is reconstructed using a dynamic method. The obtained reference values ​​are input into the control network; then, the error is calculated using the reference and actual values ​​to achieve feedforward. Bipedal motors control the motors on the bipedal legs; thrust vectoring motors control the thrust and direction of the ducted thrusters.

[0083] Figure 3 This is a block diagram of the control method for a multi-mode motion robot provided by the present invention.

[0084] like Figure 3 As shown, in terms of the control algorithm, the generated gait library function (reference gait sequence) is used as the reference trajectory. The floating base parameter state estimator is connected to the control network as a feedforward unit using a PD controller, enabling the actual joint trajectory to track the reference joint trajectory. Concurrent multi-threaded execution is used to run the state estimation and control unit, while increasing their update frequency. The state estimation program runs at 2000 Hz, while the control program runs at 1000 Hz, thus solving the problem of poor real-time performance in the walking part controller of traditional bipedal robots. Here, foot placement refers to the location in space where the robot's next foot should be placed.

[0085] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A motion method for a multi-mode motion robot, characterized in that, include: Determine the reference trajectory based on the target's motion state; The reference trajectory refers to the reference gait sequence of the motors at multiple time points required for the robot to achieve the target motion state; Determine the reference trajectory, including: A simplified model of the robot is determined; the simplified model is a linkage model. Gait planning is performed on the simplified model based on the target motion state; the gait planning includes the torque of each link; Determine the full-order model of the robot; the full-order model refers to the physical model of the robot, which includes multiple motors that act as joints; The gait planning is applied to the full-order model using a parameterization method to determine the torque of multiple motors; The full-order model is optimized based on the cost function and constraints to obtain the reference trajectory; the constraints include: During walking, the swinging foot does not contact the ground; in, This indicates the contact force between the swing foot and the ground during the switching process; This indicates the contact force between the support foot and the ground during the switching process; This indicates the instantaneous process of switching between the two feet; Indicates the instant before the switch. Indicates the instant after the switch; The supporting foot should not slip or leave the ground during walking; in, This represents the contact force between the supporting foot and the ground in the z-direction. in, Indicates the static friction coefficient of the ground. This represents the contact force between the supporting foot and the ground in the x-direction. This represents the contact force between the supporting foot and the ground in the y direction. The swinging foot does not slip during contact with the ground; in, Indicates the static friction coefficient of the ground. This represents the contact force between the swinging foot and the ground in the z-direction. This represents the contact force between the swinging foot and the ground in the x-direction. This represents the contact force between the swinging foot and the ground in the y-direction. During the switching of feet, the supporting foot leaves the ground momentarily: in, This indicates the speed of the leg swing at the instant after the switch; Based on the robot's state, the actual trajectory is determined; the actual trajectory refers to the actual gait sequence of the robot during actual movement and the predicted actual gait sequence at subsequent time points. The control trajectory is optimized by reducing the difference between the reference trajectory and the actual trajectory; the control trajectory refers to the control gait sequence that controls the robot's movement. During the robot's movement, the control trajectory is continuously optimized until the robot reaches the target motion state.

2. The motion method for a multi-mode motion robot according to claim 1, characterized in that, The expression for the cost function is: in, Represents the total cost function; Indicates the progress of each step of walking; This indicates the vertical displacement of the robot's upper body; Indicates the sub-displacement calculation coefficient; This indicates the amount of upper body tilt displacement; This indicates the amount of upper body roll displacement; Indicates the amount of upper body yaw displacement; This indicates the yaw displacement of the left hip joint; This indicates the tilt displacement of the left hip joint; This indicates the amount of rolling displacement of the left hip joint; This indicates the yaw displacement of the left hip joint; This indicates the amount of tilt displacement of the right hip joint; This indicates the amount of rolling displacement of the left hip joint.

3. The motion method for a multi-mode motion robot according to claim 1, characterized in that, Determining the actual trajectory includes: Obtain the robot's pose information; The pose information is subjected to extended Kalman filtering to obtain the actual gait sequence and the predicted actual gait sequence.

4. The motion method for a multi-mode motion robot according to claim 1, characterized in that, The optimized control trajectory includes: Obtain the foot placement of the robot; The actual trajectory obtained by the floating base parameter state estimator and the foot position are used as feedforward unit signals to be connected to the control network using the PD controller. The error between the reference trajectory and the actual trajectory is incorporated into the control network by the PD controller, so that the actual gait sequence tracks the reference gait sequence.

5. The motion method for a multi-mode motion robot according to claim 1, characterized in that, This also includes adjusting the robot's pose, including: Based on the robot's pose information, determine whether the robot's pose change is greater than a preset pose change threshold. If so, the pose change is reduced by using a ducted thruster.

Images