Intelligent accompanying robot with body

Abstract

The intelligent accompanying robot comprises a vehicle body, a traction mechanism and a control device, and is characterized in that the traction mechanism is mainly composed of a waistband, a sensing power assisting device and a traction connecting device; the sensing power assisting device comprises two walking power assisting motors installed on the waistband, a left lower limb binding belt, a right lower limb binding belt, a left walking power assisting connecting rod and a right walking power assisting connecting rod, each lower limb binding belt is connected with the corresponding walking power assisting motor through the corresponding walking power assisting connecting rod, and when a user walks, the thighs drive the walking power assisting connecting rods to move through the lower limb binding belts; the walking assistance motor is twisted to rotate to generate current, the current serves as a sensing signal to be output to the control device, and the control device judges the walking intention of a user, controls the walking assistance motor to move to provide walking assistance and meanwhile controls the vehicle body to conduct following movement. While accompanying traction assistance is achieved, walking assistance can be provided, and the single-person heavy-load transportation and assistance requirements under field complex terrains can be met.

Description

Technical Field

[0001] The present invention relates to a robot for load transportation, specifically to an attendant robot, especially an attendant robot adopting embodied intelligence technology. Background Art

[0002] When a single person needs to carry supplies during work, reducing the load that the single person himself / herself needs to bear through equipment is one of the keys to improving the combat effectiveness of individual soldiers and the work efficiency of single persons. Especially in the fields of military, fire fighting, search and rescue, transportation, etc., there are high requirements for the movement flexibility and physical fitness maintenance of a single person after carrying supplies.

[0003] Currently, the available field material transportation equipment mainly includes devices such as single-person exoskeletons, all-terrain unmanned vehicles, and robot dogs. These devices can achieve the transportation of a certain load of equipment and supplies, but they all have some problems and cannot simultaneously possess the two functions of providing assistance for the wearer's walking and providing traction assistance for the load carried by the wearer. Therefore, they cannot meet the requirements of individual soldier / single person for the attendant transportation of equipment and supplies.

[0004] Chinese invention application CN113021338A discloses an intelligent attendant robot, comprising: a body, an execution mechanism, a driving device, a detection device, and a control device; the body is used to form the shape of the intelligent attendant robot and provide support and protection for the execution mechanism, the driving device, the detection device, and the control system; the detection device is used to detect environmental information; the execution control includes analyzing the detection information and making the robot closely follow the target based on the detection information. Among them, the detection device is used to detect environmental information; it includes an infrared device, Bluetooth, a camera, etc. This device compensates for the robot drive through the detection information, predicts through the change trend based on duration or time slots, and provides strong stability for the compensated drive through continuous balance of the driving force. While providing compensation, the horizontal direction attendant control is carried out through the unified driving force to avoid the loss of the robot. The above technical solution can achieve the attendant of the robot, but it uses detection information such as infrared, Bluetooth, and cameras, and it is difficult to adapt to the application in the complex terrain in the wild.

[0005] Another consideration is to use traction accompanying power assistance. For example, Chinese invention patent CN113173043A discloses a single-soldier power-assisted tractor, including a body and a chassis device, wherein the body is mounted on the chassis device, and the chassis device includes a frame, two flexible buffer units, two walking mechanism units and an electronic control unit, wherein the two walking mechanism units are located on both sides of the frame and are connected to the frame through a flexible buffer unit. This two-wheeled portable vehicle is low-cost and highly flexible, but because its existing control method is mainly manpower or fixed-speed remote control, it is impossible to adjust the speed and lifting height of the vehicle in real time according to the speed of human movement and changes in road conditions, resulting in poor adaptability of human-machine movement, limited terrain adaptability under complex road conditions, and dragging or compression of the human-machine connection part.

[0006] One direction for improving traction assistance is the adaptability of human-machine motion, and the prediction of human motion can be improved by adding sensors. For example, Chinese invention application CN119459942A discloses a waist-hanging three-wheeled power-assisted vehicle based on multi-sensor data fusion drive, including a wearable vest, a button buckle device, a traction device, a middle rotating device, a wheel-leg device, a body fixing frame, an integrated control box and a driving rear wheel; the wearable vest is connected to the middle rotating device through a button buckle device and a traction device, and two driving rear wheels are symmetrically arranged on both sides of the bottom of the body fixing frame; an inertial measurement unit IMU is arranged on the wearable vest, the wheel-leg driving motor is equipped with an absolute encoder, a wheel-leg knee joint torque sensor is arranged in the wheel-leg device, a driving rear wheel torque sensor is arranged at the inner shaft of the driving rear wheel, and a speed encoder is arranged between the driving rear wheel housing and the driving rear wheel torque sensor. This solution uses the inertial measurement unit (IMU) installed on the wearable vest to collect the three-axis posture and acceleration of the human back, and combines multiple sensors such as the wheel-leg knee joint torque sensor, absolute value encoder, drive rear wheel torque sensor, speed encoder, etc. to obtain human motion information and the force of the drive rear wheel through relatively complex calculations, perform decision analysis, and finally obtain the input of the drive rear wheel motor. This solution can adjust the speed of the drive rear wheel in real time according to the movement of the human body, and adjust the angle of the wheel-leg drive joint in real time according to changes in road conditions, reducing damage to the human body and improving adaptability. However, its structure is complex and the calculation and processing requirements are high. In addition, it judges the movement intention by collecting the posture of the human back, and it is difficult to distinguish and judge the non-traction intention movements of the human body, which is difficult to use in actual applications.

[0007] A single-person exoskeleton can provide assistance to the wearer's walking to a certain extent, but it cannot provide assistance for the load carried by the wearer, so it does not have load-bearing capacity.

[0008] Therefore, new equipment research and development is needed to address the problem that existing individual / single-person accompanying equipment cannot simultaneously provide assistance to the wearer's walking and provide traction assistance for the load carried by the wearer. Summary of the Invention

[0009] The object of the present invention is to provide an embodied intelligent follow - along robot, which can provide walking assistance while realizing follow - along traction assistance to meet the needs of single - person large - load transportation and assistance in complex terrain in the wild.

[0010] To achieve the above - mentioned object of the invention, the technical solution adopted by the present invention is: an embodied intelligent follow - along robot, including a vehicle body, a traction mechanism and a control device. The vehicle body has a traction frame, a carrying platform and two motor - driven wheels. The traction mechanism is mainly composed of a belt, a sensing assistance device and a traction connection device. One end of the traction connection device is connected to the belt, and the other end is connected to the traction frame. A buffer component is arranged in the traction connection device. The sensing assistance device includes two walking assistance motors installed on the belt, two left - and - right lower - limb binding straps and two left - and - right walking assistance connecting rods. Each lower - limb binding strap is connected to the corresponding walking assistance motor through the corresponding walking assistance connecting rod. When the user walks, the thigh drives the walking assistance connecting rod to move through the lower - limb binding strap, twists the walking assistance motor to rotate and generate an electric current, which is output as a sensing signal to the control device. The control device judges the walking intention of the user, controls the movement of the walking assistance motor to provide walking assistance, and at the same time controls the vehicle body to perform follow - along movement.

[0011] In a further technical solution, a waist gyroscope is arranged on the belt, and the angle signal output by the waist gyroscope is connected to the control device.

[0012] In the above - mentioned technical solution, the method for judging the walking intention of the user is to respectively collect the current, voltage and motor encoder signals of the left - and - right walking assistance motors, perform conversion calculations in combination with the mechanical parameters of the walking assistance connecting rod and the walking assistance motor to obtain the biometric information of the movement amplitude and phase sequence of the two thighs during human movement, collect the rotation speeds of the two walking assistance motors, collect the angle signal output by the waist gyroscope, construct a motion pattern recognition module in the control device, use the above - collected signals as inputs, output motion intention judgment signals of forward, backward, acceleration, deceleration, left - turn, and right - turn, and train the motion pattern recognition module to realize the real - time response of the robot to the walking intention of the user.

[0013] Among them, the method for providing walking assistance is that the control device recognizes the walking intention of the user, divides the gait cycle into a stance phase and a swing phase. In the stance phase, the human body needs to bear the body weight and push the body forward. At this time, the swing amplitude of the supporting leg is small, and the control device controls the walking assistance motor to output a small - amplitude movement. In the swing phase, the human body needs to quickly move the lower limbs. The control device reduces the output torque of the walking assistance motor, and at the same time adjusts the rotation speed and motion phase to make the walking assistance motor of the follow - along robot rotate and drive the walking assistance connecting rod and the lower - limb binding strap to swing, so as to realize following the swing of the human lower limbs.

[0014] In a preferred technical solution, a model predictive control module is arranged in the control device to predict the walking state in a future period of time, optimize the control input in advance, and achieve the motion synchronous coupling between the follow-up assist robot and the wearer.

[0015] In the above technical solution, the method for controlling the vehicle body to perform following motion is that each of the two motor drive wheels is driven by an independent drive motor, and the output control signals of the control device are respectively connected to control the two motor drive wheels. When the control device identifies the walking intention of the user, it simultaneously obtains the motion direction angle value and acceleration value of the user, and realizes the uniform following, acceleration, deceleration, left turn and right turn of the vehicle body by respectively controlling the motion speed and acceleration of the two motor drive wheels.

[0016] In a preferred technical solution, an angle adjusting device is arranged between the traction frame and the bearing platform, which is used to adjust the angle between the buffer component and the traction frame to adapt to users of different heights.

[0017] In a preferred technical solution, a battery is installed below the bearing platform, and the control device is installed on the traction frame. The traction connecting device and the traction frame have hollow wire routing channels, and the input and output wires of the two walking assist motors are connected to the control device through the wire routing channels.

[0018] In the above technical solution, a quick connection interface can be arranged between the traction connecting device and the waistband. The quick connection interface includes a mechanical connection component and an electrical connection interface, which can realize quick connection and quick release, making the robot system have good flexibility.

[0019] When in use, the user binds the waistband around the waist and binds the left and right lower limb binding straps around the left and right thighs respectively. When the user walks, the thighs drive the walking assist connecting rod to move through the lower limb binding straps, and the walking assist connecting rod twists the walking assist motor to rotate, thereby generating an electric current. The control device obtains the walking intention of the user by measuring the current signal generated by the walking assist motor. Based on the signals of the left and right walking assist motors configured in the follow-up assist robot, through filtering and data processing, and combined with the output signal of the gyroscope configured on the waistband, the forward, backward, acceleration, deceleration and left and right turning intentions of the user can be identified. Thus, the walking assist motor is controlled to move according to the walking intention of the user to provide assistance for the user's walking, and the drive motor is controlled to move according to the walking intention of the user to provide assistance for the user's traction.

[0020] Therefore, during the user's walking process, the accompanying power-assisted robot can follow the user's walking speed and control the driving motor to move at the same speed. When the user accelerates, the robot also accelerates; when the user decelerates, the robot also decelerates; when the user walks at a constant speed, the robot also walks at a constant speed; and when the user turns left or right, the robot also turns left or right accordingly. The accompanying power-assisted robot can follow the user to walk in the wild environment when loaded with equipment and supplies weighing hundreds of kilograms, which can greatly save the user's physical strength.

[0021] In the control device of the accompanying power-assisted robot, the set motion mode recognition module and motion control module can have the ability of autonomous learning. Based on deep learning technology, when the robot is worn by users with different body sizes or different walking habits, it can collect data on the user's walking actions and conduct in-depth learning, and correct the stored motion data through learning, so that when the robot controls the walking assistance motor and the driving motor for walking assistance and traction assistance, it is more matched with the user's motion habits, providing a more comfortable experience for the user.

[0022] Due to the application of the above technical solutions, the present invention has the following advantages compared with the prior art: 1. The accompanying power-assisted vehicle of the present invention realizes the function of recognizing the user's motion intention by collecting the myoelectric signals of the user's lower limbs and combining with the gyroscope configured on the waistband, and can sense the user's motion intentions such as forward movement, left and right turning, acceleration, deceleration, and constant speed, and automatically control the vehicle to follow the user's walking intention for movement. Compared with the recognition through human body postures, the intention recognition is more accurate and more convenient to implement.

[0023] 2. The walking assistance motor and the driving motor of the accompanying power-assisted robot of the present invention are uniformly controlled by the control device and powered by the large-capacity battery loaded by the traction mechanism. Therefore, it has flexible stability and long battery life. When loaded with equipment and supplies weighing dozens of kilograms or even hundreds of kilograms, it can follow the user to walk in the wild environment, greatly saving the user's physical strength.

[0024] 3. The traction mechanism and the walking assistance mechanism of the accompanying power-assisted robot of the present invention are connected through the quick connection interface on the waist. This quick connection interface can be quickly connected and quickly disconnected, making the robot system have good flexibility.

[0025] 4. The accompanying power-assisted robot of the present invention has the ability of autonomous learning. Based on deep learning technology, when the robot is worn by users with different body sizes or different walking habits, it can collect data on the user's walking actions and conduct in-depth learning, and correct the stored motion data through learning, so that when the robot controls the walking assistance motor and the driving motor for walking assistance and traction assistance, it is more matched with the user's motion habits, providing a more comfortable experience for the user.

[0026] 5. The accompanying power-assisted robot of the present invention also has a buffer component, so that there is a certain buffering effect between the traction mechanism and the user's waist during walking, avoiding a large impact on the user and improving the comfort of use. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a schematic structural diagram of an embodiment of the present invention; Figure 2 is a view in another direction in the embodiment; Figure 3 is a flow block diagram of motion intention recognition in the embodiment.

[0028] Wherein: 1. Lower limb binding belt; 2. Walking assistance link; 3. Walking assistance motor; 4. Waist belt; 5. Quick connection interface; 6. Buffer component; 7. Angle adjustment mechanism; 8. Traction vehicle frame; 9. Control device; 10. Carrying platform; 11. Battery unit; 12. Motor drive wheel. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] The present invention will be further described below in conjunction with the drawings and embodiments: Embodiment: Refer to Figure 1 and Figure 2 As shown, a physical intelligent accompanying robot includes a vehicle body, a traction mechanism and a control device.

[0030] As Figure 1 , the vehicle body has a traction vehicle frame 8, a carrying platform 10 and two motor drive wheels 12. Each motor drive wheel 12 is driven by a drive motor, so that the rotation speed and acceleration of the two wheel bodies can be independently controlled. By controlling the rotation speed of the wheel bodies, the vehicle body can turn left and right, and at the same time provide corresponding compensation for the deflection caused by uneven road surfaces.

[0031] An angle adjustment device is provided between the traction vehicle frame 8 and the carrying platform 10 for adjusting the angle between the buffer component and the traction vehicle frame to adapt to users of different heights. In this embodiment, as Figure 1 shown, the traction vehicle frame 8 and the carrying platform 10 are rotatably connected and are provided with two angle adjustment bars. The angle adjustment bars, the traction vehicle frame and the carrying platform form a triangular stable structure. Among them, at least one end of the angle adjustment bar can be connected through different positioning holes, so as to change the angle of the triangle and realize the angle adjustment of the traction vehicle frame relative to the carrying platform, thereby changing the upper connection height of the traction vehicle frame.

[0032] In this embodiment, the traction mechanism is composed of a waist belt, a sensing assistance device and a traction connection device.

[0033] Wherein, one end of the traction connection device is connected to the waist belt 4, and the other end is connected to the traction vehicle frame 8. A quick connection interface 5 is provided between the traction connection device and the waist belt 4. The quick connection interface 5 includes a mechanical connection component and an electrical connection interface, enabling the quick connection and quick disconnection of the traction connection device and the waist belt. A buffer component 6 is provided inside the traction connection device, and it is connected to the traction vehicle frame 8 through an angle adjustment mechanism 7, thereby ensuring the normal traction connection between the two when the height of the connection part of the traction vehicle frame changes, so as to adapt to users of different heights.

[0034] In this embodiment, the sensing and assisting device includes two walking assisting motors 3 installed on the waist belt 4, two left and right lower limb binding straps 1, and two left and right walking assisting linkages 2. Each of the lower limb binding straps 1 is connected to the corresponding walking assisting motor 3 through the corresponding walking assisting linkage 2; a waist gyroscope is provided on the waist belt 4, and the angle signal output by the waist gyroscope is connected to the control device.

[0035] In this embodiment, a battery unit 11 is installed below the carrying platform 10, and a control device 9 is installed on the traction vehicle frame 8. The traction connection device and the traction vehicle frame 8 have hollow wire routing channels, and the input and output wires of the two walking assisting motors are connected to the control device 9 through the wire routing channels.

[0036] When the user walks, the thigh drives the walking assisting linkage to move through the lower limb binding strap, twists the walking assisting motor to rotate and generate current, which is output as a sensing signal to the control device. The control device judges the walking intention of the user, controls the movement of the walking assisting motor to provide walking assistance, and at the same time controls the vehicle body to perform a following movement.

[0037] For the method of judging the walking intention of the user, see the appendix Figure 3As shown. The companion assist robot identifies the user's motion intention by collecting and processing the signals of the left and right walking assist motors, combined with the human motion biometric information and the gyroscope angle output signal. Specifically, first, the current, voltage, and motor encoder signals of the left and right walking motors are collected, and algorithms such as Kalman filtering and wavelet denoising are used to preprocess the original signals to eliminate the influence of electromagnetic interference and environmental noise, extract stable and effective features, and through conversion calculations in combination with the mechanical structure dimensions such as the walking assist link and the walking assist motor, obtain biometric information such as the action amplitude and phase sequence of the two thighs during human motion. Through training and learning, a pattern recognition algorithm for establishing the mapping relationship between the motion intention and the signal features is realized to identify the forward and backward states, and the forward and backward signals of the user can be parsed. In addition, by accurately collecting the rotational speeds of the two walking assist motors within a unit time, the acceleration and deceleration signals of the user are obtained through the speed analysis algorithm. Finally, with the angle signal output in real time by the gyroscope configured on the belt, the left and right turn signals of the user are identified using the attitude solution algorithm. Through threshold determination and fuzzy logic reasoning, the left and right turn motion commands are accurately identified, and finally, the robot realizes the real-time response and collaborative control of the user's motion intention.

[0038] Control the motion of the walking assist motor according to the user's walking intention to provide assistance for the user's walking.

[0039] When the companion assist robot controller identifies the walking intention of the wearer according to the sensor feedback signal, the gait cycle is accurately divided into the stance phase and the swing phase. In the stance phase, the human body needs to bear the body weight and push the body forward. At this time, the swing amplitude of the supporting leg is small, and the control algorithm controls the output of this motor to move slightly. In the swing phase, the human body needs to quickly move the lower limbs. The control algorithm will reduce the motor output torque and at the same time adjust the rotational speed and motion phase, so that the walking assist motor of the companion robot rotates and drives the walking assist link and the lower limb binding belt to swing, so that it can easily follow the swing of the human lower limbs. At the same time, the companion robot controller also has model predictive control (MPC). This model predictive control is based on the dynamic model of the companion robot system, predicts the state within a period of time in the future, and optimizes the control input in advance, so that the companion robot can follow the human motion more accurately. Through this real-time and accurate control, the companion assist robot can be synchronously coupled with the motion of the wearer to achieve efficient human-machine collaborative walking and provide a comfortable and natural assistance experience for the user.

[0040] Control the motion of the drive motor according to the user's walking intention to provide assistance for the user's traction.

[0041] When the companion assist robot controller identifies the walking intention of the wearer according to the sensor feedback signal, the motion direction angle value θ and the acceleration value a of the wearer can be obtained pAmong them, a positive θ represents a left turn, and a negative θ represents a right turn; a positive acceleration value a p represents acceleration, and a p negative represents deceleration, and a p being 0 represents uniform motion. During the forward following process of the accompanying assist robot, when the direction angle θ is 0, it represents going straight. At this time, the controller, based on the acceleration a p of the wearer, outputs the same motor acceleration command a m to the two motors of the accompanying vehicle, causing the motors to start accelerating and rotating with the same acceleration, thereby realizing the accelerating forward movement of the accompanying vehicle. When a p is 0, at this time the controller does not output an acceleration value, and the motor maintains the current speed of motion, that is, uniform motion is realized. Different relative motion accelerations a p generate different motor acceleration commands a m , enabling the accompanying vehicle to follow the user to accelerate, decelerate, and move at a uniform speed in a high-response-speed manner.

[0042] When the direction angle θ is not 0, it means that the accompanying robot needs to turn. The motion controller of the accompanying robot outputs different control speed commands V1 and V2 to the two motors of the accompanying robot based on the direction angle value θ, causing the motors to rotate at different speeds. Due to the speed difference ΔV between the two motors, the accompanying vehicle realizes the turning action. Moreover, different direction angle values θ generate different motor speed differences ΔV, enabling the accompanying vehicle to have different turning speeds.

[0043] This accompanying assist robot also has the ability of autonomous learning. Based on deep learning technology, when users of different body types or different walking habits wear it, the robot can collect data on the walking actions of the users and conduct in-depth learning, and correct the stored motion data through learning, making the robot control the walking assist motor and the driving motor for walking assistance and traction assistance more in line with the users' motion habits. The specific implementation process is as follows: The accompanying assistive robot continuously obtains multi-dimensional motion feature data such as joint angles, accelerations, and direction angles based on devices such as the motor current, voltage, encoder acquisition sensors, and inertial measurement unit (IMU) carried by the accompanying robot. After noise reduction, normalization, and feature extraction by the data preprocessing module, a standardized data set is formed. Based on deep learning models such as deep neural network (DNN), long short-term memory network (LSTM), or convolutional neural network (CNN), the robot conducts spatio-temporal feature mining and pattern recognition on the collected motion data. Through continuous iterative training, the mapping relationship between different body types, walking habits, and motion control parameters is established. During the training process, transfer learning technology is adopted to use the pre-trained model parameters as the initial weights and fine-tune them in combination with new user data to accelerate model convergence. At the level of control strategy optimization, the robot dynamically corrects the torque-speed curve of the walking assist motor and the traction parameters of the drive motor according to the personalized motion features output by the deep learning model. Through the online learning mechanism, the motion control parameter database is continuously updated, so that the assist mode of the accompanying robot and the traction force are highly compatible with the user's motion habits and biomechanical characteristics, realizing the adaptive optimization of human-machine collaborative motion and improving the accuracy and comfort of assist control.

Claims

1. An embodied intelligent follow - along robot, comprising a vehicle body, a traction mechanism and a control device. The vehicle body has a traction frame, a carrying platform and two motor - driven wheels, and is characterized in that: The traction mechanism mainly consists of a waist belt, a sensing and assisting device, and a traction connection device; one end of the traction connection device is connected to the waist belt, and the other end is connected to the traction vehicle frame. A buffer component is provided inside the traction connection device; the sensing and assisting device includes two walking assisting motors installed on the waist belt, two left and right lower limb binding straps, and two left and right walking assisting connecting rods. Each lower limb binding strap is connected to the corresponding walking assisting motor through the corresponding walking assisting connecting rod. When the user walks, the thigh drives the walking assisting connecting rod to move through the lower limb binding strap, twisting the walking assisting motor to rotate and generate an electric current, which is output as a sensing signal to the control device. The control device judges the walking intention of the user, controls the movement of the walking assisting motor to provide walking assistance, and simultaneously controls the vehicle body to perform a following movement.

2. The embodied intelligent companion robot according to claim 1, wherein: A waist gyroscope is provided on the waist belt, and the angle signal output by the waist gyroscope is connected to the control device.

3. The embodied intelligent companion robot according to claim 2, wherein: The method for judging the walking intention of the user is to respectively collect the current, voltage, and motor encoder signals of the left and right walking assisting motors, perform conversion calculations in combination with the mechanical parameters of the walking assisting connecting rod and the walking assisting motor to obtain the biometric information of the movement amplitude and phase sequence of the two thighs during human movement, collect the rotation speeds of the two walking assisting motors, collect the angle signals output by the waist gyroscope, construct a motion pattern recognition module in the control device, use the above collected signals as inputs, output motion intention judgment signals of forward, backward, acceleration, deceleration, left turn, and right turn, and train the motion pattern recognition module to achieve the real-time response of the robot to the walking intention of the user.

4. The embodied intelligent companion robot according to claim 3, wherein: The method for providing walking assistance is that the control device recognizes the walking intention of the user, divides the gait cycle into a stance phase and a swing phase. In the stance phase, the human body needs to bear the body weight and push the body forward. At this time, the swing amplitude of the supporting leg is small, and the control device controls the walking assisting motor to output a small amplitude movement; in the swing phase, the human body needs to quickly move the lower limbs. The control device reduces the output torque of the walking assisting motor, and at the same time adjusts the rotation speed and motion phase to make the walking assisting motor of the accompanying robot rotate and drive the walking assisting connecting rod and the lower limb binding strap to swing, so as to achieve following the swing of the human lower limbs.

5. The embodied intelligent companion robot according to claim 4, wherein: A model predictive control module is set in the control device to predict the walking state in the next period of time and optimize the control input in advance to achieve the motion synchronous coupling of the accompanying assisting robot and the wearer.

6. The embodied intelligent mobile robot according to claim 3, characterized in that: The method for controlling the vehicle body to perform a following movement is that the output control signals of the control device are respectively connected to control two motor drive wheels. When the control device recognizes the walking intention of the user, it simultaneously obtains the motion direction angle value and acceleration value, and realizes the uniform following, acceleration, deceleration, left turn, and right turn of the vehicle body by respectively controlling the motion speeds and accelerations of the two motor drive wheels.

7. The embodied intelligent companion robot according to claim 1, wherein: An angle adjustment device is provided between the traction vehicle frame and the carrying platform.

8. The embodied intelligent companion robot according to claim 1, wherein: A battery is installed below the carrying platform, the control device is installed on the traction vehicle frame, the traction connection device and the traction vehicle frame have hollow wire routing channels, and the input and output wires of the two walking assisting motors are connected to the control device through the wire routing channels.

9. The embodied intelligent companion robot according to claim 1, characterized in that: A quick connection interface is provided between the traction connection device and the waist belt. The quick connection interface includes a mechanical connection component and an electrical connection interface, enabling the quick connection and quick disengagement of the traction connection device and the waist belt.

Images