Instant catering distribution method integrating air-ground cooperation and intelligent perception

Abstract

The invention provides an instant catering distribution method integrating air-ground cooperation and intelligent perception, and relates to the technical field of intelligent distribution, and the method comprises the steps: analyzing a catering distribution request to determine a target catering load and a physical stability constraint parameter; the double carriers are controlled to establish a mobile handover window with synchronous relative positions, and load physical transfer is completed under the state of keeping cooperative movement; microenvironment data of the load are collected in real time based on the transferred transport capacity state, and the deviation degree is calculated according to the microenvironment data and the physical stability constraint parameters; and generating a self-adaptive steady-state control instruction based on the deviation degree, adjusting the motion attitude parameters of the current bearing main body in real time, and maintaining the load state within a limited range until delivery. According to the invention, through dynamic mobile handover of air-ground transport capacity and adaptive closed-loop control based on internal object state feedback, the problems of poor transport capacity collaboration and disjunction of sensing control in the prior art are solved, and the catering distribution efficiency and physical quality in a complex environment are effectively guaranteed.

Description

Technical Field

[0001] This invention relates to the field of intelligent delivery technology, and in particular to an instant food delivery method that integrates air-ground collaboration and intelligent sensing. Background Technology

[0002] Current on-demand food delivery primarily relies on a single human operator or a single vehicle. The mainstream method involves point-to-point delivery by ground delivery personnel using electric bicycles, while some advanced solutions are beginning to explore independent delivery using drones or unmanned vehicles. Existing dispatch systems typically use static maps and real-time traffic flow data to plan routes for individual delivery capacity, aiming to achieve delivery from merchants to users in the shortest time or along the shortest route.

[0003] With the development of the low-altitude economy and autonomous driving technology, the delivery industry is transforming towards unmanned and three-dimensional operations. The future trend is to break free from the limitations of single-capacity transport and seek a combination of ground transport with high load capacity and long range, and air transport with high speed and freedom from ground traffic restrictions. Simultaneously, to ensure the quality of food, especially noodle soups and fragile foods, the perception capabilities during delivery are evolving from simple location perception to a two-way coupled perception of the environment and goods, dynamically adjusting transportation strategies based on the real-time status of the goods, such as temperature or movement.

[0004] Existing technologies have significant shortcomings when facing complex urban environments. First, they suffer from poor capacity coordination. Current air-to-ground delivery largely relies on simple relay models, such as fixed distribution points, lacking dynamic coordination based on real-time road conditions and order characteristics. This prevents efficient air-to-ground interaction during movement, such as mobile take-off and landing. Second, there is a disconnect between environmental perception and control. Current delivery methods focus solely on delivery while neglecting damage prevention. For example, when drones encounter strong air currents or unmanned vehicles traverse bumpy roads, the lack of real-time perception and feedback control mechanisms for the condition of food items can lead to physical damage upon delivery, such as spillage due to resonance. Finally, route planning is rigid. Traditional route planning fails to consider the specific physical constraints of food items, such as their heat sensitivity or susceptibility to spillage, making it difficult to achieve an optimal balance between delivery efficiency and service quality. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, the purpose of this invention is to provide an instant food delivery method that integrates air-ground coordination and intelligent sensing. This invention solves the problems in the prior art, such as poor capacity coordination leading to the inability to achieve efficient air-ground interaction, disconnect between environmental perception and control leading to damage to the physical properties of food, and rigid route planning leading to difficulty in balancing delivery efficiency and service quality.

[0006] To achieve the above objectives, the present invention provides the following solution: An on-demand food delivery method integrating air-ground collaboration and intelligent sensing includes: Obtain food delivery requests and determine the target food load and corresponding physical stability constraint parameters based on the food delivery requests; Collect real-time environmental data of the delivery area and calculate the air-ground collaborative delivery strategy based on the physical stability constraint parameters; According to the air-ground collaborative delivery strategy, ground vehicles and air vehicles are assigned from the transportation capacity resource pool as the execution entities, and the dynamic handover interval between the ground vehicles and the air vehicles in the spatiotemporal domain is planned. Control the ground vehicle and the air vehicle to move towards the dynamic handover zone, and after entering the dynamic handover zone, control the air vehicle and the ground vehicle to establish a moving handover window with relative position synchronization; Within the mobile handover window, the air vehicle and the ground vehicle are controlled to complete the physical transfer of the target catering load while maintaining coordinated movement, and the subsequent delivery route is updated according to the capacity status after the transfer is completed, so as to obtain the updated capacity status and subsequent delivery route. Based on the updated transportation capacity status and subsequent delivery routes, real-time micro-environmental status data of the target catering load is collected. The corresponding deviation is calculated based on the microenvironment state data and the physical stability constraint parameters; Generate adaptive steady-state control commands based on the deviation; The adaptive steady-state control command is used to adjust the motion attitude parameters of the ground vehicle and the air vehicle currently carrying the target catering load in real time, so as to maintain the state of the target catering load within the range defined by the physical stability constraint parameters until delivery is completed.

[0007] The present invention discloses the following technical effects: This invention provides an instant food delivery method integrating air-ground collaboration and intelligent sensing. First, addressing the problem of poor capacity coordination, this method achieves dynamic handover between ground and air vehicles while they are in motion, breaking through the efficiency bottleneck of traditional fixed-point relays and significantly improving resource utilization and three-dimensional delivery efficiency. Second, addressing the problem of disconnect between sensing and control, this method innovatively introduces real-time status feedback based on the physical constraints of the food, enabling adaptive correction of the vehicle's motion attitude, fundamentally ensuring the temperature and physical quality of the food, and effectively avoiding losses due to vibration or spillage. Finally, through risk matrix and multi-objective optimization, this method achieves an optimal balance between delivery efficiency and service quality in complex urban environments. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A flowchart illustrating an instant food delivery method integrating air-ground collaboration and intelligent sensing, provided as an embodiment of the present invention; Figure 2 The flowchart illustrates the closed-loop execution and verification of adaptive steady-state control provided in this embodiment of the invention. Detailed Implementation

[0010] 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.

[0011] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0012] like Figure 1 As shown, this invention provides an instant food delivery method that integrates air-ground collaboration and intelligent sensing, comprising: Step 100: Obtain the food delivery request, and determine the target food load and corresponding physical stability constraint parameters based on the food delivery request; Step 200: Collect real-time environmental data of the delivery area and calculate the air-ground collaborative delivery strategy based on the physical stability constraint parameters; Step 300: According to the air-ground collaborative delivery strategy, assign ground vehicles and air vehicles as the execution entities from the transportation capacity resource pool, and plan the dynamic handover interval between the ground vehicles and the air vehicles in the spatiotemporal domain. Step 400: Control the ground vehicle and the air vehicle to move towards the dynamic handover zone, and after entering the dynamic handover zone, control the air vehicle and the ground vehicle to establish a moving handover window with relative position synchronization; Step 500: Within the mobile handover window, control the air vehicle and the ground vehicle to complete the physical transfer of the target catering load while maintaining coordinated movement, and update the subsequent delivery route according to the capacity status after the transfer is completed, so as to obtain the updated capacity status and subsequent delivery route. Step 600: Based on the updated transportation capacity status and subsequent delivery routes, collect real-time micro-environmental status data of the target catering load; Step 700: Calculate the corresponding deviation based on the microenvironment state data and the physical stability constraint parameters; Step 800: Generate adaptive steady-state control commands based on the deviation; Step 900: Use the adaptive steady-state control command to adjust the motion attitude parameters of the ground vehicle and the air vehicle currently carrying the target catering load in real time, so as to maintain the state of the target catering load within the range limited by the physical stability constraint parameters until delivery is completed.

[0013] Furthermore, the specific implementation process of step 100 is as follows: This embodiment receives a food delivery request from a user terminal and performs structured parsing to extract a list of food category identifiers and their corresponding portion sizes. Subsequently, this embodiment iterates through the list of category identifiers, retrieving a unique physical morphology tag from a pre-built food physical attribute database that matches each category identifier. Unlike the coarse category divisions in traditional solutions, these tags precisely define quantifiable physical characteristics such as the proportion of liquid components, thermodynamic sensitivity level, and rigid structural strength, providing a foundation for subsequent accurate modeling.

[0014] Based on the aforementioned component data and inherent physical morphology labels, the core of this embodiment lies in performing weighted aggregation calculations on the physical characteristics of all food and beverage categories to construct a composite physical characteristic model representing the overall state of the target food and beverage load. This modeling process considers the proportional relationships and mutual influences of different categories within the load cell, and in particular, addresses the technical shortcomings of traditional solutions that evaluate multiple food items individually while neglecting the overall coupling effect. This ensures that the model accurately represents the overall centroid distribution and comprehensive vulnerability of the load, a crucial prerequisite for preventing spillage during delivery.

[0015] Finally, in this embodiment, the composite physical feature model is input into a preset transportation tolerance mapping logic to calculate the critical threshold required to maintain the physical properties of the target catering load during dynamic transportation. This threshold is precisely quantified into physical stability constraint parameters that can be directly used by the vehicle control module. These constraint parameters specifically include the maximum allowable linear acceleration, the maximum allowable tilt angular velocity, and the allowable temperature fluctuation range. This step completes the transformation from abstract catering information to executable control parameters, providing a solid and dynamic constraint foundation for subsequent adaptive closed-loop control.

[0016] Specifically, this embodiment constructs the expression of the composite physical characteristic model to generate a unified representation of the overall physical state of the target catering load, ensuring an accurate assessment of the overall vulnerability of the load. The composite physical characteristic model is defined as the final representation after integrating the physical attributes of all packaged meals. This embodiment calculates the contribution of each catering category to the overall model through weighted aggregation operations. The inherent physical morphology label vector represents the inherent attributes of a particular meal, such as the proportion of liquid components, thermodynamic sensitivity level, and rigid structural strength, and its data comes from preset information in the catering physical attribute database. This operation combines the inherent physical morphology label vector with the geometric parameters of the target catering load box. The specific source of these geometric parameters is the internal volume of the load box and the placement position of the meals; their function is to reflect the spatial constraints and center of force of the load within the box. For example, for a label vector with a high proportion of liquid components, if its geometric parameters indicate that the meal is close to the edge of the load box, the weighted aggregation operation will increase its contribution to the overall sway risk. Subsequently, this embodiment sums up the contribution of all dishes and normalizes them according to the total number of dishes and their weights to obtain the final composite physical feature model. The output of this model will be directly used as the input of the transportation tolerance mapping logic to ensure the accuracy of subsequent calculation of physical stability constraint parameters.

[0017] Specifically, this embodiment will illustrate this with a dual-category order example. Assume the target catering load contains two types of catering items. This embodiment first assigns weights to each category: the first catering category has a weight of 0.6, and the second has a weight of 0.4. When calculating the contribution of the first category, its inherent physical morphology label vector is set to 0.8, representing high thermodynamic sensitivity, while the geometric parameter of the target catering load box is set to 0.9, indicating its proximity to the box edge. In the weighted aggregation operation, this embodiment couples the label vector with the geometric parameter, calculating the category's basic contribution to the overall model as 0.72. After weighted summation of all category contributions, the final output value of the composite physical characteristic model is calculated to be 0.75, which is the quantitative indicator reflecting the overall risk and vulnerability of this batch of catering loads.

[0018] Furthermore, the specific implementation process of step 200 is as follows: This embodiment acquires real-time environmental data, including meteorological data, real-time traffic flow data, and low-altitude airspace status information, through an urban sensing network, and extracts deep features from it. Unlike conventional data reading, this embodiment performs refined airflow gradient analysis on the real-time meteorological data to identify the specific impact of local micro-meteorological changes on flight stability, thereby quantifying the dynamic risk factors in the air. Simultaneously, this embodiment calculates the congestion index from the real-time traffic flow data, focusing on identifying the frequency of road bumps and sudden stops, quantifying the dynamic risk factors on the ground. Subsequently, this embodiment performs multi-dimensional data fusion of the low-altitude airspace status information and the aforementioned dynamic risk factors on the air and ground to construct a real-time dynamic risk matrix. This matrix dynamically maps the potential threat levels to load stability at different spatiotemporal nodes within the delivery area, providing a quantitative benchmark for subsequent safety planning.

[0019] This embodiment performs crucial coupling operations, interactively mapping the real-time dynamic risk matrix with the physical stability constraint parameters determined in the preceding steps to generate a set of spatiotemporal operational constraints for the vehicle. In this process, this embodiment does not statically set thresholds but dynamically adjusts the vehicle's operational boundaries. Specifically, this embodiment dynamically limits the flight altitude of the aerial vehicle based on airflow gradient data to avoid high-risk turbulent layers. Simultaneously, based on the sensitivity requirements for vibration and tilt in the physical stability constraint parameters, this embodiment reverse-engineers the maximum acceleration limits of the vehicle in specific risky road sections or airspaces. This generation mechanism ensures that the boundary conditions for path planning depend not only on geographical accessibility but also on the physical carrying capacity of the catering load itself, achieving a logical closed loop from environmental adaptation to cargo protection.

[0020] This embodiment, based on a preset set of optimization objectives and boundary conditions, employs a multi-agent spatiotemporal collaborative planning algorithm to iteratively solve for the ground driving path of ground vehicles and the air flight path of air vehicles. This embodiment sets dual optimization objectives: prioritizing the primary objective of minimizing total delivery time, it then maximizes the secondary objective of food and beverage load stability through weight adjustment, thereby finding the optimal solution between efficiency and safety. After obtaining the initial paths, this embodiment performs millisecond-level spatiotemporal synchronization calibration on the ground driving path and the air flight path, calculating their trajectory overlap in four-dimensional spatiotemporal space to accurately determine the temporal and spatial coupling intersection point. Based on this coupling intersection point, this embodiment clearly defines the dynamic handover interval between ground vehicles and air vehicles in the spatiotemporal domain, thus achieving seamless connection between air and ground transportation capacity during movement, and ultimately obtaining a structured integration of air-ground collaborative delivery strategies.

[0021] Furthermore, the specific implementation process of step 300 is as follows: This embodiment executes a high-precision capacity matching and assignment process, which goes beyond simple vehicle availability retrieval. It delves into the load requirements, mileage requirements, and time constraints within the air-ground collaborative delivery strategy. This embodiment evaluates the matching of the remaining energy spectrum and power performance curves of each ground and air vehicle in the capacity resource pool, eliminating capacity units that cannot meet specific task constraints, thereby generating a candidate set of vehicles for task execution. Subsequently, based on the principles of optimal distance and optimal performance, this embodiment assigns specific ground and air vehicles from the candidate set as the task execution entities, and loads the air-ground collaborative delivery strategy, containing precise route data, into the navigation control modules of both vehicles, completing the task-level initialization configuration.

[0022] This embodiment focuses on the parameterized configuration and logical mapping of dynamic handover intervals. It precisely extracts the start and end spatiotemporal coordinates of the dynamic handover interval from the air-to-ground collaborative delivery strategy. Unlike the ordinary waypoint settings in conventional navigation, this embodiment defines these two key coordinate points as high-priority trigger parameters for collaborative positioning and attitude synchronization. This means that when the vehicle reaches these spatiotemporal coordinates, its control logic will be forcibly switched. This trigger parameter serves as a core activation signal, used to subsequently trigger the vehicle's underlying state transition from independent navigation mode to collaborative interlocking mode, thereby ensuring the spatiotemporal accuracy of the handover action at the logical level.

[0023] This embodiment completes the final system-level deployment and verification based on the aforementioned trigger parameters. This embodiment uses these trigger parameters to pre-configure the vehicle cooperative positioning system for ground and air vehicles, setting the communication frequency band and positioning reference when entering the handover zone to prevent signal handshake delays. Simultaneously, this embodiment pre-verifies the motion control parameters to ensure that the vehicle's dynamic response characteristics meet the control precision required to maintain relative stillness within the dynamic handover zone. Through this series of configurations and verifications, this embodiment determines the final planning and deployment of the dynamic handover zone, making full preparations at both the hardware and software levels for seamless load transfer in the physical world.

[0024] Furthermore, the specific implementation process of step 400 is as follows: In this embodiment, ground and air vehicles approach a dynamic handover zone according to a preset trajectory. During this process, this embodiment does not rely solely on absolute coordinates from the Global Positioning System (GPS), but instead calculates the vehicle's proximity status relative to the starting coordinates of the dynamic handover zone in real time. This embodiment continuously executes high-frequency spatiotemporal comparison logic, performing difference calculations between the vehicle's real-time position and preset spatiotemporal trigger parameters. Once the spatiotemporal distance between the two vehicles converges to a preset threshold range, this embodiment immediately generates a high-precision cooperative mode switching command. This command signifies that the vehicle control logic has officially switched from an independent trajectory-following mode to a dual-vehicle cooperative interaction mode.

[0025] In response to the high-precision cooperative mode switching command, this embodiment initiates the underlying navigation reference conversion program. Instead of using the Earth-fixed coordinate system as a single reference, this embodiment activates the near-field perception link between vehicles, dynamically mapping the navigation reference to a relative motion coordinate system with the ground vehicle as the origin. In this coordinate system, this embodiment reconstructs the flight control law of the air vehicle, calculating the relative velocity and attitude compensation components that can offset the motion vector of the ground vehicle, thereby ensuring that the air vehicle can accurately follow and accompany the ground vehicle with the ground vehicle as a reference.

[0026] This embodiment performs microsecond-level attitude and velocity adjustments based on the reconstructed control law to achieve full-dimensional dynamic synchronization between the airborne and ground vehicles in terms of direction of motion, relative position, and relative velocity. This embodiment monitors the relative position deviation between the two vehicles in real time and filters out environmental noise interference using a Kalman filter algorithm. When the three-axis relative error is detected to have stably converged within the tolerance range allowed by the physical transfer mechanism, this embodiment determines that the synchronization state has been achieved. This formally establishes a mobile handover window for relative position synchronization in the high-speed spatiotemporal domain, providing a stable operating environment with zero relative velocity for subsequent physical connections.

[0027] Furthermore, the specific implementation process of step 500 is as follows: This embodiment maintains microsecond-level dynamic coordination between the airborne and ground vehicles within an established, synchronized relative position handover window. This aims to eliminate vibration interference caused by relative motion and provide a stable mechanical environment for the physical connection. This embodiment sends high-priority actuation commands to the automated transfer mechanism onboard the vehicle, driving the mechanical gripping assembly or electromagnetic adsorption device to perform the physical transfer of the target catering load. During this process, this embodiment precisely controls the timing of the load's release from the donating vehicle and its gripping into the receiving vehicle's cargo hold, ensuring a smooth transition during high-speed movement and simultaneously generating a load physical transfer execution signal, marking the formal initiation of the physical contact and handover process.

[0028] This embodiment initiates a rigorous dual-state verification logic based on the physical transfer execution signal of the load to prevent accidental slippage of the load during the transfer. This embodiment uses microswitches, pressure sensors, and mechanical interlocking feedback devices located at the carrier interface to monitor the detachment status of the target catering load at the donation end and the locking status at the receiving end in parallel and in real time. This embodiment performs strict logical AND operations; the physical transfer is only considered successful when the load locking status signal at the receiving end is triggered, indicating a stable connection, and simultaneously the detachment status signal at the donation end is confirmed, indicating complete release. This mechanism effectively avoids the risk of false connections caused by unilateral signal false alarms and generates an irreversible transfer success confirmation command accordingly.

[0029] In response to the successful transfer confirmation command, this embodiment immediately triggers a dynamic reset procedure for the global transport capacity status. This embodiment collects key resource parameters of ground and air vehicles in real time at the moment of transfer completion, focusing on updating remaining battery power, real-time load data that changes abruptly due to load transfer, and the current absolute geographical location. Based on these updated physical boundary conditions, this embodiment re-evaluates the preset remaining delivery task objectives in real time and uses path planning algorithms to accurately calibrate subsequent delivery paths. Specifically, this embodiment optimizes the return or redeployment path of air vehicles based on the reduced energy consumption characteristics after unloading, or corrects the trajectory of ground vehicles based on the dynamic constraints of increased load. Finally, it structurally integrates all updated status data and path data, completing a seamless transition from the air-ground collaborative stage to the independent last-mile delivery stage.

[0030] Furthermore, the specific implementation process of step 600 is as follows: This embodiment first accurately identifies the actual transport vehicle currently carrying the target catering load based on the updated transport capacity status from previous steps. Then, it dynamically configures the sampling strategy of the onboard sensors based on the terrain or weather characteristics of the subsequent delivery route. Instead of using a fixed low-frequency sampling mode, this embodiment sends activation commands to the microelectromechanical system (MEMS) sensor array mounted on the load container based on the predicted bump or turbulence risk level along the route, switching it to a high-frequency data acquisition mode. This embodiment utilizes multi-sensor fusion technology to synchronously acquire raw data streams reflecting the microscopic physical environment of the load in real time, and performs time-series alignment and noise reduction processing to construct a digital sensing environment that accurately reflects the stress state of the catering load.

[0031] This embodiment focuses on collecting and extracting key characteristic data characterizing dynamic stability. Specifically, this embodiment calculates triaxial acceleration characteristic values ​​in real time using accelerometers, with a focus on monitoring the root mean square acceleration in the vertical direction to quantify the linear vibration intensity caused by uneven road surfaces or sudden airflow changes. Simultaneously, this embodiment utilizes gyroscopes to collect real-time triaxial angular velocity characteristic values, capturing the angular sway rate of the vehicle during turning or attitude adjustments. Furthermore, this embodiment combines inertial measurement units to calculate the real-time attitude tilt data of the load cell, including the absolute values ​​of pitch and roll angles. This data directly reflects the static spillage risk of liquid catering under gravity, providing a core mechanical basis for subsequent steady-state control.

[0032] This embodiment focuses on both the dynamic state and the thermodynamic microenvironment of the load. It uses thermal probes placed at different spatial locations within the chamber to collect real-time temperature gradient data within the load chamber, rather than focusing on a single point temperature. This embodiment analyzes the temperature difference distribution and temperature change rate between the top and bottom of the chamber to assess the dynamic impact of external wind speed or vehicle movement speed on the food insulation performance. Finally, this embodiment structurally encapsulates the aforementioned dynamic characteristic data and thermodynamic gradient data to generate standardized target food load microenvironment state data. This data package serves as the direct input for subsequent deviation calculations, ensuring the integrity of the data chain from physical perception to digital decision-making.

[0033] Furthermore, the specific implementation process of step 700 is as follows: This embodiment executes the core risk quantification calculation process, aiming to transform multi-dimensional physical sensing data into a single measurable control benchmark. First, it establishes an index mapping relationship between microenvironmental state data and physical stability constraint parameters, aligning real-time collected acceleration, angular velocity, tilt angle, and temperature gradient to their respective safety thresholds. Subsequently, instead of directly comparing numerical values, this embodiment calculates the relative percentage of real-time values ​​exceeding the safety threshold for each dimension. This embodiment introduces a non-linear, one-way penalty mechanism: risk is only considered when the real-time monitored value exceeds the critical threshold; if the monitored value is within the safety range, the risk contribution for that dimension is zero. This approach effectively avoids data noise in a safe state interfering with the decision-making system.

[0034] The deviation in this model is a dimensionless composite value used to characterize the overall degree to which the current transportation state deviates from a safe and stable state; a larger value indicates a higher risk of cargo damage. The dimensional index in the model refers to the four main physical dimensions included in the microenvironmental state data: real-time triaxial acceleration characteristic values, real-time triaxial angular velocity characteristic values, real-time attitude tilt data of the load cell, and temperature gradient data inside the load cell. The risk weight coefficient in the model is a numerical factor reflecting the importance of a specific dimension to overall stability. The value of this coefficient is based on the inherent attributes of the food category. For example, for noodle soups, which are highly sensitive to spills, the weight coefficient for tilt data is set to a higher value, while for frozen foods, the weight coefficient for temperature gradient is correspondingly increased.

[0035] In this embodiment, the real-time feature values ​​used in the calculation are directly derived from the micro-environmental state data collected by sensors and processed by noise reduction, representing the actual force situation in the physical world. The critical threshold, on the other hand, is directly derived from the physical stability constraint parameters calculated based on the composite physical feature model in the preceding steps, representing the physical limit that the catering load can withstand. The core of this calculation logic lies in calculating the difference between the absolute value of the real-time feature value and the absolute value of the critical threshold, then dividing this difference by the critical threshold to obtain the relative over-limit ratio, and finally multiplying the over-limit ratios of each dimension by the corresponding risk weight coefficient and summing them up. Through this weighted over-limit normalization calculation, this embodiment can accurately identify which physical factor is threatening catering quality and provide a quantitative basis for subsequently generating targeted adaptive control commands.

[0036] Specifically, this embodiment uses a high-risk scenario of drone delivery of a "beef noodle soup" as an example to explain the parameters in the above calculation model. In this scenario, among the four physical dimensions represented by the dimension index, this embodiment focuses on selecting "real-time attitude tilt data of the load cell" as the third dimension and "real-time triaxial acceleration characteristic value" as the first dimension for calculation. From the physical stability constraint parameters calculated in the previous steps, the critical threshold of the first dimension (i.e., the maximum allowable linear acceleration) is set to 2.0 m / s², and the critical threshold of the third dimension (i.e., the maximum allowable tilt angle) is set to 15 degrees. Given the extremely high spill sensitivity of noodle soup, the risk weight coefficient of the third dimension, derived from the pre-set attribute database, is assigned a high value of 0.5 to reflect its decisive influence on overall stability, while the risk weight coefficient of the first dimension is set to 0.3.

[0037] When encountering lateral gusts during delivery, micro-environmental status data from onboard sensors showed: the real-time characteristic value of the first dimension was 1.5 meters per second squared, which did not exceed the threshold; however, the real-time characteristic value of the third dimension instantly reached 18 degrees, exceeding the safety limit. Based on these values, the system performed coupled calculations: since the first dimension did not exceed the limit, its risk contribution value was 0; the relative excess ratio of the third dimension was calculated as 0.2 (i.e., the excess of 3 degrees divided by the baseline of 15 degrees), which was multiplied by a weighting coefficient of 0.5 to obtain a risk component of 0.1 for this dimension. Assuming that the other dimensions were in a stable state, the system finally accumulated a deviation value of 0.1. This specific value directly quantifies the current spill risk level and serves as the input signal to trigger subsequent attitude correction control.

[0038] Furthermore, the specific implementation process of step 800 is as follows: This embodiment implements a refined risk classification and strategy matching process, aiming to transform abstract deviation values ​​into control logic with physical meaning. Instead of simply setting a single alarm threshold, this embodiment pre-sets multi-level graded control thresholds, including those for minor disturbances, significant oscillations, and critical instability. This embodiment compares the deviation calculated in previous steps with these graded thresholds using multi-level hysteresis to accurately determine the current load risk level. Based on this risk level and the characteristics of the current carrier type, this embodiment addresses within a pre-set control strategy library and dynamically selects the most suitable steady-state control strategy. For example, when an airborne vehicle is identified as being at a high risk level, this embodiment prioritizes an attitude decoupling control strategy, allowing the aircraft fuselage to tilt to counteract crosswinds while forcibly maintaining the horizontal stability of the mounted gimbal; for ground vehicles, this embodiment selects an active suspension damping adjustment strategy to absorb road impacts.

[0039] This embodiment, based on the selected steady-state control strategy, enters the core parameter calculation stage and performs deviation-based inverse compensation calculation. This embodiment combines the specific value of the deviation with the boundary difference of the physical stability constraint parameters, using a PID controller or model predictive control algorithm to calculate the target motion parameter adjustment amount aimed at offsetting external environmental disturbances. This embodiment specifically quantifies this adjustment amount into a set of control parameters for disturbance compensation. This parameter set is not a general navigation command, but rather fine-tuning parameters for stability. For example, when calculating acceleration compensation, this embodiment dynamically adjusts the control gain according to the magnitude of the deviation, outputting a reverse acceleration command accurate to the millisecond level or a damping coefficient for a specific frequency, ensuring that the control intervention effectively suppresses disturbances without causing new oscillations due to over-correction.

[0040] This embodiment completes the final implementation of the control logic, standardizing the encapsulation and encoding of the execution instructions. This embodiment reads the underlying hardware interface protocol of the current vehicle and, through a built-in protocol adaptation layer, encapsulates the abstract steady-state control strategy and quantized control parameter set. This embodiment converts the control signals into binary data packets conforming to the vehicle controller area network bus standard or the UAV flight control link standard, generating the final adaptive steady-state control instruction. This instruction has the highest execution priority, can directly penetrate conventional navigation logic, and drive the vehicle's actuators to make an immediate response, thereby achieving lossless transmission from algorithmic decision-making to physical actions.

[0041] Furthermore, the specific implementation process of step 900 is as follows: like Figure 2As shown, in this embodiment, the control process begins by sending an adaptive steady-state control command to the vehicle and then monitoring whether the motion control interface has successfully received the command. If it is determined that the command has not been received (no), the system executes a retransmission mechanism to immediately resend the control command to ensure signal delivery; if it is confirmed that the command has been received (yes), the actuator adjusts the flight attitude and suspension damping in real time according to the command. This adjustment directly produces the corrected vehicle motion attitude parameters. Subsequently, the system enters the critical closed-loop verification stage to determine whether the microenvironmental state of the target catering load has stabilized and met the standards. If the state does not meet the standards (no), the system triggers an adaptive correction mechanism and re-enters the loop of parameter calculation and command generation; only when the state is confirmed to be stable and meets the standards (yes) does the control process officially end, completing this steady-state adjustment.

[0042] This embodiment constructs a command transmission link based on a low-latency communication protocol to accurately distribute the generated adaptive steady-state control commands to the actual carrier currently carrying the target catering load. This embodiment uses an authentication mechanism to lock the underlying motion control interface of the target vehicle, ensuring that the control signal can cover the original cruise commands with a millisecond-level response speed, regardless of whether the vehicle is a drone flying at high altitude or an unmanned vehicle traveling on the ground. This establishes a real-time intervention channel from cloud decision-making to edge execution, ensuring that correction signals can be transmitted to the execution end as soon as external disturbances occur.

[0043] This embodiment drives the motion control interface to physically map and execute the quantified target motion parameter adjustments in the commands. For air vehicles, this embodiment corrects the differential speed of each rotor and the fuselage tilt angle in real time to actively counteract turbulence caused by airflow and generate a reverse attitude balancing torque. For ground vehicles, this embodiment dynamically adjusts the hydraulic damping coefficient or air spring stiffness of the active suspension system to absorb the impact energy from uneven road surfaces and smooth vehicle vibrations. Through this process, this embodiment transforms the digitized control parameters into corrected vehicle motion attitude parameters in the physical world, achieving real-time reconstruction of the vehicle's motion characteristics.

[0044] This embodiment, while performing corrective actions, initiates a rigorous closed-loop verification procedure, continuously collecting the latest microenvironmental state data and comparing it with physical stability constraint parameters. If the load state is detected as not returning to the specified range, this embodiment will continuously perform the next round of deviation calculation and parameter adjustment, forming a continuous adaptive correction cycle until the monitoring data indicates that the catering load has entered a stable safe threshold range. Based on this steady state, this embodiment monitors the current geographical location information in real time, and only releases the final delivery authority if the location accurately matches the preset delivery address and the load state continues to meet the standards, thereby completing a high-quality delivery process.

[0045] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0046] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A real-time catering delivery method fusing air-ground coordination and intelligent perception, characterized in that, The method comprises the following steps: acquiring a catering delivery request and determining a target catering load and corresponding physical stability constraint parameters according to the catering delivery request; collecting real-time environmental data of a delivery area and calculating an air-ground collaborative delivery strategy according to the physical stability constraint parameters; assigning a ground vehicle and an air vehicle as execution subjects from a transport resource pool according to the air-ground collaborative delivery strategy and planning a dynamic handover interval of the ground vehicle and the air vehicle in a time-space domain; controlling the ground vehicle and the air vehicle to move to the dynamic handover interval and, after entering the dynamic handover interval, controlling the air vehicle and the ground vehicle to establish a moving handover window with relative position synchronization; in the moving handover window, controlling the air vehicle and the ground vehicle to complete physical transfer of the target catering load in a state of maintaining collaborative motion and updating a subsequent delivery path according to a transport state after the transfer is completed to obtain an updated transport state and the subsequent delivery path; based on the updated transport state and the subsequent delivery path, collecting micro-environmental state data of the target catering load in real time; calculating a corresponding deviation degree according to the micro-environmental state data and the physical stability constraint parameters; generating adaptive steady-state control instructions based on the deviation degree; using the adaptive steady-state control instructions to adjust motion attitude parameters of the ground vehicle and the air vehicle currently carrying the target catering load in real time to maintain the state of the target catering load within the range defined by the physical stability constraint parameters until delivery is completed.

2. The instant catering delivery method of claim 1, wherein The method comprises the following steps: receiving the catering delivery request from a user terminal and structurally analyzing the catering delivery request to extract a list of catering category identifiers and component data corresponding to each catering category; traversing the list of catering category identifiers to retrieve an inherent physical form label uniquely matched with each catering category identifier in a pre-configured catering physical attribute database, wherein the inherent physical form label includes a liquid component proportion, a thermodynamic sensitivity level, and a rigid structure strength; based on the component data, performing weighted aggregation calculation on each retrieved inherent physical form label to construct a composite physical feature model representing the overall state of the target catering load; inputting the composite physical feature model into a pre-configured transportation tolerance mapping logic to calculate a critical threshold value required to maintain the physical properties of the target catering load and obtaining the physical stability constraint parameters; the physical stability constraint parameters include an allowed maximum linear acceleration, an allowed maximum angular velocity, and an allowed temperature fluctuation range; the expression of the composite physical feature model is: ； wherein, is the composite physical feature model; is the total number of food and beverage categories; is the weight of the th food and beverage category; is the intrinsic physical form label vector of the th food and beverage category; is the geometric parameter of the target food and beverage load box; is the weighted aggregation operation representing the feature vector.

3. The method according to claim 1, wherein, The method comprises the following steps: acquiring real-time environmental data through a city perception network, wherein the real-time environmental data includes meteorological data, real-time traffic flow data, and low-altitude airspace state information; performing airflow gradient analysis on the real-time meteorological data to obtain an air dynamic risk factor; performing congestion index calculation on the real-time traffic flow data to obtain a ground dynamic risk factor; obtaining a real-time dynamic risk matrix according to the low-altitude airspace state information, the air dynamic risk factor and the ground dynamic risk factor; performing coupling operation on the real-time dynamic risk matrix and the physical stability constraint parameter to generate a vehicle space-time operation constraint set, wherein the vehicle space-time operation constraint set includes: the air vehicle flight height limit based on the airflow gradient, and the vehicle maximum motion acceleration limit based on the physical stability constraint parameter; iteratively solving a ground driving path of the ground vehicle and an air flight path of the air vehicle according to a multi-agent space-time collaborative planning algorithm based on a preset optimization target set and a preset boundary condition, wherein the preset optimization target set includes: a primary optimization target and a secondary optimization target, the primary optimization target is to minimize the total delivery time, and the secondary optimization target is to maximize the catering load stability, and the preset boundary condition is the vehicle space-time operation constraint set; performing space-time synchronous calibration on the ground driving path and the air flight path obtained by iterative solving, determining a time and space coupling intersection, and defining the dynamic intersection interval of the ground vehicle and the air vehicle in the space-time domain based on the coupling intersection, so as to structurally integrate the air-ground collaborative delivery strategy.

4. The method according to claim 1, wherein, assigning the ground vehicle and the air vehicle as execution subjects from a transport resource pool according to the air-ground collaborative delivery strategy, and planning the dynamic intersection interval of the ground vehicle and the air vehicle in the space-time domain, including: performing matching evaluation on the load capacity and endurance capacity of all ground vehicles and air vehicles in the transport resource pool according to the load demand, mileage requirement and time limit set in the air-ground collaborative delivery strategy, to generate a vehicle candidate set for task execution; assigning the ground vehicle and the air vehicle as task execution subjects from the vehicle candidate set, and loading the air-ground collaborative delivery strategy into the ground vehicle and the air vehicle respectively; extracting the starting and ending space-time coordinates of the dynamic intersection interval from the air-ground collaborative delivery strategy, and configuring the starting and ending space-time coordinates as trigger parameters for collaborative positioning and attitude synchronization of the ground vehicle and the air vehicle; according to the trigger parameters, completing vehicle collaborative positioning system configuration and motion control parameter verification of the ground vehicle and the air vehicle, and determining the final planning and deployment of the dynamic intersection interval.

5. The method of claim 1, wherein the method further comprises: controlling the ground vehicle and the air vehicle to move to the dynamic intersection interval, and after entering the dynamic intersection interval, controlling the air vehicle and the ground vehicle to establish a mobile intersection window with relative position synchronization, including: controlling the ground vehicle and the air vehicle to drive at high speed to the dynamic intersection interval according to preset flight path planning data; real-time collecting and monitoring the absolute position of the ground vehicle and the air vehicle during high-speed driving, and calculating the vehicle proximity state relative to the starting coordinates of the dynamic intersection interval; continuously compare real-time positions of the ground carrier and the aerial carrier with time-space trigger parameters set for the dynamic handover zone until a high-precision cooperative mode switching instruction is generated; control flight attitude and speed of the aerial carrier according to the high-precision cooperative mode switching instruction to realize dynamic synchronization of the aerial carrier and the ground carrier in terms of movement direction, relative position and relative speed, and establish a mobile handover window in which relative positions are synchronized.

6. The method of claim 1, wherein the method further comprises: in the mobile handover window, control the aerial carrier and the ground carrier to complete physical transfer of the target catering load in a state of maintaining cooperative movement, and update a subsequent delivery path according to a transport capacity state after the transfer is completed, to obtain an updated transport capacity state and the subsequent delivery path, including: in the mobile handover window in which relative positions are synchronized, perform physical transfer of the target catering load to obtain a load physical transfer execution signal; real-time monitor a detachment state and a locking state of the target catering load according to the load physical transfer execution signal; when a signal of the load locking state is triggered and a signal of the detachment state is confirmed, determine that the physical transfer is successful, and generate a transfer success confirmation instruction; according to the transfer success confirmation instruction, collect real-time resource parameters of the ground carrier and the aerial carrier to obtain the updated transport capacity state, wherein the real-time resource parameters include: remaining power, real-time load, and current actual position; based on the updated transport capacity state and a preset remaining delivery task target, accurately calibrate the subsequent delivery path to obtain the updated transport capacity state and the calibrated subsequent delivery path; structurally integrate the updated transport capacity state and the calibrated subsequent delivery path to obtain the updated transport capacity state and the subsequent delivery path.

7. The method of claim 1, wherein the method further comprises: the micro-environment state data of the target catering load includes: real-time three-axis acceleration characteristic values, real-time three-axis angular velocity characteristic values, real-time attitude inclination data of a load box body, and temperature gradient data inside the load box body.

8. The method of claim 1, wherein the method further comprises: the calculation expression of the deviation degree is: ； Wherein, Dev is the deviation degree; is a dimension index, representing the main feature dimension corresponding to the real-time three-axis acceleration characteristic value, the real-time three-axis angular velocity characteristic value, the real-time attitude tilt data of the load box, and the temperature gradient data inside the load box in the microenvironment state data; is the first dimension risk weight coefficient, used to reflect the importance of the dimension to the overall stability; is the real-time characteristic value of the microenvironment state data in the first dimension; is the critical threshold of the physical stability constraint parameter in the first dimension.

9. The instant catering delivery method of fusion air-ground coordination and intelligent perception according to claim 1, characterized in that, generating an adaptive steady-state control instruction based on the deviation degree includes: compare the deviation degree with a preset hierarchical control threshold to determine a load risk level; select a steady-state control strategy based on the load risk level; calculate a target motion parameter adjustment amount for offsetting external disturbances according to requirements of the selected steady-state control strategy and in combination with a specific value of the deviation degree to obtain a control parameter set for compensating for disturbances; encapsulate and encode the selected steady-state control strategy and the control parameter set to generate the adaptive steady-state control instruction.

10. The instant catering delivery method of fusion air-ground coordination and intelligent perception according to claim 1, characterized in that, use the adaptive steady-state control instruction to real-time adjust motion attitude parameters of the ground carrier and the aerial carrier currently carrying the target catering load to maintain the state of the target catering load within a range defined by the physical stability constraint parameter until delivery is completed, including: send the adaptive steady-state control instruction to a motion control interface of the ground carrier and the aerial carrier currently carrying the target catering load through a low-delay communication network; The motion control interface adjusts the flight attitude and suspension damping in real time according to the target motion parameter adjustment quantity quantified in the adaptive steady-state control instruction, to obtain a corrected vehicle motion attitude parameter; Based on the corrected vehicle motion attitude parameter, the micro-environment state of the target catering load is continuously verified in a closed loop, and the vehicle motion attitude parameter is continuously adaptively corrected according to the physical stability constraint parameter, until the load state is stable and meets the standard, and delivery is completed.

Images