An embodied intelligent mobility robot and its decision-making and control method

Abstract

The present invention discloses an embodied intelligent mobility robot and a decision-making control method thereof, which mainly relates to the field of artificial intelligence, specifically an embodied intelligent mobility robot and a decision-making control method thereof. It comprises a driving chassis, an adjustment structure connected to the driving chassis with quick release, a seat articulated to the adjustment structure, armrests connected to the seat, a chair back connected to the seat rotationally, a gyroscope, and a pressure sensor; the driving chassis comprises a chassis, a plurality of driving motors arranged on the chassis, driving wheels connected to the driving motors, and a battery pack arranged on the chassis, and the adjustment structure automatically adjusts the angle posture of the seat and the chair back according to the angular position of the center of gravity of the seat and the change in gravity. The beneficial effect of the present invention is that it can more intelligently assist people with mobility difficulties to move and complete daily tasks in complex environments, and is safer and more convenient during the mobility process. The AI ​​large language model can replace people in making some decisions, and is more intelligent and simpler to operate.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence hardware, and in particular to an embodied intelligent mobility robot and a decision-making control method thereof. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] A robot is an intelligent machine capable of semi- or fully autonomous operation. It can be programmed and automatically controlled to perform tasks such as work or movement. In the field of mobility assistance, to make travel more convenient for people with limited mobility, the low-automatic nature of traditional wheelchairs is being upgraded to enable them to better assist people with limited mobility and better integrate into their lives. Embodied intelligent robots are the product of a deep integration of embodied intelligence technology with intelligent mobility tools. They are no longer simply a means of transportation; they are "intelligent entities" with autonomous perception, cognition, decision-making, and action capabilities. They can more safely, naturally, and proactively assist people with limited mobility (such as the elderly and people with disabilities) to navigate complex environments and complete daily tasks. Existing mobility robots merely solve the problem of mobility, but fail to address mobility issues more safely and easily. Existing mobility robots, when used at home, need to be stored indoors after use. Since they cannot be folded, they take up space indoors and are inconvenient to carry when needed for activities, significantly impacting the user experience. Therefore, it is necessary to ensure that the mobility robot can be folded without affecting the safety of the mobility robot, which is an issue that needs to be urgently addressed. Summary of the Invention

[0004] The purpose of the present invention is to provide an embodied intelligent mobility robot and a decision-making control method thereof, which can more intelligently assist people with mobility impairments to move and complete daily tasks in complex environments, making the mobility process safer and more convenient.

[0005] To achieve the above-mentioned purpose, the present invention is implemented through the following technical solutions:

[0006] An embodied intelligent mobility robot includes a body and an embodied sensing system mounted on the body;

[0007] The vehicle body includes a driving chassis, an adjustment structure connected to the driving chassis with a quick release, a seat hinged to the adjustment structure, an armrest connected to the seat, a seat back connected to the seat with a rotation, a gyroscope, and a pressure sensor;

[0008] The armrest is provided with a microphone, the drive chassis includes a chassis, a plurality of drive motors disposed on the chassis, drive wheels connected to the drive motors, and a battery pack disposed on the chassis, the adjustment structure automatically adjusts the angular posture of the seat and the backrest according to the angular position of the seat's center of gravity and changes in gravity, and the adjustment structure can also fold the seat and backrest;

[0009] The embodied perception system includes multiple binocular cameras arranged in the front and rear directions of the vehicle body, multiple ultrasonic radars arranged on the drive chassis, and a central decision-making unit; the central decision-making unit is connected to an AI large language model and an AI image processing module; the AI ​​large language model is connected to a microphone, and recognizes and converts the semantic information input by the microphone and makes behavioral decisions based on the semantic information; the central decision-making unit constructs and updates the map model of the surrounding environment in real time based on multiple ultrasonic radars and monitors the object model information around the vehicle body in real time, and selects a path based on the environmental map planning and user instructions; multiple binocular cameras capture images in multiple directions and measure distance, and the image information is then identified and analyzed by the AI ​​image processing module. The objects and environmental information in the image are detected, and the distance between the vehicle body and the object information is detected. The AI ​​large language model makes behavioral decisions on behalf of the passengers based on the current vehicle speed and detected target information, and controls the vehicle body to move according to the decision information.

[0010] The adjustment structure includes a seat adjustment structure and a seat back adjustment structure; the seat adjustment structure includes two quadrilateral connecting rod mechanisms and an electric push rod, the quadrilateral connecting rod mechanism includes a connecting seat connected to the bottom of the seat, a first connecting beam hinged to the connecting seat at one end, a second connecting beam hinged to the connecting seat at one end, and a clamping seat; the first connecting beam is parallel to the second connecting beam, and a plurality of limit grooves are provided at the bottom of the clamping seat, and the clamping seat is hinged to the other ends of the first and second connecting beams; a connecting rod is provided between the two first connecting beams, and a first electric push rod is rotatably connected to the connecting rod, and the push rod end of the first electric push rod is hinged to the bottom of the seat.

[0011] The seat back adjustment structure includes a second electric push rod hingedly arranged under the seat, a first linkage member hingedly arranged under the seat, and a second linkage member hingedly arranged with the first linkage member;

[0012] A push rod end of the second electric push rod is hinged to the first linkage member and the second linkage member, and a non-hinged end of the second linkage member is hinged to the chair back.

[0013] Two transverse beams are provided between the two beam frames of the chassis, two track seats are sleeved on the transverse beams, the track seats are provided with guide grooves, and the clamping seats are provided with guide columns adapted to the guide grooves;

[0014] The clamping seat is provided with a connecting shaft, and two limiting claws are rotatably connected on the connecting shaft. The limiting claws are provided with a slot, and the slot is adapted to the transverse beam. The clamping seat is provided with a control mechanism, and the control mechanism drives the two limiting claws to rotate synchronously to achieve limiting and separation.

[0015] The control mechanism includes a quick-release switch arranged under the base, a pull wire connected to the quick-release switch, a guide plate arranged on the connecting shaft, a retaining spring arranged on one side of the limit claw, and a connecting piece arranged on the two limit claws; the pull wire is connected to the connecting piece, and the connecting piece is connected to the two limit claws. When the quick-release switch drives the pull wire to retract, the limit claw rotates on the connecting shaft, and when the quick-release switch is released, the retaining spring drives the limit claw to reset.

[0016] The bottom of the guide groove is an inclined surface, the notch of the limiting groove located on the front side faces the ground, and the notch of the limiting groove located on the rear side faces the rear side.

[0017] A decision-making and control method for an embodied intelligent mobility robot:

[0018] S1: When an obstacle appears on the route, distance measurement and object recognition are performed based on the real-time image information provided by the two binocular cameras and ultrasonic radar on the front side;

[0019] S2, the AI ​​image processing module determines whether the object is dynamic or static based on the data information;

[0020] S2.1 If a dynamic object enters the primary sensing range, the vehicle's current speed is first reduced to level 1. After the speed reduction, the vehicle's maximum motion time before reaching the safe speed limit is calculated.

[0021] S2.2 then analyzes the motion trajectory of the dynamic object to determine whether the object leaves the primary sensing range before reaching the maximum braking distance. If the object does not leave, the system decelerates in multiple stages based on the maximum motion time until the object enters the final sensing range, then brakes and waits for the object to leave.

[0022] S2.3 If it is a static object, identify it based on the surrounding sensing range and choose to bypass it;

[0023] S3, when an obstacle appears on the retreat route, distance measurement and object recognition are performed based on the real-time image information of the rear binocular camera, and step S2 is repeated.

[0024] In step S2, the AI ​​model determines that the object information also includes relative dynamic and relative static information based on the data information;

[0025] S2.11: If it is a relatively dynamic object, the vehicle is judged based on its speed. If the speed of the relatively dynamic object is less than the current speed, the vehicle is subjected to perception range identification and trajectory prediction to determine whether the exceeding condition is met. If the speed of the relatively dynamic object is greater than the current speed and the vehicle is relatively static, the vehicle can maintain the current speed and continue driving.

[0026] In step S2, the AI ​​image processing module determines based on the data information that there is also an area with a height difference of more than 6 cm. If this situation is detected within the perception range, step S2 is repeated.

[0027] Compared with the prior art, the beneficial effects of the present invention are:

[0028] It can more intelligently assist people with limited mobility in navigating complex environments and completing daily tasks, making transportation safer and more convenient. The AI ​​language model can replace some human decision-making, making it more intelligent and easier to operate. The overall control process is safer, with multiple sensors and a binocular camera jointly identifying surrounding obstacles and environmental models. A gravity sensor and gyroscope can adjust the seat and backrest angle in real time during operation, providing not only a more comfortable experience for the operator but also greater safety assurance.

[0029] The folding structure of the seat and the backrest allows for folding and storage of the space. The seat and chassis parts can be quickly disassembled, which improves the overall convenience of the device and its easy storage function, thereby broadening the usage scenarios. After separating the seat and chassis, the seat and chassis can be placed in the trunk of a car for separate storage, making it convenient to carry when traveling. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Attachment Figure 1 It is an overall view of the present invention.

[0031] Attachment Figure 2 It is a rear side view of the present invention.

[0032] Attachment Figure 3 It is a partial view of a seat in the present invention.

[0033] Attachment Figure 4 It is a partial view of the regulating structure in the present invention.

[0034] Attachment Figure 5 It is a partial view of the regulating structure in the present invention.

[0035] Attachment Figure 6 This is a partial view of the card seat in the present invention.

[0036] Attachment Figure 7 This is a view of the guide groove and the limit groove in the present invention.

[0037] Reference numerals shown in the accompanying drawings:

[0038] 1. Seat; 2. Armrest; 3. Seat back; 4. Binocular camera; 5. Ultrasonic radar; 6. Connecting seat; 7. First connecting beam; 8. Second connecting beam; 9. Snap-fit ​​seat; 10. Limiting slot; 11. Connecting rod; 12. First electric push rod; 13. Second electric push rod; 14. First linkage member; 15. Second linkage member; 16. Transverse beam; 17. Track seat; 18. Guide slot; 19. Guide column; 20. Connecting shaft; 21. Limiting claw; 22. Snap-fit ​​slot; 23. Quick release switch; 24. Pull wire; 25. Guide plate; 26. Retaining spring; 27. Connecting member. DETAILED DESCRIPTION

[0039] Below in conjunction with specific embodiment, further set forth the present invention.Should be understood that these embodiments are only used to illustrate the present invention and are not used in limiting the scope of the present invention.In addition, should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall equally within the scope limited by the appended claims of the application.

[0040] See the attached drawings in the manual Figure 1-Figure 7 The present invention describes an embodied intelligent mobility robot and its decision-making and control method. The application can be divided into a body part and an embodied perception system part. The body includes a driving chassis, an adjustment structure quickly connected to the driving chassis, a seat 1 articulated to the adjustment structure, an armrest 2 connected to the seat 1, a seat back 3 rotatably connected to the seat 1, a gyroscope (alarm for excessive inclination), and a pressure sensor; a control joystick and a microphone are provided on the armrest 2, and the joystick can be used to control the vehicle body to move forward, backward, and turn, and the passenger can independently choose between intelligent driving and self-operation.

[0041] The driving chassis includes a chassis, multiple driving motors arranged on the chassis, driving wheels connected to the driving motors, and a battery pack arranged on the chassis; the two driving wheels on the front side are omnidirectional wheels, which can realize the functions of turning on the spot and turning around.

[0042] When passing through uphill / downhill sections, in order to improve the stability of the vehicle body, the adjustment structure needs to adjust the angle posture of seat 1 and seat back 3 based on the data feedback from the gyroscope (center of gravity angle position of seat 1) and the pressure sensor (gravity change). For example: when going uphill, the gyroscope detects that the center of gravity angle is offset by 10°. In order to prevent the vehicle body from tilting backward and causing the passengers to feel like they are tipping over, thereby affecting the user experience, the gyroscope combines the pressure sensor data to synchronously fine-tune seat 1 and seat back 3 downward by 5° to 10°. At the same time, in order to prevent the vehicle from tipping over due to excessive slope when going uphill, a pair of anti-roll wheels are also provided on the rear side of the vehicle body.

[0043] The adjustment structure includes a seat adjustment structure and a seat back adjustment structure. The two adjustment structures can also separate the seat 1 and the seat back 3 from the chassis as a whole, and fold the seat back 3 and the seat 1 together, thereby reducing space occupation and facilitating travel and carrying.

[0044] Seat adjustment structure:

[0045] The seat adjustment structure includes two quadrilateral linkage mechanisms and an electric push rod. The quadrilateral linkage mechanism includes a connecting seat 6 connected to the bottom of the seat 1, a first connecting beam 7 hinged to the connecting seat 6 at one end, a second connecting beam 8 hinged to the connecting seat 6 at one end, and a clamping seat 9.

[0046] The first connecting beam 7 is parallel to the second connecting beam 8 . The bottom of the clamping seat 9 is provided with multiple limiting slots 10 . The clamping seat 9 is hinged to the other ends of the first and second connecting beams. The ends of the first and second connecting beams are hinged to the connecting seat 6 and the clamping seat 9 , respectively, to form a quadrilateral linkage mechanism. To automatically control the angle adjustment of the seat 1 , a connecting rod 11 is provided between the two first connecting beams 7 . A first electric push rod 12 is rotatably connected to the connecting rod 11 . The push rod end of the first electric push rod 12 is hinged to the bottom of the seat 1 . When the push rod of the first electric push rod 12 is displaced, the quadrilateral linkage mechanism drives the seat 1 to adjust forward and backward and in height.

[0047] Seat back adjustment structure:

[0048] The backrest adjustment structure includes a second electric push rod 13 hingedly disposed below the seat 1, a first linkage 14 hingedly disposed below the seat 1, and a second linkage 15 hinged to the first linkage 14. The push rod end of the second electric push rod 13 is hinged to the first linkage 14 and the second linkage 15, and the non-hinged end of the second linkage 15 is hinged to the backrest 3. When the push rod of the second electric push rod 13 extends, it drives the first linkage 14 to move toward the seat 1, and at the same time drives the second linkage 15 to move. Since the backrest 3 is hinged to the seat 1, the second linkage 15 rotates synchronously with the second linkage 15 during rotation, thereby causing the backrest 3 to rotate relative to the seat 1 about the hinge point. The seat adjustment structure and the backrest adjustment structure are not entirely based on data adjustment. The armrest 2 of the seat 1 is equipped with buttons, which the user can use to adjust the angle of the seat 1 and the backrest 3.

[0049] The aforementioned quick-release connection between the seat 1 and the driving chassis is actually the quick-release connection between the clamping seat 9 and the driving chassis. The specific structure is as follows:

[0050] Two transverse beams 16 are positioned between the two chassis beams. A connecting shaft 20 is provided on the connecting shaft 20, and two limiting claws 21 are rotatably connected to the connecting shaft 20. Each limiting claw 21 has a slot 22 that mates with the transverse beam 16. A control mechanism is provided on the connecting shaft 9, which drives the two limiting claws 21 to rotate synchronously to achieve limiting and separation. This control mechanism includes a quick-release switch 23 located below the base, a pull wire 24 connected to the quick-release switch 23, a guide plate 25 provided on the connecting shaft 20, a retaining spring 26 provided on one side of the limiting claw 21, and a connector 27 provided on the two limiting claws 21.

[0051] The pull wire 24 is connected to the connecting piece 27, and the connecting piece 27 is connected to the two limit claws 21. When the quick release switch 23 drives the pull wire 24 to contract, the limit claw 21 rotates on the connecting shaft 20, so that the limit claw 21 releases the limit between the transverse beam 16. At this time, the seat 1 can be separated from the driving chassis as a whole. When the quick release switch 23 is released, the pull wire 24 is relaxed and reset under the combined action of its own elastic force and the retaining spring 26.

[0052] If the quick-release structure described above is used, the connection will be unstable and the limiting effect will be poor, affecting the safety of passengers. In addition, after the seat 1 is separated from the chassis, it will not have a guiding function during installation. Therefore, further improvements are required:

[0053] Two track seats 17 are sleeved on the transverse beam 16, and the two track seats 17 are located on both sides of the clamping seat 9. A guide groove 18 is provided on the inner side of the track seat 17, and a guide column 19 is provided on the clamping seat 9 at the bottom of the guide groove 18 to match the guide groove 18. The guide groove 18 limits and guides the guide column 19, making it faster to install the seat 1 between the entire body and the chassis and allowing for quick positioning during installation. In order to further improve the stability of the quick-release structure, further improvements have been made:

[0054] The bottom of the guide groove 18 is an inclined surface. The aforementioned snap-in seat 9 is provided with a plurality of limiting grooves 10. The limiting grooves 10 located at the front face the ground, while the limiting grooves 10 located at the rear face the rear. After the seat 1 is assembled with the base, when the center of gravity of the seat 1 is facing forward, the limiting grooves 10 located at the front and the locking grooves 22 of the limiting claws 21 jointly provide a positional limit due to the inclined bottom of the guide groove 18. When the center of gravity of the seat 1 is facing rearward, the limiting grooves 10 located at the rear, the inclined edges of the guide groove 18, and the limiting claws 21 jointly provide a positional limit, thereby achieving more uniform force distribution and greater stability.

[0055] For this application, the seat 1 and the seat back 3 can be stored and folded by using the seat adjustment structure and the seat back adjustment structure, saving the space occupied by the entire device. The quick-release function greatly improves the convenience of use and easy storage of the entire device, thereby broadening the usage scenarios. After separating the seat 1 and the chassis, the seat 1 and the chassis can be placed in the trunk of the car for separate storage, which is convenient for carrying when traveling.

[0056] Embodied Perception System:

[0057] The embodied perception system includes a plurality of binocular cameras 4 arranged in the front and rear directions of the vehicle body, a plurality of ultrasonic radars 5 arranged on the driving chassis, and a central decision unit;

[0058] The central decision-making unit is connected to an AI large language model and an AI image processing module; the main core components of the central decision-making unit are the above two, and in fact it also includes other components. However, these other components are not innovative structures and are well known to those skilled in the art, so they are only listed below and will not be repeated. The central decision-making unit also includes a computing power integration board, data storage, motor controller, GPS micro positioning module, etc. Satellite positioning is achieved through GNSS+IMU, and synchronous positioning and modeling are performed through the SLAM solution, which is used in areas without positioning.

[0059] The AI ​​language model is connected to the microphone, and the semantic information input by the microphone is recognized and converted, and behavioral decisions are made based on the semantic information. For example, based on the map model information loaded by the central decision-making unit, the microphone is instructed to go to the bedroom, go to a certain park, or turn left. The route is selected based on the environmental map planning and user instructions.

[0060] The central decision unit constructs and updates the map model of the surrounding environment in real time based on multiple ultrasonic radars 5 and monitors the object model information around the vehicle body in real time. The ultrasonic radar 5 is equipped with a shell above the four wheel hubs of the vehicle body and next to the binocular camera 4 on the rear side of the seat 1 to perform real-time monitoring of multiple directions around the vehicle body.

[0061] Two binocular cameras 4 are provided in the forward direction, located under the armrests 2 of the two seats 1, and a binocular camera 4 is provided on the rear side. Multiple binocular cameras 4 capture images in multiple directions and perform distance measurement. The image information is then recognized and analyzed by the AI ​​image processing module for objects and environmental information in the image. The deep learning recognition capability of binocular camera vision is used to simulate a person's "eyes", achieving better accuracy, higher angular resolution, and a smaller size. It is compatible with near- and far-range detection functions and can perform real-time detection of the vehicle's driving environment and other vehicle targets under all working conditions. The YOLO algorithm can accurately identify obstacles, pedestrians, vehicles, traffic signal signs, etc., and detect the distance between the vehicle body and object information. The AI ​​large language model makes behavioral decisions on behalf of the passengers based on the current vehicle speed and detected target information, and controls the vehicle body to act according to the decision information.

[0062] The overall decision-making process is based on a safe, simple, and efficient driving strategy. Since this application is widely applicable to daily life scenarios such as homes, communities, hospitals, airports, and parks, the driving speed needs to be controlled within 10-15 km / h. This can not only meet the needs of transportation but also maintain a safe driving speed. It also has lower requirements for perception hardware, further reducing costs. The specific decision-making control method is as follows:

[0063] S1: When an obstacle appears on the route, distance measurement and object recognition are performed based on the real-time image information provided by the two binocular cameras and ultrasonic radar on the front side;

[0064] S2, the AI ​​image processing module determines whether the object is dynamic or static based on the data information;

[0065] S2.1 If a dynamic object enters the primary perception range, the current vehicle speed is first reduced to level 1. After the speed reduction, the maximum movement time before the safe speed displacement to the dynamic object is calculated; the primary perception range is: 2.5m-3m directly in front, 0-1m wide, and the level 1 speed is 2km / h.

[0066] S2.2 then analyzes the motion trajectory of the dynamic object to determine whether the dynamic object leaves the first-level perception range before reaching the maximum braking distance. If it does not leave, multi-level deceleration is performed according to the maximum movement time until the dynamic object enters the final perception range and then brakes to wait for the dynamic object to leave. The multi-level deceleration process gradually narrows the perception range and gradually reduces the vehicle speed until it stops. The next level of perception range: 1.5-2.5m in front and 0-1m in width. When an obstacle is identified, the speed is reduced to 1km / h; the next level of perception range: 0.3-0.8m in front and 0-1m in width. When an obstacle is identified, the speed is gradually reduced to 0.5km / h; the final level of perception range: 0-0.3m in front and 0.65m in width. The entire vehicle is braked to a stop.

[0067] Throughout the entire decision-making and control process, the vehicle maintains a safe speed and distance, gradually decelerating through multiple levels of perception. The vehicle remains in motion throughout the entire process, avoiding sudden stops. This reduces rider frustration and ensures both safe and comfortable driving. The perception range must be bidirectional: forward and left / right distance. A single-directional range may not be sufficient to determine vehicle width or allow passage through a hallway.

[0068] S2.3 If it is a static object, determine the object based on the spatial information surrounding it and choose to bypass it;

[0069] In step S3, if an obstacle appears on the reverse path, the vehicle performs distance measurement and object recognition based on real-time image information from the rear binocular camera, and repeats step S2. The sensing range and vehicle speed in the reverse direction change as follows: if an obstacle is detected within a sensing range of 0.5-1m and a width of 0-1m, the vehicle's speed gradually decreases to 1km / h. If an obstacle is detected within a sensing range of 0.1-0.3m and a width of 0.8m, the vehicle's speed gradually decreases to 0.5km / h. The vehicle stops completely when an obstacle appears within a sensing range of 0.1m and a width of 0.65m.

[0070] In step S2, the AI ​​model determines that the object information includes not only dynamic and static information but also relative dynamic and relative static information based on the data information;

[0071] S2.11: If it is a relatively dynamic object, the vehicle is judged based on its speed. If the speed of the relatively dynamic object is less than the current speed, the vehicle is subjected to perception range identification and trajectory prediction to determine whether the exceeding condition is met. If the speed of the relatively dynamic object is greater than the current speed and the vehicle is relatively static, the vehicle can maintain the current speed and continue driving.

[0072] During driving, there may be a situation where there is a height difference of more than 6cm, such as a pit in the road or a staircase. In this case, you need to choose to detour or brake. If the detour conditions are met within the sensing range, the detour is chosen. If the detour conditions are not met, the brake is required. There are two situations: Cliff ahead: During forward movement, the front end of the vehicle will be 15cm away from the cliff to achieve intelligent braking. Within the sensing range of 1 to 2 meters in front and a width of 0 to 1 meter, if the cliff is detected, the speed will be reduced to 1km / h. Within the sensing range of 0.5 to 1 meter in front and a width of 0 to 0.8 meters, if the cliff is detected, the speed will be reduced to 0.5km / h. The entire vehicle will brake until the front edge of the vehicle is within 0 to 0.15m and a width of 0.8m. There is a cliff behind the vehicle: During forward movement, the rear end of the vehicle will be 15cm away from the cliff to achieve intelligent braking. If a cliff is detected within a sensing range of 0.5 to 1.5 meters and 0 to 1 meter wide behind the vehicle, the vehicle's speed will be reduced to 1 km / h. If a cliff is detected within a sensing range of 0.15 to 0.5 meters and 0 to 0.8 meters wide behind the vehicle, the vehicle's speed will be reduced to 0.5 km / h. The vehicle will stop completely when the front edge of the vehicle is within a sensing range of 0 to 0.15 meters and 0.8 meters wide. This prevents excessive height differences that could affect driving safety, such as falls.

Claims

1. An embodied intelligent mobility robot, characterized by: Including the vehicle body and the embodied sensing system mounted on the vehicle body; The vehicle body comprises a driving chassis, an adjustment structure connected to the driving chassis in a quick-release manner, a seat (1) hinged to the adjustment structure, an armrest (2) connected to the seat (1), a seat back (3) rotatably connected to the seat (1), a gyroscope, and a pressure sensor; The armrest (2) is provided with a microphone, the driving chassis comprises a chassis, a plurality of driving motors arranged on the chassis, driving wheels connected to the driving motors, and a battery pack arranged on the chassis, the adjusting structure automatically adjusts the angle posture of the seat (1) and the backrest (3) according to the angular position of the center of gravity of the seat (1) and the change of gravity, and the adjusting structure can also fold the seat (1) and the backrest (3) for placement; the adjusting structure comprises a seat adjusting structure and a backrest adjusting structure; the seat adjusting structure comprises two quadrilateral connecting rod mechanisms, a first electric push rod (12) connected to the quadrilateral connecting rod mechanism, the quadrilateral connecting rod mechanism comprises a connecting seat (6) connected to the bottom of the seat (1), a first connecting beam (7) hinged to the connecting seat (6) at one end, a second connecting beam (8) hinged to the connecting seat (6) at one end, and a clamping seat (9); the first connecting beam (7) is parallel to the second connecting beam (8), a plurality of limiting grooves (10) are provided at the bottom of the said clamping seat (9), and the said clamping seat (9) is hinged to the other end of the first and second connecting beams; a connecting rod (11) is provided between the two first connecting beams (7), and a first electric push rod (12) is rotatably connected to the said connecting rod (11), and the push rod end of the first electric push rod (12) is hinged to the bottom of the seat (1); the said chair back adjustment structure includes a second electric push rod (13) hingedly arranged under the seat (1), a first linkage (14) hingedly arranged under the seat (1), and a second linkage (15) hinged to the first linkage (14); the push rod end of the second electric push rod (13) is hinged to the first linkage (14) and the second linkage (15), and the non-hinged end of the second linkage (15) is hinged to the chair back (3); The embodied perception system includes a plurality of binocular cameras (4) arranged in the front and rear directions of the vehicle body, a plurality of ultrasonic radars (5) arranged on the driving chassis, and a central decision unit; The central decision-making unit is connected to an AI large language model and an AI image processing module; the AI ​​large language model is connected to a microphone, and the semantic information input by the microphone is recognized and converted and behavioral decisions are made based on the semantic information; The central decision unit constructs and updates the map model of the surrounding environment in real time based on multiple ultrasonic radars (5) and monitors the object model information around the vehicle in real time, and selects a path based on the environmental map planning and user instructions; Multiple binocular cameras (4) capture images in multiple directions and perform distance measurement. The image information is then used by the AI ​​image processing module to identify and analyze the objects and environmental information in the image, and detect the distance between the vehicle body and the object information. The AI ​​large language model makes behavioral decisions on behalf of the passengers based on the current vehicle speed and detected target information, and controls the vehicle body to act according to the decision information.

2. The embodied intelligent mobility robot according to claim 1, characterized in that: Two transverse beams (16) are provided between the two beam frames of the chassis, two track seats (17) are sleeved on the transverse beams (16), the track seats (17) are provided with guide grooves (18), and the clamping seat (9) is provided with guide posts (19) adapted to the guide grooves (18); The clamping seat (9) is provided with a connecting shaft (20), and two limiting claws (21) are rotatably connected on the connecting shaft (20), and a clamping groove (22) is provided on the limiting claws (21), and the clamping groove (22) is adapted to the transverse beam (16). The clamping seat (9) is provided with a control mechanism, and the control mechanism drives the two limiting claws (21) to rotate synchronously to achieve limiting and separation.

3. The embodied intelligent mobility robot according to claim 2, characterized in that: The control mechanism comprises a quick-release switch (23) provided below the base, a pull wire (24) connected to the quick-release switch (23), a guide plate (25) provided on the connecting shaft (20), a retaining spring (26) provided on one side of the limiting claw (21), and a connecting piece (27) provided on the two limiting claws (21); The pull wire (24) is connected to a connecting member (27), and the connecting member (27) is connected to two limit claws (21). When the quick release switch (23) drives the pull wire (24) to retract, the limit claw (21) rotates on the connecting shaft (20). When the quick release switch (23) is released, the retaining spring (26) drives the limit claw (21) to reset.

4. The embodied intelligent mobility robot according to claim 3, characterized in that: The bottom of the guide groove (18) is an inclined surface, the notch of the limiting groove (10) located on the front side faces the ground, and the notch of the limiting groove (10) located on the rear side faces the rear side.

5. The decision-making and control method for an embodied intelligent mobility robot according to any one of claims 1 to 4, characterized in that: S1, when an obstacle appears on the route, distance measurement and object recognition are performed based on the real-time image information provided by the two binocular cameras (4) and ultrasonic radar (5) on the front side; S2, the AI ​​image processing module determines whether the object is dynamic or static based on the data information; S2.1 If a dynamic object enters the primary sensing range, the vehicle's current speed is first reduced to level 1. After the speed reduction, the vehicle's maximum motion time before reaching the safe speed limit is calculated. S2.2 then analyzes the motion trajectory of the dynamic object to determine whether it leaves the primary sensing range before reaching the maximum braking distance. If it does not, the system decelerates in multiple stages and reduces the sensing range based on the maximum motion time until it enters the final sensing range, where it brakes to a stop. S2.3 If it is a static object, identify it based on the surrounding sensing range and choose to bypass it; S3, when an obstacle appears on the retreat route, distance measurement and object recognition are performed based on the real-time image information of the rear binocular camera (4), and the deceleration process in step S2 is repeated.

6. The decision-making and control method for an embodied intelligent mobility robot according to claim 5, characterized in that: In step S2, the AI ​​model determines that the object information also includes relative dynamic and relative static information based on the data information; S2.11, if it is a relatively dynamic object, determine the speed of the relatively dynamic object. If the speed of the relatively dynamic object is less than the current speed, perform perception range identification and trajectory prediction to determine whether the exceeding condition is met. If the relative dynamic object is greater than the current speed and is in a relatively static state, maintain the current speed and continue driving.

7. The decision-making and control method for an embodied intelligent mobility robot according to claim 6, characterized in that: In step S2, the AI ​​image processing module determines based on the data information that the process also includes an area with a height difference of more than 6 cm. If this situation is detected within the perception range, the speed reduction process in step S2 is repeated.

Images