Distribution robot

Abstract

The invention provides a distribution robot which comprises a display screen, a storage structure located below the display screen, a chassis structure located below the storage structure and an environment monitoring structure and is used for obtaining environment information so as to obtain moving tracks of moving to all target positions through path planning based on the environment information. The environment monitoring structure comprises a depth camera, a laser radar and a pit measuring radar; the depth camera is fixed on the outer surface of the storage structure; the laser radar is fixed between the storage structure and the chassis structure; the pit measuring radar is fixed on the outer surface of the chassis structure. The cost map is obtained through the environment monitoring structure, the initial moving trajectory is planned based on the cost map, and the initial moving trajectory is optimized through various constraint cost values, so that the optimized distribution robot is kept stable in the transportation process.

Description

Technical Field

[0001] This application belongs to the field of robotics technology and relates to an intelligent mobile robot, particularly a delivery robot. Background Technology

[0002] Delivery robots are used in many fields such as logistics, catering, hotels, and healthcare. They are especially useful in the healthcare field, where there is a large and frequent demand for the transportation of supplies, such as food delivery to wards, drug transfer, medical equipment dispensing, medical dressings and sample delivery. The introduction of delivery robots can free medical staff from tedious transportation work, reduce unnecessary personnel contact and movement, help maintain a clean environment, reduce the probability of cross-infection, and improve the overall operational efficiency and safety management level of hospitals.

[0003] However, due to the complex ground conditions in hospitals, which may involve ramps, sliding door tracks, elevator connections, and slightly uneven ground, as well as the acceleration, deceleration, and turning of robots during their movement, the delivered items may easily shift, shake, or even spill. This not only causes waste and pollutes the hospital environment, but also affects the timing and dosage of patients' meals and medications, and may even lead to medical safety incidents.

[0004] Therefore, how to obtain a delivery robot that can transport goods smoothly is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] The purpose of this application is to provide a delivery robot that solves the problems in the prior art where delivery robots are not stable, and the transported items are prone to shifting, spilling, tipping over, or even causing adverse effects.

[0006] In a first aspect, this application provides a delivery robot for transporting at least one item to be delivered to at least one target location, comprising: a display screen for displaying an interactive interface; a storage structure located below the display screen for placing each of the items to be delivered; a chassis structure located below the storage structure, including at least two drive units for driving the delivery robot to move to each of the target locations; and an environmental monitoring structure for acquiring environmental information to obtain a trajectory for moving to each of the target locations based on the environmental information through path planning, wherein the environmental monitoring structure includes a depth camera, a lidar, and a pit-detecting radar; the depth camera is fixed to the outer surface of the storage structure; the lidar is fixed between the storage structure and the chassis structure; and the pit-detecting radar is fixed to the outer surface of the chassis structure.

[0007] In one embodiment of this application, for any target location, an initial running trajectory is generated based on the environmental information to move to the target location, and various constraint costs are calculated based on the initial running trajectory. The initial running trajectory is then optimized based on each constraint cost to obtain the final running trajectory.

[0008] In one embodiment of this application, the various constraint costs include the endpoint constraint cost. The calculation method of the endpoint constraint cost includes: obtaining the distance between the endpoint of the initial running trajectory and the corresponding target position as the endpoint error; obtaining the corresponding error weight based on the endpoint error, combined with the stopping zone radius and the transition zone radius; the stopping zone radius is the maximum error distance value between the delivery robot and the corresponding target position when the robot stops; the transition zone radius is the minimum distance value at which the delivery robot approaches the target position at maximum speed; obtaining the error weight and the endpoint error to obtain the effective error; and calculating the endpoint constraint cost based on the effective error; wherein, the smaller the effective distance, the smaller the endpoint constraint cost.

[0009] The endpoint constraint in this application achieves a smooth transition from soft to hard constraints by constructing an endpoint constraint cost function based on error weights. This approach improves optimization convergence stability, avoids oscillations near the endpoint, and enhances trajectory executability.

[0010] In one embodiment of this application, when the endpoint error is greater than or equal to the transition zone radius, the error weight is the maximum weight value; when the endpoint error is less than the transition zone radius but greater than the stop zone radius, calculating the error weight includes: obtaining the difference between the endpoint error and the stop zone radius as a first difference, and obtaining the difference between the transition zone radius and the stop zone radius as a second difference, and using the square of the ratio of the first difference and the second difference as the error weight; when the endpoint error is less than the stop zone radius, the error weight is the minimum weight value.

[0011] In one embodiment of this application, the various constraint values ​​include a backward constraint value. When there is backward motion in the initial running trajectory, the backward constraint value is obtained based on the movement distance of the backward motion.

[0012] In one embodiment of this application, the storage structure includes a storage compartment and a dynamic weighing unit disposed directly below the storage compartment for acquiring information on changes in the items to be delivered within the storage compartment.

[0013] In one embodiment of this application, the dynamic weighing unit includes a sensor array.

[0014] In one embodiment of this application, the storage compartment includes: an outer compartment fixedly connected to both the display screen and the chassis structure, and an inner compartment located inside the outer compartment and connected to the outer compartment via multiple sets of motors; the multiple sets of motors are used to rotate the inner compartment; the dynamic weighing unit further includes an inertial measuring instrument and a gyroscope fixed below the inner compartment.

[0015] In one embodiment of this application, the storage compartment further includes an inner tray and an outer tray. The inner compartment is provided with a slide rail, and the outer tray is slidably connected to the slide rail. The inner tray is used to place each of the items to be delivered. When the delivery robot transports the items to be delivered, it is placed in the outer tray. When the delivery robot is loading the items to be delivered, it is taken out and placed outside the storage compartment.

[0016] In one embodiment of this application, the storage structure further includes a camera fixed to the inner surface of the inner compartment.

[0017] As described above, this application provides a delivery robot that obtains a cost map through an environmental monitoring structure, plans an initial operating trajectory based on the cost map, and optimizes the initial operating trajectory through various constraint costs. This ensures that the optimized delivery robot remains stable during transportation, avoiding tipping over. Furthermore, the delivery robot includes an inner compartment and an outer compartment. The inner compartment is fixed inside the outer compartment by a motor, which rotates the inner compartment. This ensures that even when the delivery robot as a whole shakes, the inner compartment remains stable, preventing spillage of medicines or food, thereby achieving efficient and convenient delivery of goods. Attached Figure Description

[0018] Figure 1 The diagram shown is a structural schematic of a delivery robot according to an embodiment of this application.

[0019] Figure 2 The diagram shown is a structural schematic of a storage structure according to an embodiment of this application.

[0020] Figure 3 The diagram shown is a schematic representation of a sensor array according to an embodiment of this application.

[0021] Explanation of reference numerals in the attached figures

[0022] 100: Display screen; 200: Storage structure; 210: Storage compartment; 211: Camera; 212: Outer compartment; 213: Inner compartment; 2131: Slide rail; 214: Motor; 215: Inner tray; 216: Outer tray; 220: Dynamic weighing unit; 221: Sensor; 222: Automatic door; 223: Automatic door lock; 300: Chassis structure; 401: Depth camera; 402: LiDAR; 403: Pit detection radar; 404: Ultrasonic sensor. Detailed Implementation

[0023] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0024] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0025] Delivery robots can effectively reduce labor costs and free up workers from tedious transportation tasks in scenarios with high-volume and frequent material delivery needs. However, in complex transportation environments, maintaining stability during delivery by robots is challenging, which can easily lead to items shifting, shaking, or even spilling, affecting the delivery efficiency.

[0026] To address the technical problems existing in the prior art, the following embodiments of this application provide a delivery robot, including an environmental monitoring structure. The robot acquires environmental information through a depth camera, lidar, and pit-finding radar, so that it can avoid obstacles during transportation and plan a good running trajectory to reduce the shaking of the robot during transportation, thereby achieving stable and safe transportation of goods and improving the effectiveness of the delivery robot.

[0027] The following embodiments of this application provide a delivery robot, which can be applied to transportation scenarios such as logistics, catering, hotels, and medical care, but this application does not specifically limit its use. The following description will take the use of a delivery robot in a hospital for the delivery of medicines or meals as an example.

[0028] The principle and implementation method of a delivery robot according to this embodiment will be described in detail below with reference to the accompanying drawings, so that those skilled in the art can understand the delivery robot of this embodiment without creative effort.

[0029] like Figure 1 As shown, the delivery robot provided in this application is used to transport at least one item to be delivered to at least one target location. For example, in a medical transportation scenario, it is used to transport various medicines or meals to different hospital beds for patients to pick up.

[0030] Specifically, the delivery robot includes: a display screen, a storage structure below the display screen, and a chassis structure below the storage structure. The display screen is used to show the interactive interface to guide patients or nurses to retrieve the corresponding medications or meals, avoiding accidents caused by misplacing items. The storage structure is used to hold the items to be delivered, such as medications or meals. The chassis structure includes at least two drive units for moving the delivery robot to each target location, i.e., each hospital bed, to facilitate patients in retrieving their meals or medications.

[0031] Based on this, the delivery robot can deliver each item to its designated location for patients to pick up their meals or medications. However, due to the complex environment of hospitals and the high volume of foot traffic, to prevent the delivery robot from colliding with obstacles or tipping over, thus avoiding spillage of items, this embodiment includes an environmental monitoring structure to acquire environmental information. This allows the delivery robot to plan its path based on this information, enabling it to smoothly reach its designated locations. Specifically, the environmental monitoring structure includes a depth camera, a lidar, and a pit-detection radar. The depth camera is fixed to the outer surface of the storage structure to acquire information about obstacles at higher elevations, such as within the robot's coverage area. The lidar is fixed between the storage structure and the chassis structure to acquire information about obstacles on the ground surface. The pit-detection radar is fixed to the outer surface of the chassis structure to acquire information about the flatness of the ground. Based on the environmental information acquired by the depth camera, lidar, and pit-detection radar, a cost map is generated. Path planning is then performed based on this cost map to obtain the initial operating trajectory.

[0032] Optionally, the environmental monitoring structure may also include an ultrasonic sensor for short-range measurement of the distance between the obstacle and the robot.

[0033] Furthermore, to improve the smoothness of robot operation, it is necessary to reduce the number of turns and the degree of curvature in the initial trajectory. Specifically, the cost map generated based on environmental information is a grid map and includes the initial cost value of each grid point. For any point on the cost map, if the initial cost value of the point and its eight surrounding points is less than a preset initial cost value threshold, the point is considered a simplifiable point. It is determined whether each point on the initial trajectory is a simplifiable point, and when multiple consecutive simplifiable points exist, these simplifiable points are extracted. When the trajectory length connecting these simplifiable points exceeds a preset simplifiable line segment length threshold, these simplifiable points are straightened, i.e., the first and last points of these consecutive simplifiable points are connected, and the resulting straight line is used as the simplified trajectory to simplify the initial trajectory and reduce the number of turns and the degree of curvature. Those skilled in the art can set the initial cost value threshold and the simplifiable line segment length threshold according to actual needs; this embodiment does not impose specific limitations.

[0034] It should be noted that, in order to further ensure the smooth operation of the delivery robot, the initial running trajectory is optimized. Specifically, the cost of various constraints is calculated based on the initial running trajectory, and the sum of the costs of various constraints is used as the total constraint cost to optimize the initial running trajectory. The trajectory corresponding to the minimum total constraint cost is taken as the final running trajectory.

[0035] For example, various constraint costs include safety distance constraint costs. Specifically, for a cost map constructed based on environmental information obtained from an environmental monitoring structure, and thereby obtaining the single-point safety distance cost corresponding to each point on the initial running trajectory based on obstacle distribution, when the minimum distance between the point and the obstacle is less than or equal to a preset safety distance, the single-point safety distance cost is calculated based on the difference between the minimum distance between the point and the obstacle and the safety distance; when the minimum distance between the point and the obstacle is greater than the preset safety distance and less than or equal to a preset expansion distance, the single-point safety distance cost is calculated based on the difference between the minimum distance between the point and the obstacle and the expansion distance; when the minimum distance between the point and the obstacle is greater than the preset expansion distance, the single-point safety distance cost of the point is a preset value, for example, 0. Specifically, those skilled in the art should know how to calculate the single-point safety distance cost of each point based on the preset safety distance and expansion distance, and this embodiment does not impose specific limitations here. For example, the preset safety distance is the radius of the tangent circle of the robot body, and the preset expansion distance is 0.5m greater than the safety distance. The sum of the single-point safety distance values ​​corresponding to each point on the initial running trajectory is taken as the safety distance value of the initial running trajectory. The trajectory with the minimum sum of the safety distance value and the values ​​of other types of constraints is taken as the final running trajectory, so as to achieve optimization.

[0036] In some optional implementations, various constraint costs include endpoint constraint costs, used to constrain the endpoint position of the initial trajectory. Specifically, the endpoint constraint cost is calculated using the Proximal Hybrid Regularized Augmented Lagrangian Method (PHR-ALM) to calculate the cost of equality constraints, including:

[0037] S1, obtain the endpoint error vector between the endpoint of the initial running trajectory and the corresponding target position.

[0038] For example, the endpoint error vector is:

[0039]

[0040] in, For the endpoint error vector, This is the endpoint of the current initial running trajectory. The desired endpoint is the endpoint of the trajectory under ideal conditions.

[0041] Optionally, the endpoint error vector is used to characterize the deviation between the current trajectory endpoint and the target endpoint. It can be a two-dimensional position error, a three-dimensional position error, or a higher-dimensional extended error vector that includes attitude components. This embodiment does not make specific limitations here.

[0042] It should be noted that the endpoint constraint in this embodiment is actually an equality constraint objective in the optimization model, that is, the endpoint error vector is required to approach zero; however, in actual execution, considering factors such as positioning error, limited control frequency, robot braking capability, and mechanism response lag, the equality constraint can be continuously relaxed by setting an error tolerance range, thereby improving actual reachability.

[0043] S2 constructs the stopping region and transition region, and calculates the error weights;

[0044] Based on the endpoint error vector e, we define the error norm r, where r can be the Euclidean norm of e, satisfying:

[0045] ;

[0046] Furthermore, set the stop zone radius. and the radius of the transition zone And satisfy > .

[0047] Wherein, the radius of the stopping area A threshold used to characterize the positional accuracy threshold in which the robot believes it has reached its destination; transition zone radius. Used to characterize the transition range from relaxed constraints to strict constraints.

[0048] The purpose of setting a transition zone is to avoid immediately applying a strong endpoint stop constraint when the robot gradually approaches the endpoint. Instead, the endpoint constraint strength is gradually weakened as the endpoint error decreases, allowing the robot sufficient time to complete posture adjustment, speed decay, and position correction. This avoids the problem of "not being able to reach the destination" due to an excessively small threshold or "not being able to stop accurately" due to an excessively large threshold.

[0049] When the error norm is less than the transition zone radius but greater than the stopping zone radius, the error weight is calculated, including: obtaining the difference between the endpoint error and the stopping zone radius as the first difference, and obtaining the difference between the transition zone radius and the stopping zone radius as the second difference. The square of the ratio of the first difference and the second difference is used as the error weight, i.e.:

[0050]

[0051] in, For error weights, Let the error norm be... The radius of the stopping area, The radius of the transition region is given. It should be understood that the above quadratic function is only one of the preferred embodiments. In practice, linear functions, cubic functions, piecewise functions, exponential functions, sigmoid functions, or other monotonic functions can also be used to implement the transition modulation concept of this application.

[0052] When the endpoint error is less than the radius of the stopping zone, the error weight is the minimum weight value, which is 0 for example.

[0053] S3, based on the error weights and the endpoint error vector, obtain the effective endpoint error.

[0054] Specifically, the effective error is calculated as follows:

[0055]

[0056] in, This is used to characterize the endpoint deviation after soft-hard transition modulation. By weighting the endpoint error, the endpoint constraint cost can change smoothly as it approaches the endpoint, avoiding trajectory jitter caused by abrupt cost changes.

[0057] S4, calculate the endpoint constraint cost based on the effective error.

[0058] Based on the augmented Lagrange method, the effective endpoint error is analyzed. Cost calculations are performed to obtain the endpoint constraint cost.

[0059] In some alternative implementations, the endpoint constraint cost can be constructed in the following form:

[0060]

[0061] in, The endpoint constraint is the value. Coordinates are The effective error of the point, Coordinates are The penalty parameter for the point is updated in the following way: , The penalty parameter growth factor, Coordinates are The Lagrange multiplier vector of the point is updated in the following way: .

[0062] In other alternative implementations, proximal terms, regularization terms, or additional stabilizing terms may be introduced to construct a numerically stable PHR-ALM optimization form, that is, adding a proximal regularization term related to the current position or the change in the current trajectory to the Lagrange multiplier term and the penalty term. This embodiment does not limit this.

[0063] Based on this, the endpoint position error can be continuously mapped to the endpoint constraint cost and embedded into a unified optimization framework, which can then be used for subsequent trajectory iteration optimization. At the same time, by setting transition and stopping zones, it is prevented from approaching the target position too quickly before reaching it and then stopping abruptly after reaching the target position. This allows the delivery robot to reach the target position in a precise, stable, and smooth manner, preventing the medicine or food from shaking or spilling.

[0064] Furthermore, to further improve the robot's stability upon reaching the endpoint, an endpoint angle constraint is also included, that is, constraining the direction of the robot's angular velocity when it reaches the endpoint region. Specifically, when At that time, punishment will be imposed, among which, The derivative of the angle at the initial moment with respect to time is the initial angular velocity for this trajectory optimization. This refers to the angular velocity published in the previous iteration, specifically the final angular velocity published during the last trajectory optimization. The penalty function is:

[0065]

[0066] in, , A pre-set smoothing factor; The parameters of the pre-set penalty function curve; For another pre-set penalty function curve parameter.

[0067] The weight value of the penalty function for switching the angular velocity direction is set as follows:

[0068] When the endpoint error is greater than twice the radius of the transition zone, the weight is 0;

[0069] When the endpoint error is greater than the transition zone radius but less than twice the transition zone radius, the weight is:

[0070]

[0071] in, For the penalty function weights, The magnitude of the endpoint error, i.e. , The radius of the stopping area, The radius of the transition zone;

[0072] When the endpoint error is less than the transition zone radius, the error weight is 1.

[0073] Based on this, the oscillation behavior at the destination can be suppressed, thereby improving the operational stability of the delivery robot and preventing medicine or food from spilling.

[0074] In some optional implementations, the various constraint costs include a backward constraint cost. When backward motion exists in the initial trajectory, the backward constraint cost is obtained based on the distance of that backward motion. Specifically, the backward constraint cost is calculated as follows:

[0075]

[0076] in, To reduce the value of constraints, The derivative of the initial trajectory with respect to time is the initial linear velocity during this trajectory optimization. For the linear velocity published in the previous cycle, when When the value is negative, it means that the delivery robot is switching between forward and backward. In this case, the value of the product is used as the cost of the backward constraint.

[0077] When the robot switches between forward and backward movement, that is... When the value is negative, a penalty is applied. The penalty function is described in the previous section on the penalty function for switching angular velocity directions. Replace with For details, please refer to the aforementioned content, which will not be repeated here.

[0078] Based on this, by reducing the forward and backward switching behavior of the delivery robot by constraining the backward value, we can avoid frequent reversals that could cause the robot to bounce and thus spill medicine or food.

[0079] In some alternative implementations, such as Figure 2 As shown, the storage structure includes a storage compartment and a dynamic weighing unit located directly below the storage compartment to monitor changes in the weight of each item to be delivered within the compartment. Specifically, the dynamic weighing unit is positioned below the storage compartment to track changes in the weight of the entire compartment, thereby indicating whether any medicines or meals within the compartment have been removed. This information helps the delivery robot to restart its journey after stopping at the target location.

[0080] For example, such as Figure 3 As shown, the dynamic weighing unit includes a sensor array. Since the delivery robot can deliver multiple medications or meals at a time, a corresponding sensor is placed directly below each location in the storage compartment where medications or meals are placed to accurately detect the retrieval status of medications or meals at each location. For example, the sensor array is a pressure sensor array, which detects changes in pressure at various locations in the storage compartment to indicate whether medications or meals at those locations have been retrieved.

[0081] Furthermore, cameras can be installed on the inner surface of the upper part of the storage compartment to capture images of the items to be delivered inside. Based on machine vision technology, the retrieval status of medicines or meals in the storage compartment can be obtained, assisting the delivery robot to stop at the target location and then set off again.

[0082] In some alternative implementations, to further ensure that medicines or food in the storage compartment will not spill, the dynamic weighing unit is also used to obtain the tilt status of the storage compartment.

[0083] Specifically, such as Figure 2 As shown, the storage compartment includes an outer compartment and an inner compartment. The outer compartment is fixedly connected to the display screen and chassis structure. The inner compartment is located inside the outer compartment and is connected to the outer compartment via multiple sets of motors. The inner compartment is equipped with a tray for placing items to be delivered. The motors drive the inner compartment to rotate. When the outer compartment moves and shakes with the delivery robot as a whole, the motor sets adjust the inner compartment so that it rotates in the opposite direction of the outer compartment's shaking, ensuring that the inner compartment remains stable.

[0084] The dynamic weighing unit also includes an inertial measurement unit (not shown in the figure) fixed below the inner compartment. The inertial measurement unit includes a gyroscope. The inertial measurement unit is used to obtain the sway angle and direction of the inner compartment and outputs the detected data to the processing center of the delivery robot (not shown in the figure). The processing center then sends corresponding commands to each motor so that the motor moves and drives the inner compartment to move, thereby keeping the inner compartment stable.

[0085] To facilitate understanding by those skilled in the art, the following will illustrate how the movement of the motor keeps the interior compartment stable. Please refer to... Figure 2When the outer compartment sways to the left, meaning the delivery robot as a whole sways to the left, the inner compartment also sways to the left. The sway angle is detected by the inertial measurement unit (IMU) and gyroscope and transmitted to the processing center. Based on the acquired data, the processing center sends commands to each motor. The left-side motor extends, driving the inner compartment to rotate to the right. The extent of the left-side motor's extension is calculated based on the sway angle detected by the IMU and gyroscope; the larger the sway angle, the longer the left-side motor extends. This ensures that the inner compartment ultimately remains stable, thus preventing spillage of medicine or food.

[0086] In some alternative implementations, such as Figure 2 As shown, the trays inside the cabin include an inner tray and an outer tray. The outer tray is slidably connected to the side wall of the inner cabin via a slide rail, while the inner tray is used to place various meals or medicines. The inner and outer trays are not fixedly connected, allowing the inner tray to be placed inside the outer tray during transport. When medicines or meals need to be filled, the inner tray can be removed, placed, and then placed back into the outer tray. This convenient operation facilitates filling items for delivery, promotes cleanliness, improves delivery efficiency, and reduces operational complexity.

[0087] Furthermore, the inner tray is equipped with a handle for easy removal.

[0088] In some alternative implementations, such as Figure 1 As shown, the outer compartment is equipped with an automatic door and an automatic door lock. When the delivery robot is transporting medicines or items, the automatic door lock closes and the automatic door closes. When the delivery robot arrives at the target location, the automatic door lock automatically opens and the automatic door pops open so that the patient can take the medicine or food.

[0089] In summary, the delivery robot provided in this application acquires environmental information through depth cameras, lidar, and pit-finding radar to generate a cost map, plan the initial operating trajectory, and optimize the initial operating trajectory through various constraint costs, thereby improving the stability of the delivery robot during transportation. Furthermore, it is equipped with an inner and outer compartment connected by motors, allowing the inner compartment to rotate via motors, further ensuring its stability and preventing spillage of medicines or food. This effectively improves the delivery robot's delivery performance and has high industrial application value.

[0090] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0091] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A delivery robot for transporting at least one item to be delivered to at least one target location, characterized in that, include: A display screen is used to show the interactive interface; A storage structure, located below the display screen, is used to place each of the items to be delivered; The chassis structure, located below the storage structure, includes at least two drive units for driving the delivery robot to each of the target positions; An environmental monitoring structure is used to acquire environmental information and, based on the environmental information, to obtain the running trajectory to each of the target positions through path planning. The environmental monitoring structure includes a depth camera, a lidar, and a pit-detecting radar. The depth camera is fixed on the outer surface of the storage structure. The lidar is fixed between the storage structure and the chassis structure. The pit-detecting radar is fixed on the outer surface of the chassis structure.

2. The delivery robot according to claim 1, characterized in that, For any of the target locations, an initial trajectory is generated based on the environmental information to move to the target location, and various constraint costs are calculated based on the initial trajectory. The initial trajectory is then optimized based on each constraint cost to obtain the final trajectory.

3. The delivery robot according to claim 2, characterized in that, The various constraint costs include the endpoint constraint cost, and the calculation method for the endpoint constraint cost includes: The distance between the endpoint of the initial trajectory and the corresponding target position is obtained as the endpoint error; Based on the endpoint error, and combined with the stopping zone radius and the transition zone radius, the corresponding error weight is obtained; the stopping zone radius is the maximum error distance value between the delivery robot and the corresponding target position when the delivery robot stops; the transition zone radius is the minimum distance value at which the delivery robot approaches the target position at maximum speed. Based on the error weight and the endpoint error, the error weight is obtained, and the effective error is obtained; The endpoint constraint cost is calculated based on the effective error; wherein, the smaller the effective distance, the smaller the endpoint constraint cost.

4. The delivery robot according to claim 3, characterized in that, When the endpoint error is greater than or equal to the transition zone radius, the error weight is the maximum weight value; When the endpoint error is less than the transition zone radius but greater than the stop zone radius, the error weight is calculated, including: obtaining the difference between the endpoint error and the stop zone radius as a first difference, and obtaining the difference between the transition zone radius and the stop zone radius as a second difference, and taking the square of the ratio of the first difference and the second difference as the error weight; When the endpoint error is less than the radius of the stopping zone, the error weight is the minimum weight value.

5. The delivery robot according to claim 3, characterized in that, The various constraint costs also include the endpoint angle cost, which is the product of the starting angular velocity of the delivery robot in the current trajectory optimization and the angular velocity published in the previous cycle. When the cost of the endpoint angle is less than zero, a penalty is applied based on a preset penalty function.

6. The delivery robot according to claim 1, characterized in that, The storage structure includes a storage compartment and a dynamic weighing unit located directly below the storage compartment to obtain information on changes in the items to be delivered within the storage compartment.

7. The delivery robot according to claim 6, characterized in that, The dynamic weighing unit includes a sensor array.

8. The delivery robot according to claim 7, characterized in that, The storage compartment includes: an outer compartment fixedly connected to both the display screen and the chassis structure, and an inner compartment located inside the outer compartment and connected to the outer compartment via multiple sets of motors; the multiple sets of motors are used to rotate the inner compartment; the dynamic weighing unit also includes an inertial measuring instrument and a gyroscope fixed below the inner compartment.

9. The delivery robot according to claim 8, characterized in that, The storage compartment also includes an inner tray and an outer tray. The inner compartment is equipped with a slide rail, and the outer tray is slidably connected to the slide rail. The inner tray is used to place each of the items to be delivered. When the delivery robot transports the items to be delivered, it is placed in the outer tray. When the delivery robot is loading the items to be delivered, it is taken out and placed outside the storage compartment.

10. The delivery robot according to claim 8, characterized in that, The storage structure also includes a camera fixed to the inner surface of the inner compartment.

Images