An omnidirectional mobile projection type AI education electronic pet robot

Abstract

The application discloses an omnidirectional mobile projection type AI education electronic pet robot, relates to the technical field of AI education robots, and comprises a bionic trunk shell and a bionic head shell arranged at the top end of the bionic trunk shell. The bottom of the bionic trunk shell is provided with a chassis, and a plurality of Mecanum wheels are arranged below the chassis. The chassis is provided with an intelligent computing mechanism. The application has the beneficial effects that: through the setting of the chassis, the plurality of Mecanum wheels, the heterogeneous computing core in the intelligent computing mechanism and the projection module in the bionic head shell, when the robot works, the heterogeneous computing core realizes real-time SLAM space sensing and human body following control, drives the Mecanum wheels to realize translation, rotation and autonomous obstacle avoidance navigation in any direction, and simultaneously, the projection module moves to the desktop in front of the user or the wall surface nearby to realize dynamic projection.

Description

Technical Field

[0001] This invention relates to the field of AI educational robot technology, specifically to an omnidirectional mobile projection-type AI educational electronic pet robot. Background Technology

[0002] Most home education robots on the market are either fixed-base smart terminals or have limited mobility, typically integrating voice interaction modules, touch screens, and pre-installed learning applications. These robots often have a cartoonish or minimalist technological design and can play educational videos, conduct simple question-and-answer interactions, or assist children in completing exercises. Meanwhile, the projection devices used in home settings are mainly standalone mini projectors, often placed on a table or fixedly mounted on a wall or ceiling, used in conjunction with a screen or white wall to magnify and display educational content, cartoons, or online courses.

[0003] Existing home education robots typically rely on user-triggered interaction: children wake them up by voice or select learning content by touching the screen, and the robot plays courses, asks questions, and judges whether the answers are correct according to a preset logic process. Some robots have simple following functions, but most rely on infrared or ultrasonic sensors to achieve rough movement and cannot perform precise spatial positioning and autonomous navigation. Independent projection devices require pre-fixed positions and manual focusing and correction of trapezoidal distortion. When in use, the content of a computer, mobile phone, or tablet is projected onto a wall or screen via wired or wireless means, and the image remains static on that plane and cannot adjust the projection angle or position with the user's movement.

[0004] Therefore, existing educational robots have limited mobility, mostly being fixed or only able to move in a straight line, unable to achieve omnidirectional flexible following and indoor autonomous obstacle avoidance navigation, resulting in a narrow interaction range. In addition, the corresponding projection equipment needs to be fixedly installed and relies on a flat wall or special screen, unable to achieve dynamic distortion-free projection on any plane, and cannot move and project as the user's position changes. Summary of the Invention

[0005] The purpose of this invention is to provide an omnidirectional, projection-type AI educational electronic pet robot to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an omnidirectional mobile projection-type AI educational electronic pet robot, comprising a bionic torso shell and a bionic head shell disposed on the top of the bionic torso shell, a chassis disposed at the bottom of the bionic torso shell, and multiple Mecanum wheels disposed below the chassis; an intelligent computing mechanism, comprising a heterogeneous computing core disposed inside the bionic torso shell, interactive modules disposed on both sides of the bionic torso shell, a projection module and a sensing unit disposed inside the bionic head shell, the sensing unit comprising a depth camera and an inertial measurement unit, and cleaning components disposed around the projection module.

[0007] Preferably, there are two chassis, which are respectively located on both sides of the bottom of the bionic torso shell. One of the chassis has a Mecanum wheel installed under it that is a left-handed wheel, and the other chassis has a Mecanum wheel installed under it that is a right-handed wheel.

[0008] Preferably, the heterogeneous computing core includes an NPU chip and an FPGA chip, which are electrically connected via a PCIe bus.

[0009] Preferably, the interaction module includes an LED light strip and a touch sensor, the touch sensor covering the surface of the annular LED light strip.

[0010] Preferably, the projection module is a DLP micro laser projection component and is equipped with a TOF autofocus sensor.

[0011] Preferably, the cleaning component includes a protective shell, which is rectangular in shape and is fixedly connected to the surface of the bionic head shell.

[0012] Preferably, a lifting block is slidably connected to the inner wall of the protective shell, and multiple nozzles are installed inside the lifting block, with the air outlet of the nozzles located above the lifting block.

[0013] Preferably, an air pump is installed on one side of the lifting block, and multiple nozzles are connected to the air pump through air pipes. Multiple air vents are provided on the side of the protective shell near the air pump.

[0014] Preferably, a winding rod is rotatably connected to the inner top wall of the protective shell, and a servo motor is installed inside the protective shell, with the drive end of the servo motor fixedly connected to one end of the winding rod.

[0015] Preferably, a dustproof cloth is wound around the surface of the winding rod, and the bottom end of the dustproof cloth is fixedly connected to the upper surface of the lifting block.

[0016] The beneficial effects of this invention are:

[0017] 1. By setting up a chassis, multiple Mecanum wheels, a heterogeneous computing core in the intelligent computing mechanism, and a projection module inside the bionic head shell, when the robot is working, the heterogeneous computing core performs real-time SLAM spatial perception and human following control, driving the Mecanum wheels to achieve translation, rotation, and autonomous obstacle avoidance navigation in any direction. At the same time, the projection module moves with the robot to the user's table or nearby wall for dynamic projection. This optimizes the problems of limited mobility of existing educational robots, inability to follow flexibly in all directions, and the need for fixed installation of projection equipment that depends on a flat wall and cannot move and project according to the user's position. It achieves the effect of expanding the interaction range and realizing an immersive teaching experience of "the screen follows the person and projection is distortion-free on any plane".

[0018] 2. By setting up a cleaning component, when the robot is in non-projection mode, the servo motor drives the winding rod to reverse, causing the dustproof cloth to unfold downwards and cover the light outlet of the projection module. At the same time, the lifting block drives the bottom of the dustproof cloth to descend to the bottom of the protective shell, forming a closed dustproof barrier. This effectively reduces the contact of external dust, lint, and foreign objects with the projection lens, reduces lens contamination or scratches, and maintains the clarity of the projected image and the lifespan of the module for a long time.

[0019] 3. By setting up a lifting block, multiple nozzles, an air pump, and a dustproof cloth, when the robot needs to project, the servo motor drives the winding rod to rotate forward, causing the dustproof cloth to roll upward. The lifting block rises synchronously, and the air pump supplies air to multiple nozzles, causing the nozzles to spray high-speed airflow onto the surface of the rising dustproof cloth and the area around the projection window. The airflow blows away the fine dust and fibers attached to the dustproof cloth or projection window. This achieves automatic cleaning before the projection window is opened, without manual intervention, improves the effect of each projected image, and enhances user convenience and the presentation quality of educational content. Attached Figure Description

[0020] Figure 1 This is a structural diagram of an omnidirectional projection-type AI educational electronic pet robot.

[0021] Figure 2 This is a diagram of the bottom structure of an omnidirectional projection-type AI educational electronic pet robot.

[0022] Figure 3 This is a schematic diagram showing the installation of the projection module and cleaning components of an omnidirectional, projection-type AI educational electronic pet robot.

[0023] Figure 4 This is a schematic diagram of the internal structure of the cleaning component of an omnidirectional projection-type AI educational electronic pet robot.

[0024] Figure 5This is a schematic diagram showing the connection of the lifting block, dust cover, and winding rod of an omnidirectional projection-type AI educational electronic pet robot.

[0025] Figure 6 This is an architecture diagram of an omnidirectional, projection-based AI educational electronic pet robot.

[0026] Figure 7 This is a software flowchart for an omnidirectional, projection-based AI educational electronic pet robot.

[0027] The following numbers are labeled in the diagram: 100, Bionic torso shell; 200, Bionic head shell; 300, Chassis; 310, Mecanum wheel; 400, Intelligent computing mechanism; 410, Heterogeneous computing core; 420, Interaction module; 430, Projection module; 440, Cleaning component; 441, Protective shell; 442, Lifting block; 443, Dustproof cloth; 444, Winding rod; 445, Servo motor; 446, Nozzle; 447, Air pump. Detailed Implementation

[0028] 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Example 1:

[0030] Reference Figures 1 to 7 This is the first embodiment of the present invention, which provides an omnidirectional mobile projection-type AI educational electronic pet robot, including a bionic torso shell 100 and a bionic head shell 200 disposed on the top of the bionic torso shell 100. A chassis 300 is disposed at the bottom of the bionic torso shell 100, and multiple Mecanum wheels 310 are disposed below the chassis 300. An intelligent computing mechanism 400 is provided, which includes a heterogeneous computing core 410 disposed inside the bionic torso shell 100. Interaction modules 420 are disposed on both sides of the bionic torso shell 100. A projection module 430 and a sensing unit are disposed inside the bionic head shell 200. The sensing unit includes a depth camera and an inertial measurement unit. A cleaning component 440 is disposed around the projection module 430.

[0031] Example 2:

[0032] Reference Figure 1 , Figure 2 , Figure 6 as well as Figure 7 This is the second embodiment of the present invention, which is based on the previous embodiment.

[0033] Specifically, there are two chassis 300s, each located on one side of the bottom of the bionic torso shell 100. One chassis 300 has a left-handed Mecanum wheel 310 mounted underneath it, while the other chassis 300 has a right-handed Mecanum wheel 310 mounted underneath it. The heterogeneous computing core 410 includes an NPU chip and an FPGA chip, which are electrically connected via a PCIe bus. The interaction module 420 includes an LED light strip and a touch sensor, which covers the surface of the ring-shaped LED light strip. The projection module 430 is a DLP micro-laser projection assembly equipped with a TOF autofocus sensor.

[0034] In this embodiment, the perception unit includes a depth camera mounted on the bionic head shell 200 and a nine-axis inertial measurement unit (IMU) mounted inside the bionic torso shell 100. When the robot is activated, the depth camera acquires RGB images and depth information from the environment in real time, obtaining the user's position, obstacle distances, and 3D point cloud data of the projected target plane. Simultaneously, the IMU samples the robot's three-axis acceleration, angular velocity, and magnetometer data at high frequency to estimate the robot's real-time attitude and trajectory. The FPGA in the heterogeneous computing core 410 filters, denoises, and performs synchronous preprocessing on the raw data from the depth camera and IMU. Subsequently, the NPU chip runs ORB-SLAM3 or RTAB-Map visual SLAM algorithms to fuse the depth images with the IMU data, constructing a 3D point cloud map of the environment and calculating the robot's six-degree-of-freedom pose. Based on the map and pose, SLAM... The AM module outputs the robot's precise position relative to the environment and identifies passable areas and applicable projection surfaces. In human-following mode, after the depth camera detects the target user, the NPU runs the YOLO target detection and DeepSORT tracking algorithms. Combined with the robot's own motion information provided by the IMU, it calculates the distance and azimuth angle of the target relative to the robot, thereby generating motion control commands. In dynamic projection correction mode, the depth camera acquires the 3D point cloud of the target projection plane, extracts the plane's normal vector and region range through plane fitting, and combines the robot's current pose provided by the IMU and SLAM. The FPGA calculates the homography transformation matrix in real time to achieve distortion-free pre-deformation of the projected image. Throughout the process, the depth camera and IMU work together to provide the necessary environmental perception and self-positioning information for the robot's omnidirectional movement, human-following, autonomous obstacle avoidance, and dynamic projection.

[0035] The robot integrates a real-time SLAM mapping and path planning module. When the robot first enters a home environment, it acquires omnidirectional environmental depth data through a depth camera, runs the ORB-SLAM3 or RTAB-Map visual SLAM framework, and performs fusion localization in conjunction with a nine-axis IMU to construct a 3D point cloud map of the environment and mark furniture areas, movable spaces, and applicable projection areas. Based on the map, the robot uses the A* or Dijkstra algorithm for global path planning and utilizes the Dynamic Window (DWA) method for local real-time obstacle avoidance to ensure smooth obstacle avoidance while following the user and guarantee walking safety.

[0036] The dynamic projection calibration algorithm is another core technology of this invention: the depth camera acquires the three-dimensional point cloud of the target projection plane in real time, extracts the position, normal vector and projection area range of the plane through plane fitting, and combines the robot's current pose provided by SLAM. The FPGA chip calculates the homography transformation matrix or surface deformation mesh in real time at a frequency of 30Hz, performs pre-deformation processing on the educational content image, and then sends it to the DLP projection module 430 for projection, thereby achieving distortion-free projection on any inclined plane, curved surface or even non-flat surface, and the projected image remains stable during the robot's following motion.

[0037] The heterogeneous computing power scheduling mechanism ensures the real-time performance of each task: the NPU chip is dedicated to running deep neural network inference tasks such as YOLOv11 object detection, gesture recognition, sentiment analysis, and voice interaction; the FPGA chip is dedicated to executing four-wheel motion control PID closed loop, real-time distortion correction of projected images, and depth image preprocessing; while the CPU core is responsible for SLAM mapping, path planning, and educational resource recommendation algorithms, and the small core handles system scheduling and peripheral drivers; each unit exchanges data with zero copy through shared memory, and with dynamic load balancing, the robot can maintain low latency response while performing omnidirectional following, projection correction, and AI interaction;

[0038] Each Mecanum wheel 310 is directly driven by an independent DC geared motor, with four motors connected one-to-one with each of the four Mecanum wheels 310. The motor driver receives PWM pulse width modulation signals from the FPGA chip and adjusts the PWM duty cycle to change the motor speed. Simultaneously, a Hall encoder mounted on the motor shaft detects the actual speed in real time and feeds it back to the FPGA. The FPGA runs a PID closed-loop control algorithm, converting the deviation between the desired speed and the actual speed into a PWM adjustment amount, thereby achieving precise control of the speed of each Mecanum wheel 310. Based on this, the robot calculates the speed and direction commands of each of the four wheels in real time according to the desired direction of motion using the inverse kinematics algorithm of the Mecanum wheel 310, and sends them to the four motor drivers respectively, so that each wheel rotates in coordination with different speeds and directions. Utilizing the resultant force formed by the oblique rollers unique to the Mecanum wheel 310, the robot can move in any direction.

[0039] Example 3:

[0040] Reference Figure 3 , Figure 4 as well as Figure 5 This is the third embodiment of the present invention, which is based on the first two embodiments.

[0041] Specifically, the cleaning component 440 includes a protective shell 441, which is rectangular in shape and fixedly connected to the surface of the bionic head shell 200. A lifting block 442 is slidably connected to the inner wall of the protective shell 441. Multiple nozzles 446 are installed inside the lifting block 442, with their outlets located above it. An air pump 447 is installed on one side of the lifting block 442, and the multiple nozzles 446 are connected to the air pump 447 via air pipes. Multiple air vents are located on the side of the protective shell 441 near the air pump 447. A winding rod 444 is rotatably connected to the inner top wall of the protective shell 441. A servo motor 445 is installed inside the protective shell 441, and its drive end is fixedly connected to one end of the winding rod 444. A dustproof cloth 443 is wound around the surface of the winding rod 444, and the bottom end of the dustproof cloth 443 is fixedly connected to the upper surface of the lifting block 442.

[0042] In this embodiment, the cleaning component 440 begins operation: In the non-projection state, the lifting block 442, which is slidably connected to the inner wall of the protective shell 441 in the cleaning component 440, is in a low position. The servo motor 445 drives the winding rod 444 to reverse, causing the dustproof cloth 443 to unfold downwards and cover the light outlet of the projection module 430, forming a dustproof barrier. When projection is required, the servo motor 445 drives the winding rod 444 to rotate forward, the dustproof cloth 443 rolls up, and the lifting block 442 rises synchronously. At the same time, multiple nozzles 446 installed inside the lifting block 442 are connected to the air pump 447 through air pipes. The air pump 447 starts working and supplies air to the nozzles 446. The air outlets of the nozzles 446 are located above the lifting block 442, spraying high-speed airflow onto the surface of the rising dustproof cloth 443 and the area around the projection window, blowing away the dust adhering to the dustproof cloth 443 or the projection window. After cleaning away dust and fibers, the projection module 430 begins operation; the lifting block 442 descends naturally under gravity, and the sliding mating surface is made of a low-friction material; the projection module 430 is a DLP micro laser projection component equipped with a TOF autofocus sensor, which can project educational content onto the target plane without distortion; at the same time, the interaction module 420, located on both sides of the bionic torso shell 100, includes an LED light strip and a touch sensor covering the surface of the ring-shaped LED light strip. The LED light strip displays the robot's working status or emotional expression, and the touch sensor detects the user's touch, stroking, or patting actions, realizing emotional interaction between humans and robots; thus, in the process of the robot moving omnidirectionally to follow the user, actively cleaning the projection window, and projecting educational content, an immersive teaching experience of "the screen follows the person and the environment is the screen" is achieved.

[0043] When the robot is powered on, the heterogeneous computing core 410 in the intelligent computing mechanism 400 inside the bionic torso shell 100 begins operation. This heterogeneous computing core 410 includes an NPU chip and an FPGA chip electrically connected via a PCIe bus. The NPU chip runs AI algorithms such as YOLO object detection, SLAM visual localization, emotion recognition, and voice interaction, while the FPGA chip performs motion control and sensor data preprocessing. The robot acquires environmental depth images and the user's position using a depth camera mounted on the bionic head shell 200, and achieves omnidirectional movement using multiple Mecanum wheels 310 located under the chassis 300. Two chassis 300s are located on either side of the bottom of the bionic torso shell 100; one chassis 300 is equipped with a left-handed Mecanum wheel 310, and the other with a right-handed Mecanum wheel 310. Under the control of the heterogeneous computing core 410, the robot can perform real-time human following, autonomous obstacle avoidance navigation, and translation and rotation in any direction. When the robot follows the user to the front table or nearby... Behind the wall, the cleaning component 440 automatically opens and cleans the projection window. Then, the projection module 430 starts working. This projection module 430 is a DLP micro laser projection component equipped with a TOF autofocus sensor, which can correct distortion in real time according to the projection surface, projecting educational content onto the target plane without distortion. At the same time, the interaction module 420, located on both sides of the bionic torso shell 100, includes an LED light strip and a touch sensor covering the surface of the ring-shaped LED light strip. The LED light strip displays the robot's working status or emotional expression, and the touch sensor detects the user's touch, stroking, or patting actions. Combined with gestures recognized by the depth camera and voice commands collected by the microphone array, the NPU chip in the heterogeneous computing core 410 performs multimodal emotion analysis and intent understanding in real time, and actively pushes educational content suitable for the user's current state, realizing an intelligent teaching mode of "information finding people". Thus, in the process of the robot's omnidirectional movement and following, dynamic projection, and active interaction, it provides users with an immersive and personalized educational companionship experience.

[0044] It should be noted that the Mecanum wheel 310, air pump 447, nozzle 446 and servo motor 445 mentioned above are all devices with relatively mature existing technology. The specific model can be selected according to actual needs. At the same time, the air pump 447 and servo motor 445 are powered by built-in power supplies. The specific power supply method can be selected according to the situation, which will not be elaborated here.

[0045] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0046] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An omnidirectional, projection-based AI educational electronic pet robot, characterized in that: include: The bionic torso shell (100) and the bionic head shell (200) disposed at the top of the bionic torso shell (100) are provided with a chassis (300) at the bottom of the bionic torso shell (100) and a plurality of Mecanum wheels (310) are disposed below the chassis (300). The intelligent computing mechanism (400) includes a heterogeneous computing core (410) disposed inside a bionic torso shell (100), an interaction module (420) disposed on both sides of the bionic torso shell (100), a projection module (430) and a sensing unit disposed inside the bionic head shell (200), the sensing unit including a depth camera and an inertial measurement unit, and a cleaning component (440) disposed around the projection module (430).

2. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 1, characterized in that: The number of chassis (300) is two and they are respectively set on both sides of the bottom of the bionic torso shell (100). One of the chassis (300) has a Mecanum wheel (310) installed under it that is a left-handed wheel, and the other chassis (300) has a Mecanum wheel (310) installed under it that is a right-handed wheel.

3. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 1, characterized in that: The heterogeneous computing core (410) includes an NPU chip and an FPGA chip, which are electrically connected via a PCIe bus.

4. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 1, characterized in that: The interaction module (420) includes an LED light strip and a touch sensor, which covers the surface of the annular LED light strip.

5. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 1, characterized in that: The projection module (430) is a DLP micro laser projection component and is equipped with a TOF autofocus sensor.

6. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 1, characterized in that: The cleaning component (440) includes a protective shell (441) which is rectangular in shape and is fixedly connected to the surface of the bionic head shell (200).

7. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 6, characterized in that: The inner wall of the protective shell (441) is slidably connected to a lifting block (442), and multiple nozzles (446) are installed inside the lifting block (442). The air outlet of the nozzles (446) is located above the lifting block (442).

8. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 7, characterized in that: An air pump (447) is installed on one side of the lifting block (442), and multiple nozzles (446) are connected to the air pump (447) through air pipes. Multiple air vents are provided on the side of the protective shell (441) near the air pump (447).

9. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 8, characterized in that: The inner top wall of the protective shell (441) is rotatably connected to a winding rod (444), and a servo motor (445) is installed inside the protective shell (441). The drive end of the servo motor (445) is fixedly connected to one end of the winding rod (444).

10. The omnidirectional mobile projection-type AI educational electronic pet robot according to claim 9, characterized in that: The surface of the winding rod (444) is wrapped with a dustproof cloth (443), and the bottom end of the dustproof cloth (443) is fixedly connected to the upper surface of the lifting block (442).

Images