Hot pot robot system

Abstract

The invention relates to the technical field of catering service, and provides a chafing dish robot system which comprises a mechanical arm module, a first mechanical arm, a second mechanical arm, a control module and a control module, the first mechanical arm executes dish scalding operation and kitchen ware control action, and the second mechanical arm executes dish clamping operation and dish placing action; the multi-sensor module comprises a visual perception unit for collecting dish images and hot pot positions, a thermal inductance monitoring unit for monitoring the surface temperature of the hot pot in real time, and a liquid level monitoring unit for detecting the liquid level of soup; the central controller is used for identifying dish types and dish space poses, automatically determining instant-boiling time, generating soup adjustment and fire adjustment instructions and planning a cooperative movement path of the mechanical arm; the firepower control unit is used for adjusting the firepower gear of the hot pot according to the firepower gear adjusting instruction; and the man-machine interaction module is used for receiving a user instruction and feeding back a system execution state. According to the method, the self-adaptive regulation and control of the hot pot state and the personalized user experience optimization can be ensured while the cooking efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of catering services, and in particular to a hot pot robot system. Background Art

[0002] In recent years, with the accelerated advancement of smart technology in my country's catering industry, some restaurants have introduced service robots to perform basic tasks such as in-restaurant food delivery and tableware collection. In the hot pot sector, a few technical solutions attempt to use single-arm mechanical structures to achieve targeted dish delivery or simple soup base replenishment. These solutions employ preset motion trajectories to transfer ingredients, or rely on manually preset timings to complete single steps.

[0003] However, existing hot pot robots generally have functional limitations and interaction defects: First, single-arm structures or non-cooperative dual-arm systems can only perform sequential operations. For example, they need to wait for manual instructions to fish out after gripping, and cannot achieve parallel operations of boiling and fishing, resulting in low efficiency; second, they rely on fixed-angle cameras or preset parameters, which makes it difficult to capture changes in soup level, fluctuations in fire intensity and the doneness of dishes in real time, which easily leads to the risk of drying out or overcooking; third, the environmental perception dimension is single, such as only vision or timer, lacking multi-source sensor fusion capabilities, and unable to adjust strategies according to the dynamic working conditions of the hot pot; fourth, human-computer interaction is rigid, users need to manually set parameters and the system has no learning ability, making it difficult to meet personalized taste requirements.

[0004] Therefore, there is an urgent need for a hot pot robot system that can improve cooking efficiency while ensuring adaptive control of the hot pot state and personalized user experience optimization. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention provides a hot pot robot system that can improve cooking efficiency while ensuring adaptive control of hot pot status and personalized user experience optimization.

[0006] The present invention provides a hot pot robot system, comprising the following modules:

[0007] The robotic arm module includes a first robotic arm and a second robotic arm, wherein the first robotic arm is configured to perform dish blanching operations and kitchenware manipulation actions, and the second robotic arm is configured to perform dish picking operations and dish arrangement actions;

[0008] The multi-sensor module includes a visual perception unit, a thermal monitoring unit, and a liquid level monitoring unit. The visual perception unit is used to collect images of dishes and hot pot location information, the thermal monitoring unit is used to monitor the hot pot surface temperature in real time, and the liquid level monitoring unit is used to detect the hot pot soup level in real time.

[0009] The central controller is used to identify the dish category and spatial position based on the dish image; generate soup adjustment instructions and heat level adjustment instructions based on the hot pot soup level and hot pot surface temperature; determine the scalding time based on the hot pot surface temperature and heat level; and generate the coordinated motion path of the robotic arm module;

[0010] A firepower control unit is used to adjust the firepower level of the hot pot according to the firepower level adjustment instruction;

[0011] The human-computer interaction module is used to receive user instructions and provide real-time feedback on the system execution status.

[0012] Furthermore, the dish category and spatial position are identified, including:

[0013] Control the visual perception unit to shoot the food picking area, extract the current frame feature descriptor based on the shot image, identify the food based on the cosine similarity between the current frame feature descriptor and the pre-stored template feature descriptor, and output the food category and food spatial pose when the similarity exceeds the preset threshold.

[0014] Furthermore, the process of generating the soup adjustment instruction specifically includes:

[0015] Control the visual perception unit to perform real-time depth scanning of the hot pot to obtain a scene depth map;

[0016] Based on the scene depth map, a segmentation algorithm is used to separate the hot pot area from the desktop background, and the hot pot center coordinates and hot pot radius are calculated;

[0017] Determine the inner area of ​​the hot pot edge, the depth of the hot pot edge, and the liquid level depth based on the hot pot center coordinates and the hot pot radius;

[0018] Estimate the current liquid level based on the inner area of ​​the hot pot edge, the depth of the hot pot edge, and the liquid surface depth;

[0019] Compare the current liquid level with the soup threshold. If it is less than the soup threshold, generate a water-adding instruction.

[0020] Furthermore, the system also includes a water-adding robotic arm, which is used to automatically perform water-adding actions after receiving a water-adding instruction.

[0021] Furthermore, the coordinated motion path includes a scalding motion path of the first robotic arm and a clamping motion path of the second robotic arm.

[0022] Furthermore, the scalding motion path of the first robotic arm includes:

[0023] A straight line interpolation path from the dish placement position to the edge of the hot pot;

[0024] The periodic shaking path of the arc trajectory in the hot pot area;

[0025] The lifting path after scalding.

[0026] Furthermore, the gripping motion path of the second robotic arm includes:

[0027] During the scalding process, the first robotic arm is pre-positioned to the preset standby coordinates;

[0028] Generate a Cartesian straight line scooping path based on the estimated posture of the food in the pot;

[0029] After taking out the food, lift it to a fixed height to avoid the edge of the pot and move it to the plate position.

[0030] Furthermore, the collaborative motion path also includes implementing collision avoidance control, including: using a local obstacle avoidance algorithm to detect in real time whether a conflict occurs in the robot arm path, and achieving dynamic collision avoidance by adjusting the preset standby coordinates of the second robot arm and / or the scalding trajectory of the first robot arm.

[0031] Furthermore, the scalding time is determined according to the surface temperature of the hot pot and the fire level, including:

[0032] Call preset recommended blanching time based on dish type;

[0033] The preset recommended scalding time is proportionally corrected according to the real-time hot pot surface temperature and the fire level to obtain the corrected scalding time;

[0034] The corrected blanching time is corrected again based on the doneness preference set by the user to obtain the final blanching time.

[0035] Furthermore, the preset recommended boiling time is proportionally corrected according to the real-time hot pot surface temperature and the fire level, and the corrected boiling time includes:

[0036]

[0037] Among them, t adj Indicates the corrected scalding time, t rec Indicates the preset recommended boiling time, T ref Indicates the reference temperature, T s Indicates the surface temperature of the hot pot, φ f Indicates the current firepower level, φ ref Indicates the reference firepower level.

[0038] The embodiments of the present invention have the following technical effects:

[0039] The present invention is based on the decoupling design of the first robotic arm being responsible for the scalding operation and the second robotic arm being responsible for gripping and plating, which solves the limitations of traditional single-arm sequential operations, realizes parallel processing of dish processing and cooking links, and significantly improves the efficiency of the cooking process; the visual perception unit is used to locate the position of the hot pot and the spatial posture of the dish, and the thermal monitoring unit is combined with the real-time detection of the temperature change of the hot pot surface and the liquid level monitoring unit to dynamically detect the liquid level of the hot pot soup, so as to realize the status monitoring of the hot pot cooking, solve the blind spots of traditional single sensor monitoring, and realize precise control of firepower and liquid level; the central controller automatically determines the scalding time based on the dish recognition result, the spatial posture of the dish, the surface temperature of the hot pot and the firepower level, and synchronously generates the robotic arm motion path and firepower adjustment instructions to realize adaptive control of the cooking process; the human-computer interaction module receives user preference instructions in real time and feeds back the instruction execution status, so as to dynamically match the mechanical operation with user needs and enhance the user's personalized experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 1 is a schematic structural diagram of a hot pot robot system provided by an embodiment of the present invention;

[0042] Figure 2 1 is a flow chart of a control method for a hot pot robot provided by an embodiment of the present invention;

[0043] Figure 3 This is a schematic structural diagram of a hot pot robot provided by an embodiment of the present invention;

[0044] Figure 4 It is a structural schematic diagram of another hot pot robot provided by an embodiment of the present invention.

[0045] Figure Number:

[0046] 100-first robotic arm;

[0047] 200-second robotic arm;

[0048] 300-Visual Perception Unit;

[0049] 400-Liquid level monitoring unit;

[0050] 500-fire control unit;

[0051] 600-hot pot;

[0052] 700-main control panel;

[0053] 800-Raw material area. DETAILED DESCRIPTION

[0054] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention are described clearly and completely below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are also within the scope of protection of the present invention.

[0055] The present invention proposes a hot pot robot system, such as Figure 1 As shown, the hot pot robot system specifically includes: a robotic arm module, a multi-sensor module, a central controller, a firepower control unit and a human-computer interaction module.

[0056] The robotic arm module includes a first robotic arm 100 and a second robotic arm 200. The first robotic arm is configured to perform dish blanching operations and kitchenware manipulation, while the second robotic arm is configured to perform dish picking operations and dish arrangement operations.

[0057] The multi-sensor module includes a visual perception unit 300, a thermal monitoring unit, and a liquid level monitoring unit 400. The visual perception unit is used to collect images of dishes and the location information of the hot pot 600, the thermal monitoring unit is used to monitor the surface temperature of the hot pot in real time, and the liquid level monitoring unit is used to detect the liquid level of the hot pot in real time.

[0058] The visual perception unit includes a multi-camera RGB-D camera array and an arm-mounted camera; the thermal monitoring unit includes a thermal infrared probe; and the liquid level monitoring unit includes a liquid level sensor. The multi-sensor module combines the visual perception unit, thermal monitoring unit, and liquid level monitoring unit to enhance the system's ability to respond to dynamic scenarios. The multi-sensor fusion algorithm ensures that the hot pot edge and liquid level can be stably identified even when lighting changes or oil stains on the soup surface interfere.

[0059] The central controller is used to identify the dish category and spatial position based on the dish image using a template matching algorithm; generate soup adjustment instructions and heat level adjustment instructions based on the hot pot soup level and the hot pot surface temperature; determine the scalding time based on the hot pot surface temperature and the heat level; and generate the coordinated motion path of the robotic arm module;

[0060] The heat control unit 500 is connected to the induction cooker or gas stove and is used to adjust the heat level of the induction cooker or gas stove according to the heat level adjustment instruction generated by the central controller. The heat level can include low, medium, high, etc.

[0061] The human-computer interaction module is used to receive user instructions and provide real-time feedback on system status.

[0062] This embodiment utilizes a decoupled design, with the first robotic arm responsible for scalding and the second for gripping and plating. This design also utilizes spatial trajectory decoupling, with the robotic arm's motion time slots complementing each other (for example, while the first robotic arm is scalding, the second robotic arm moves to a standby point). This overcomes the limitations of traditional single-arm sequential operations, enabling parallel processing of dish processing and cooking, shortening the task cycle and significantly improving cooking process efficiency. Complementary verification is achieved through the visual perception unit, combined with the thermal and liquid level monitoring units, overcoming the limitations of a fixed viewing angle. This allows for accurate extraction of hot pot edge and liquid level data even under varying lighting conditions and steam interference, enabling monitoring of the hot pot cooking status. This addresses the blind spots of traditional single-sensor monitoring and enables precise control of heat and liquid level. A central controller automatically determines the scalding duration based on dish recognition results, dish spatial position, hot pot surface temperature, and heat level, and simultaneously generates robotic arm motion paths and heat adjustment commands, enabling adaptive control of the cooking process. A human-computer interaction module receives user preference commands in real time and provides feedback on command execution status, dynamically matching mechanical operations with user needs and enhancing a personalized user experience.

[0063] In the embodiment of the present invention, the arrangement of the first robotic arm 100 and the second robotic arm 200 can be as follows: Figure 3 The method shown in FIG. 1 is independent of each other and can be set at different positions. Figure 4 The arrangement shown is for the left and right arms of the robot, respectively, and is not limited here. In actual operation, in addition to automatically controlling the corresponding parameters of the hot pot through the cooperation of the above-mentioned units, manual intervention control can also be performed through the main control panel 700, for example, the fire power level can be manually adjusted.

[0064] In some embodiments, the process of the central controller identifying the dish category and the spatial position of the dish using the template matching algorithm specifically includes:

[0065] Control the RGB-D camera on the arm of the visual perception unit to detect the food picking area (such as Figure 3 The raw material area 800 in the image processing unit is used to capture multi-angle template features of common dishes (such as beef, tripe, vegetables, fish balls, etc.), generate a pre-stored template feature library of dishes under multiple angles and multiple lighting conditions, and automatically adjust the threshold online with real-time resolution to ensure recognition accuracy in different containers and lighting environments;

[0066] After the food is served, the visual perception unit is controlled to capture the food collection area and the food is identified and its pose estimated using a fast template matching algorithm. The fast template matching algorithm can generate template features based on SURF+ descriptor matching or a lightweight convolutional network.

[0067] For example, in this embodiment, the cosine similarity between the current frame feature descriptor and the pre-stored template feature descriptor is extracted for dish recognition. The specific calculation formula is as follows:

[0068]

[0069] Among them, S represents the similarity value, i represents the i-th pair of matching features, N represents the number of feature matching pairs, f i c represents the feature descriptor of the i-th current frame, f i t Represents the i-th pre-stored template feature descriptor;

[0070] When the similarity exceeds the preset threshold, the dish category and spatial position of the dish are output;

[0071] Assuming the preset threshold is 0.8, when S≥0.8, the recognition is considered successful, and the dish category and dish spatial pose are output. The dish spatial pose includes the center coordinates of the dish (x p ,y p ,z p ) and the yaw angle θ p .

[0072] Furthermore, the initial fire level of the fire control unit is determined by the system based on the user's settings and the type of soup base (clear soup / spicy);

[0073] During the hot pot boiling process, the multi-sensor module measures the hot pot surface temperature T in real time through the thermal infrared probe. s , if T s >T high (For example, assuming that the upper temperature threshold of the hot pot is 100°C), the firepower is judged to be too high, and the central controller sends a "downshift" command to the firepower control unit; if T s <T low (For example, assuming that the lower limit threshold of the hot pot temperature is 80°C), it is determined that the firepower is too low, and the central controller sends a "downshift" command to the firepower control unit.

[0074] In some embodiments, the initial temperature of the hot pot and the user's setting preferences can be combined to design a PID type heat regulator so that T s Stable within ±2°C of target temperature fluctuation.

[0075] In some embodiments, the process of the central controller generating the soup adjustment instruction specifically includes:

[0076] Control the visual perception unit's downward-looking RGB-D camera array to perform real-time depth scanning of the hot pot to obtain a scene depth map;

[0077] In some embodiments, the visual liquid level and infrared temperature can be fused through Kalman filtering to buffer the noise of a single sensor, thereby further improving the liquid level estimation accuracy and temperature judgment accuracy, so that the liquid level estimation accuracy reaches ±3mm and the temperature judgment accuracy reaches ±0.5℃.

[0078] According to the scene depth map, the segmentation algorithm is used to separate the hot pot area from the desktop background and calculate the hot pot center coordinates (x h ,y h ,z h ) and hot pot radius R h ;

[0079] The hot pot center coordinates and hot pot radius determine the inner area of ​​the hot pot edge, the hot pot edge depth, and the liquid level depth;

[0080] Estimate the current liquid level based on the inner area of ​​the hot pot edge, the depth of the hot pot edge, and the liquid surface depth;

[0081] The specific calculation formula is as follows:

[0082]

[0083] Among them, h l Indicates the real-time depth of the soup in the hot pot, Ω h Indicates the inner area of ​​the hot pot edge, that is, the annular area of ​​the hot pot inner wall close to the edge, D 锅沿 (x,y) represents the depth of the hot pot edge, that is, the depth value of the pot edge in three-dimensional space. D(x,y) represents the liquid surface depth, that is, the depth value of the soup surface in three-dimensional space. (x,y) represents the plane coordinates of the pixel point in the depth map.

[0084] Determine the current liquid level and compare it with the soup threshold. If it is less than the soup threshold, generate a water-adding instruction.

[0085] For example, assuming that the soup threshold h min is 5cm, then when h l When the water level is less than 5cm, the system will prompt “The soup is too low, please add water” and send audio and video prompts to add water.

[0086] In some embodiments, the system further includes a water adding robot arm, which is used to automatically perform a water adding action after receiving a water adding instruction. For example, the water adding robot arm may move to a preset water cup position coordinate (x wt ,y wt ,z wt ) to perform the water extraction operation, and move the water joint (water injection port) to the edge of the pot to perform the water addition operation. Preferably, in order to avoid water splashing and forming a vertical water column, the z coordinate of the pot edge where the water joint is positioned is slightly higher than the actual coordinate of the pot edge. For example, it can be (x h ,yh ,z h +2cm), and let the water flow vertically into the pot.

[0087] In some embodiments, the coordinated motion path generated by the central controller includes the scalding motion path of the first robotic arm and the clamping motion path of the second robotic arm, dividing each task into a micro-action sequence of "scalding → avoiding → scooping → placing on the plate" to ensure that the paths of the two arms are continuous and do not interfere with each other.

[0088] In some embodiments, the trajectory planning of the first arm mainly includes the action sequence of "putting food into the pot → scalding → shaking → shaking off", and the path planning includes:

[0089] From the dish placement position (x p ,y p ,z p ) to the edge of the hot pot entrance (x e ,y e ,z e ) of the linear interpolation path of the linear interpolation;

[0090] From the entrance of the hot pot edge to the specified depth (x h ,y h ,z h +δ), where δ is used to control the scalding depth;

[0091] In the hot pot area, the arc trajectory (radius R r , angle α r ) of the periodic shaking path; wherein, αr can adopt a sinusoidal or elliptical path to ensure uniform heating of the dishes;

[0092] The lifting path after the scalding is completed is to rise to the edge of the pot to wait for subsequent scooping; for example, in order to avoid hitting the edge of the pot, the lifting path after the scalding is completed can be set to a vertical lifting path.

[0093] In some embodiments, the second arm trajectory planning is used to complete the "pre-retrieval → foundation → retrieval → placement" tasks in parallel with the first arm. The path planning includes:

[0094] During the hot-scalding process, the first robotic arm is pre-positioned to the preset standby coordinate (x w ,y w ,z w ), wait for the signal that the hot water is finished;

[0095] Generate a Cartesian straight line scooping path based on the spatial pose estimation of the food in the pot;

[0096] After taking out the food, lift it to the plate position at a fixed height to avoid the edge of the pot (x d ,y d ,z d); wherein, the fixed lifting height can be set according to actual conditions, for example, it can be set to 10cm.

[0097] In some embodiments, the collaborative motion path also includes implementing collision avoidance control, including: using a local obstacle avoidance algorithm to detect in real time whether a conflict occurs in the robot arm path, and achieving dynamic collision avoidance by adjusting the preset standby coordinates of the second robot arm and / or the scalding trajectory of the first robot arm.

[0098] Exemplarily, the local obstacle avoidance algorithm can be implemented using a Rapidly-exploring Random Tree-Connect (RRT-Connect) algorithm.

[0099] Furthermore, all path planning needs to ensure continuous velocity and acceleration constraints to prevent soup splashing or sudden stops at the end of the robotic arm.

[0100] Compared with traditional single-arm or manual timing solutions, in the embodiment of the present invention, the first and second robotic arms operate in parallel, which can realize the synchronous operation of "scalding and clamping", thereby improving the overall scalding efficiency; the first robotic arm continuously turns the ingredients in the pot with a parameterized path (arc or sinusoidal shaking) to ensure uniform heating; the second robotic arm is ready in advance, without waiting, reducing task switching time.

[0101] In some embodiments, when there are multiple hot pots running at the same time, the central controller can also dispatch the firepower control unit to separately regulate the firepower of each pot and the division of labor of the robotic arms to achieve parallel service in the same queue.

[0102] In some embodiments, the process of dynamically calculating the scalding time includes:

[0103] Call preset recommended blanching time based on dish type;

[0104] For example, the preset recommended scalding time t rec ={Beef: 30s; Tripe: 45s; Vegetables: 20s; Fish Balls: 25s};

[0105] The preset recommended scalding time is proportionally corrected according to the real-time hot pot surface temperature and the fire level to obtain the corrected scalding time;

[0106] The specific calculation formula is as follows:

[0107]

[0108] Among them, t adj Indicates the corrected scalding time, t rec Indicates the preset recommended boiling time, T refIndicates the reference temperature. For example, the reference temperature can be set to 100°C. s Indicates the surface temperature of the hot pot, φ f Indicates the current firepower level, φ ref Indicates the reference firepower level.

[0109] The corrected blanching time is then recalibrated based on the user's set doneness preference to obtain the final blanching time.

[0110] For example, if the user prefers “eight mature”, then t target =t adj ×0.8;

[0111] If the user prefers "well done", then t target =t adj ×1.2;

[0112] Among them, t target Represents the final blanching time after secondary correction. Adaptive calculation and personalized adjustment of blanching time ensures more even cooking of each dish, reduces user complaints of overcooked or undercooked food, and improves the user experience.

[0113] Furthermore, the timer accuracy is controlled within ±1s, and after the blanching is completed, the system automatically prompts "Please prepare to scoop out the food."

[0114] In some embodiments, the central unit uses sensors to fuse visual, thermal, liquid level, and force information to update the hot pot status in real time; automatically adjusts system strategies based on real-time user operational preferences (such as commonly used boiling time and taste preferences); and develops a "user habit database" to support learning of each user's doneness and taste preferences, giving the system a "memory effect" and improving satisfaction with repeated use.

[0115] Furthermore, the system can also integrate a customer preference learning module to optimize and adjust the preset recommended boiling time by recording the boiling time and evaluation of multiple dishes. rec and the corrected blanching time t adj accuracy.

[0116] Based on the above hot pot robot system, the embodiment of the present invention also provides a control method for the hot pot robot, see Figure 2 , specifically including the following steps:

[0117] S102, collecting images of dishes through a visual perception module;

[0118] S104: performing similarity matching between the dish image and pre-stored template features, identifying the dish category and outputting the spatial pose of the dish;

[0119] S106, integrating the thermal monitoring data, the fire level and the liquid level monitoring data, dynamically calculating the scalding time and generating a fire adjustment instruction and a soup adjustment instruction, and adjusting the hot pot fire level according to the fire adjustment instruction;

[0120] S108. Generate a scalding motion path for the first robotic arm and a gripping motion path for the second robotic arm, and control the first robotic arm to perform a scalding operation and control the second robotic arm to perform a scooping operation based on the scalding time.

[0121] Before S102, it also includes system initialization:

[0122] Upon system startup, the system performs a self-check process. The visual perception unit completes preheating and spatial calibration to ensure accurate and reliable measurement benchmarks. Multiple robotic arms simultaneously perform a zero-return calibration before automatically moving to a pre-set capture standby position, ensuring the food collection area and hot pot area are fully visible and unobstructed by mechanical structures. Users select the dish category and doneness preference through the human-computer interface. Based on the selection, the system loads pre-stored dish template feature data and the corresponding recommended scalding time parameters, completing the initial configuration of the cooking task.

[0123] S104 specifically includes:

[0124] The visual perception unit continuously scans the food collection area. When a target object is detected, a template matching algorithm is activated to extract the feature descriptors of the current image frame and compare them with a pre-stored template library. If the feature similarity reaches the recognition threshold, the system outputs the dish category name and spatial pose coordinates. The second robotic arm plans a Cartesian straight line path based on the pose data to pick up the dish and smoothly move it to the positioning point of the alternative dish plate. This synchronously triggers the first robotic arm to move from its standby position along an obstacle avoidance path to directly above the alternative dish plate for scalding. The multi-arm action timing is coordinated and coordinated by the central controller for seamless connection.

[0125] S106 specifically includes:

[0126] The thermal monitoring unit provides real-time feedback on the hot pot's surface temperature distribution, while the liquid level monitoring unit tracks changes in the soup's height. The central controller integrates sensor data and calculates the dynamic boiling time. Initial recommendations are first applied based on the dish type, followed by a proportional correction based on the real-time temperature and power level. Finally, a secondary optimization is performed based on user preferences for doneness. The boiling timer is displayed synchronously on the human-computer interface. If the liquid level falls below a safe threshold or the temperature exceeds a reasonable range, the system generates an audible and visual alarm and waits for user authorization. Based on the instructions, the system automatically dispatches the water-adding robotic arm to refill the water or adjust the power control unit level.

[0127] S108 specifically includes:

[0128] The first robotic arm carries the dish and immerses it into the hot pot broth, while periodically oscillating to simulate manual scalding. After the scalding countdown ends, the first robotic arm lifts and tilts to remove excess broth from the dish. The second robotic arm then begins high-speed movement from its standby position to the hot pot retrieval point. It then grasps the dish along an optimal obstacle-avoiding path, lifts it to a safe height, and moves laterally to the user-specified plate position for precise plating. Upon completion, a voice prompt announces the completion status. The robotic arm then returns to its camera standby position, and the system returns to standby mode, awaiting the next task.

[0129] The present invention matches a multi-angle dish feature template library with real-time RGB-D images, rapidly compares the current frame with the template features using a cosine similarity algorithm, and calculates spatial pose in combination with depth information. This resolves blind spots in fixed-viewpoint recognition and adapts to complex scenarios such as tableware reflection and stacking occlusion, ensuring gripping and positioning accuracy. A depth camera scans the geometric contours of the pot edge and combines it with a thermal infrared camera to monitor the liquid surface temperature in real time. The liquid level is dynamically estimated using depth difference. A Kalman filter is used to fuse temperature time series data to suppress sensor noise. A coupled model of soup consumption and firepower intensity is constructed to implement abnormal liquid level warnings and adaptive power level adjustment. Real-time collision detection and local avoidance ensure that while the first robotic arm is performing boiling and shaking operations, the second robotic arm remains on standby and quickly switches to scooping, preventing soup splashing and mechanical collisions. While the first robotic arm is continuously flipping the food, the second robotic arm synchronously completes the scooping operation, shortening the task idle period. Based on historical user operation data and doneness preferences, the benchmark boiling time for dishes is adjusted. Environmental parameters and user preferences are integrated to achieve personalized taste control, thereby improving boiling efficiency, safety, and a personalized experience.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A hot pot robot system, characterized in that: include: The robotic arm module includes a first robotic arm and a second robotic arm, wherein the first robotic arm is configured to perform dish blanching operations and kitchen utensil manipulation actions, and the second robotic arm is configured to perform dish picking operations and dish arrangement actions; The multi-sensor module includes a visual perception unit, a thermal monitoring unit, and a liquid level monitoring unit. The visual perception unit is used to collect images of dishes and hot pot location information, the thermal monitoring unit is used to monitor the hot pot surface temperature in real time, and the liquid level monitoring unit is used to detect the hot pot soup level in real time. The central controller is configured to identify the dish category and spatial position of the dish based on the dish image; generate a soup adjustment instruction and a heat level adjustment instruction based on the hot pot soup level and the hot pot surface temperature; and determine a scalding time based on the hot pot surface temperature and the heat level; generating a coordinated motion path of the robotic arm module; A firepower control unit, used to adjust the firepower level of the hot pot according to the firepower level adjustment instruction; The human-computer interaction module is used to receive user instructions and provide real-time feedback on the system execution status.

2. The hot pot robot system according to claim 1, characterized in that: The identification of dish categories and dish spatial postures includes: Control the visual perception unit to shoot the food picking area, extract the current frame feature descriptor based on the shot image, identify the food based on the cosine similarity between the current frame feature descriptor and the pre-stored template feature descriptor, and output the food category and food spatial pose when the similarity exceeds the preset threshold.

3. The hot pot robot system according to claim 1, characterized in that: The process of generating the soup adjustment instruction includes: Controlling the visual perception unit to perform real-time depth scanning of the hot pot to obtain a scene depth map; According to the scene depth map, a segmentation algorithm is used to separate the hot pot area from the desktop background, and the coordinates of the hot pot center and the hot pot radius are calculated; Determine the inner area of ​​the hot pot edge, the depth of the hot pot edge, and the liquid level depth based on the hot pot center coordinates and the hot pot radius; Estimate the current liquid level based on the inner area of ​​the hot pot edge, the depth of the hot pot edge, and the liquid surface depth; The current liquid level is compared with the soup threshold, and if it is less than the soup threshold, a water addition instruction is generated.

4. The hot pot robot system according to claim 3, characterized in that: The system also includes a water adding robot arm, which is used to automatically perform a water adding action after receiving a water adding instruction.

5. The hot pot robot system according to claim 1, characterized in that: The collaborative motion path includes a scalding motion path of the first robotic arm and a clamping motion path of the second robotic arm.

6. The hot pot robot system according to claim 5, characterized in that: The scalding motion path of the first robotic arm includes: A straight line interpolation path from the dish placement position to the edge of the hot pot; The periodic shaking path of the arc trajectory in the hot pot area; The lifting path after scalding.

7. The hot pot robot system according to claim 5, characterized in that: The gripping motion path of the second robotic arm includes: During the scalding process, the first robotic arm is pre-positioned to the preset standby coordinates; Generate a Cartesian straight line scooping path based on the estimated posture of the food in the pot; After taking out the food, lift it to a fixed height to avoid the edge of the pot and move it to the plate position.

8. The hot pot robot system according to claim 7, characterized in that: The collaborative motion path also includes implementing collision avoidance control, including: using a local obstacle avoidance algorithm to detect in real time whether a conflict occurs in the robot arm path, and achieving dynamic collision avoidance by adjusting the preset standby coordinates of the second robot arm and / or the scalding trajectory of the first robot arm.

9. The hot pot robot system according to claim 1, characterized in that: Determining the scalding time according to the surface temperature of the hot pot and the fire level includes: Call preset recommended blanching time based on dish type; Proportionally correcting the preset recommended hot pot time according to the real-time hot pot surface temperature and the fire level to obtain a corrected hot pot time; The corrected blanching time is corrected again based on the doneness preference set by the user to obtain the final blanching time.

10. The hot pot robot system according to claim 9, characterized in that: The preset recommended scalding time is proportionally corrected according to the real-time hot pot surface temperature and the fire level, and the corrected scalding time is: Among them, t adj Indicates the corrected scalding time, t rec Indicates the preset recommended boiling time, T ref Indicates the reference temperature, T s Indicates the surface temperature of the hot pot, φ f Indicates the current firepower level, φ ref Indicates the reference firepower level.

Images