Microgravity near space planning method and system for service robot in space station cabin

Abstract

The invention discloses a microgravity near space planning method and system for a service robot in a space station cabin, and relates to the technical field of human body near space navigation planning. The method is provided for solving the problems that an existing algorithm cannot accurately navigate in a human body near space, and influences of human body subjective comfort on robot navigation planning are ignored. According to the technical key points, kinematics modeling is carried out on human joints in a microgravity environment, and a dynamic collision avoidance potential field of a human body near space is generated according to a modeling result. A subjective comfort potential field of a human body near space is designed based on subjective comfort evaluation of a person in a perception range. Combining two potential fields, namely a dynamic collision avoidance potential field of the human body near space and a subjective comfort potential field of the human body near space, to finish path planning of the robot in the human body near space under the microgravity environment. The method is suitable for human body near space sensing and path planning scenes in a narrow space and a microgravity environment, and safety and comfort planning of movement of the robot in the human body near space is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of human near-body space navigation planning, and particularly relates to a micro-gravity near-body space planning method for a space station cabin service robot. BACKGROUND

[0002] The research on space robots has been booming since the beginning of this century. Today, humans have developed various types of robots in space stations, covering all aspects of space research. Among them, the most representative and intelligent are space station cabin service robots. Such robots are divided into space flying robots and space hovering robots, and their main service object is astronauts, aiming to assist them in completing various tasks in the cabin, including data analysis, environment and personnel behavior monitoring, and accompanying flight shooting. Thanks to the rapid development of artificial intelligence technology in recent years, space station cabin service robots will play an increasingly important role in human space research.

[0003] Obstacle avoidance capability is one of the most critical capabilities of space station cabin service robots, and its effect is directly related to the safety of cabin personnel and equipment. Therefore, in the research process, various path planning methods for space station cabin robots are constantly updated and improved. However, as we all know, human movement is characterized by a wide range, high speed, and complex trajectory under the control of various joints and muscles. Therefore, in view of the characteristics of human behavior, teams have successively carried out research on robot obstacle avoidance and path planning related to humans on the ground using unmanned aerial vehicles or robotic arms as platforms. Although such methods take into account the movement of the human body to some extent, they are limited to treating the human body as a simple geometric shape (such as a sphere) for processing. The premise of this is that the robot has enough space to avoid, but obviously this is not applicable in the cramped environment of a space station cabin.

[0004] Under the influence of the space micro-gravity environment, human movement will change significantly. Many experiments on human movement in space have shown that due to the existence of micro-gravity environment, the control ability of muscles on the body's limbs is weakened. Specifically, when stationary, the human body basically does not need muscles to maintain its posture, and when moving, the speed of the human body is also lower than on the ground. In contrast, on the ground, whether stationary or moving, the human body needs to deploy all the muscles to resist the action of gravity. This difference has a huge impact on the human body's posture and movement trend. In addition, for a person moving on the ground, the degrees of freedom mainly exist in two translational directions and one rotational direction. However, for a person moving in space, since he is suspended in the air, the degrees of freedom in all six directions (three translational directions and three rotational directions) are present. Obviously, human movement in space will be more complex and less predictable.

[0005] Currently, obstacle avoidance methods for near-space environments are limited to human movement under normal gravity, while research on related methods in microgravity environments is lacking. Existing patent literature on near-space planning in space station scenarios is scarce. For example, document number CN118550226B discloses a method, system, device, and medium for vectorized representation of mission planning state space, relating to the field of state space description technology for mission planning. Its method includes: obtaining the current state of any action device on the space station; determining the state space vector corresponding to any action device on the space station based on the current state, whereby the state space vector is represented by a quintuple. This invention uses a quintuple representation method to represent the state space vector, adding a description of a predetermined state within a future time interval associated with the current state. This enables modeling of complex constraints between actions under multi-timeline parallel execution, expanding the application scope of state space description methods and providing a foundation for solving complex temporal planning problems using state space planning methods. It possesses good flexibility and practicality. However, it does not address the issue of near-space path planning for service robots within the space station, nor does it provide a solution.

[0006] In addition to predicting and avoiding human movement, the influence of human subjective consciousness on action should also be considered when planning close-range spatial paths. In daily life, humans tend to maintain a certain distance from objects in their surroundings. In other words, when an object crosses this boundary and approaches, people feel uncomfortable and try to avoid it. Besides distance, the speed of the object is also an influencing factor. When an object approaches rapidly, people usually panic and change their actions to avoid it. Existing methods are feasible when there is a large available space, but they may be unsolvable in confined environments such as inside a space station. Therefore, it is necessary to design planning methods for confined three-dimensional spaces. Summary of the Invention

[0007] The technical problem to be solved by this invention is:

[0008] To address the problems of existing algorithms failing to provide accurate navigation in close proximity to the human body and neglecting the impact of human subjective comfort on robot navigation planning, this invention proposes a microgravity close proximity space planning method and system for service robots within space stations. This enables safe and comfortable planning of robot movement within close proximity to the human body. This invention is applicable to close proximity space perception and path planning scenarios in confined spaces and microgravity environments.

[0009] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0010] A microgravity close-range space planning method for in-vehicle service robots on a space station, comprising the following steps:

[0011] The method is used for path planning of human body near space in microgravity environment, makes the robot successfully reach the target point after passing through the human body near space from the starting point, and makes the robot effectively avoid the subjective perception area of the human body and the space possibly occupied by the limbs in the future, and the method comprises the following steps:

[0012] Step one, kinematics modeling of human body joints in microgravity environment is carried out, and a dynamic collision avoidance potential field of human body near space is generated according to the modeling result;

[0013] Step two, a subjective comfort potential field of human body near space is designed based on the subjective comfort evaluation of the human body in the perception range;

[0014] Step three, the path planning of the robot in the human body near space in the microgravity environment is completed by combining the dynamic collision avoidance potential field of the human body near space and the subjective comfort potential field of the human body near space.

[0015] The microgravity near space planning method for the space station cabin service robot provided by the application is more specific, and the method comprises the following steps:

[0016] Step one, kinematics modeling of human body joints in microgravity environment is carried out, and a dynamic collision avoidance potential field of human body near space is generated according to the modeling result, and the specific process is as follows:

[0017] The human body structure is identified and subdivided into multiple parts through an image algorithm, the relationship between the limbs of the human body can be regarded as a radial structure which decreases step by step from the trunk outward, the trunk is regarded as a first-level limb, and the head, the upper arm and the upper leg connected with the first-level limb are regarded as second-level limbs; similarly, the lower arm and the lower leg are regarded as third-level limbs; a first-level limb coordinate system is constructed with the center of gravity of the human body as the origin, and second-level limb coordinate systems and third-level limb coordinate systems are respectively constructed with the joints as the origins; all the above coordinate systems are right-handed systems; the movement of each level of limbs includes consistent translational movement and rotational movement with the connection point of the previous level as the center, the translational movement of the human body is calculated through the movement trajectory of the center of gravity, and the rotational movement of the limbs can be regarded as the rotation of a rigid body, and the kinematics modeling of the human body joints in the microgravity environment is based on the above.

[0018] The movement of the human body in the microgravity environment is inflated according to the limb contour, the movement volume of the space station cabin service robot itself is considered, in order to avoid collision between the robot and the human body, the above prediction result is inflated based on the generation method of the electric potential field, and a human body dynamic collision avoidance potential field is obtained;

[0019] Step two, a subjective comfort potential field of human body near space is designed based on the subjective comfort evaluation of the human body in the perception range, and the specific process is as follows:

[0020] Assumption: based on the compensation of auditory perception to visual perception, the perception range of human body near space is approximately spherical, and the center of the sphere coincides with the center of the human head;

[0021] Based on the above assumption of the spherical subjective perception range with the human head as the center, a subjective comfort potential field of the human body near space is established, and the subjective comfort potential field is used to make the robot in the potential field subject to a subjective repulsive force, so that the planned route avoids the human head;

[0022] Step three, combine the dynamic collision avoidance potential field of the human body near space and the subjective comfort potential field of the human body near space, complete the path planning of the robot in the human body near space in the microgravity environment, specifically:

[0023] Synthesize the collision avoidance potential field and the subjective comfort potential field of the human body near space, equivalent the robot to a particle moving in the potential field, obtain the resultant force suffered by the robot, when the robot is too far away from the target, use a conical function to reduce the value of the attractive force to effectively avoid the attractive force being too large; when the robot is relatively close to the target, use a parabolic function to ensure that the robot smoothly reaches the target point, finally, the planned path is obtained through the iteration of the positions of adjacent time points.

[0024] The judgment method of whether the robot is too far away from the target or relatively close to the target is: define a global distance Rg, use a conical function when greater than Rg, and use a parabola when less than or equal to; or, the target distance is taken in the range of 1m-1.5m, and the range is adjusted according to the actual effect.

[0025] The present application has the following beneficial technical effects:

[0026] The present application is based on a service robot in a space station cabin, and proposes a human body near space perception and path planning method in a narrow space and a microgravity environment. The main contributions of this paper are as follows:

[0027] 1) A method for generating a human body dynamic collision avoidance potential field in a narrow environment in a cabin in a microgravity state is proposed, which is based on muscle-joint dynamics in a microgravity environment to generate an occupied space of the limbs. And the human body dynamic collision avoidance potential field is generated by expanding along the outline of the occupied space of the limbs.

[0028] 2) A subjective comfort potential field with the human head as the center is constructed, which is used for correcting the flight trajectory pointing to the human head.

[0029] 3) The human body motion prediction result is fused to generate a dynamic collision avoidance potential field and a subjective comfort potential field, and the human body near space path planning in a narrow space and a microgravity environment is realized, and a cabin robot flight route meeting the safety requirements and comfort requirements is obtained.

[0030] In conclusion, the application improves the safety and comfort of the path planning of the human body in the narrow microgravity environment, and has high engineering application value. BRIEF DESCRIPTION OF DRAWINGS

[0031] The application can be better understood by reference to the following description taken in connection with the accompanying drawings, which are included as a part of this specification, illustrate preferred embodiments of the application and, together with the detailed description, serve to explain the principles and advantages of the application.

[0032] Figure 1 A schematic diagram for the main limbs and joints of the human body.

[0033] Figure 2 A schematic diagram for the world system and the limb system of the human body.

[0034] Figure 3 A schematic diagram for the translation and rotation of the limbs by using axial vectors.

[0035] Figure 4 A result diagram of the space occupation probability of the human body movement in the microgravity environment.

[0036] Figure 5 A schematic diagram of the dynamic collision avoidance potential field of the human body in the microgravity environment.

[0037] Figure 6 A schematic diagram of the influence of the subjective potential field of the human body in the near-body space.

[0038] Figure 7 A schematic diagram of the effect of the subjective potential field of the human body.

[0039] Figure 8 A schematic diagram of the path planning result of the near-body space of the human body in the microgravity environment.

[0040] Fig. 9 is a comparison diagram of the planning effect of the near-body space potential field and the artificial potential field, in which (a) is the planning effect of the near-body space potential field, and (b) is the planning effect of the artificial potential field. The path planning is performed from nine directions around the human body; in the two diagrams, the red serial number represents the starting point, the green serial number represents the ending point, and the same serial number of the starting point and the ending point represents the same planning. DETAILED DESCRIPTION

[0041] GIVE COMBINATION Figure 1 -9, the implementation process of the microgravity near-body space planning method for the space station cabin service robot is described as follows:

[0042] Step one: kinematics modeling of the human body joints in the microgravity environment and construction of the human body near-body space collision avoidance potential field:

[0043] Human body is a complex motion system, there are a large number of joints and muscles, and there are a large number of coupled motion relationships among them. When processing human motion, we do not need to consider every joint or muscle movement, but need to combine and simplify it. Nowadays, image algorithms can accurately identify and subdivide human structure into multiple parts. In order to simplify the calculation, the human body is divided into head, trunk, large arm, large leg, small arm (including hand) and small leg (including foot) six parts, as shown in Figure 1 The relationship of each limb of human body can be regarded as a radial structure which decreases step by step from the trunk outward. If the trunk is regarded as the first level limb, the head, the large arm and the large leg connected with the first level limb are the second level limbs. Similarly, the small arm and the small leg are the third level limbs.

[0044] In the present application, the center of gravity of human body is taken as 56% of the height. As shown in Figure 2 S is the world system. Taking the center of gravity of human body as the origin, the front direction of the body is the xt axis to construct the first level limb coordinate system L t . Taking each joint as the origin, the second level limb coordinate systems L s1 , L s2 , L n , L h1 , L h2 and the third level limb coordinate systems L e1 , L e2 , L k1 , L k2 are constructed. The above coordinate systems are right-handed systems. In order to unify the representation method, the x axis direction of the second level limb system and the third level limb system is along the central axis of the limb to the joint of the next level.

[0045] The motion of each level limb includes consistent translation motion and rotation motion with the connection point of the upper level as the center. The translation motion of human body can be calculated by the motion trajectory of the center of gravity. The rotation motion of the limb can be regarded as the rotation of a rigid body, and the commonly used calculation methods include Euler method and quaternion method. Since the Euler method has multiple solutions when calculating the coupled motion of multiple rigid bodies, the quaternion method is adopted to represent the rotation motion of the limb.

[0046] In order to facilitate calculation, the present application simplifies each part of the human body into a simple geometric form. Among them, the trunk is simplified as a cuboid, and the remaining limbs are simplified as a cylinder. In the quaternion operation, the axial vector of the limb is extracted to represent its motion. As shown in Figure 3 The position of the limb at the initial moment is represented by the axial vector . After a sampling time, the limb moves to the position of . Among them, the translation motion is represented by , the rotation motion is represented by , and the rotation axis is . This is the result of rotation only.

[0047] Pick and The unit vector of the angle bisector is ,but

[0048] (1)

[0049] Where s n1 With s n2 Let s1 and s2 represent the unit vectors, respectively. Then the rotation quaternions of s1 and s2 are:

[0050] (2)

[0051] This allows us to obtain the rotation quaternions of each limb at each sampling moment. The motion of the human body in a microgravity environment is then expanded according to the limb contours, such as... Figure 4 As shown. Figure 4 The colors in the image are used to distinguish the probability of different areas of the human body occupying space at different times in the future. Darker colors represent the current time, and the change from red to green indicates that the time is from near to far. The probability also decreases as time increases.

[0052] Considering the robot's own moving volume, to avoid collisions with the human body, the predicted results are expanded using the same method for generating electric potential fields to obtain the dynamic collision avoidance potential field for the human body. Assume... The robot's current location is point R, and the coordinates of R in the world system are... .right Time to The space occupied by the human body at any given time is calculated, resulting in a form as follows: Figure 4 The point set in the middle is denoted as Г. The coordinates of any point W in the point set in the world system are... Let the collision avoidance repulsion vector be... Then the robot experiences a collision avoidance repulsive force potential function W at point R. for

[0053] (3)

[0054] in, This represents the Euclidean distance between points W and R. This represents the range of the repulsive potential at point W. Represents the collision avoidance repulsion coefficient, which is related to Related, the expression is

[0055] (4)

[0056] in, represents the repulsive force gain of collision avoidance and , the sampling time at which the representative point W is located, is a constant and The corresponding repulsive force of collision avoidance can be expressed as

[0057] (5)

[0058] Equation (4) shows that the farther the sampling time at which the representative point W is located is from , the smaller the Equation (5) shows that when calculating the size of the repulsive force of collision avoidance that a point R in the calculation space is subjected to, the center of the sphere is R, , the diameter of the sphere, only the points in contained in the sphere will have a repulsive effect on R, and its expression is

[0059] (6)

[0060] Figure 5 The generation effect of the dynamic collision avoidance potential field of the human body's near-body space with a collision avoidance radius of 20 cm in a microgravity environment is shown. It can be seen that the collision avoidance potential field completely wraps the occupied space along its contour, which can effectively avoid collision between the robot and the human body during movement.

[0061] Step two: human body near-body space subjective comfort potential field.

[0062] The movement of the human body is controlled by the consciousness of the human body. When a person is moving, he will perceive the changes in the surrounding environment in real time and react. When the cabin flying robot enters the human body's perception area, factors such as the size of the speed, the direction of movement, and the distance from the human body will all affect the human's perception and judgment. Therefore, we construct the human body's near-body space subjective potential field.

[0063] The human body mainly perceives moving objects in the surrounding environment through vision and hearing. The human visual perception range is approximately conical, and when the rotation of the head is considered, the combined perception range will be even wider. In this study, for the sake of simplicity, we assume that based on the compensation of auditory perception to visual perception, the human body's near-body space perception range is approximately spherical, with the center of the sphere coinciding with the center of the human head.

[0064] Since a person will feel danger and panic when an object is close to himself, and feel safe and relaxed when an object is far away from himself. Therefore, based on the assumed spherical subjective perception range with the human head as the center, we establish the human body's near-body space subjective comfort potential field. The effect of the subjective comfort potential field is similar to that of the collision avoidance potential field. The robot in the potential field will be subjected to a subjective repulsive force, so that the planned route will avoid the human head as much as possible.

[0065] Figure 6The basic principle of the subjective potential field of the human body's near space. The center of the human head at the current time is point O, and the position of the robot is point R in front of the human body. The predicted trajectory of the human head center in the future is curve , which will pass through points , , , in turn. The robot at point R is subjected to the subjective repulsive potential function

[0066] (7)

[0067] where represents the subjective repulsive constant. represents the distance vector of R and , and represents the Euclidean distance between point and point R. represents the subjective perception range, i.e., the repulsive potential range of point . The corresponding subjective repulsive force can be represented as

[0068] (8)

[0069] Similarly, the subjective repulsive force of each point in the predicted trajectory on the robot can be calculated … , and the weighted sum of the above is , whose expression is

[0070] (9)

[0071] where represents the position of the human head center in the predicted trajectory, represents the weighting coefficient, whose expression is

[0072] (10)

[0073] where represents the weighted gain constant, and have the same meaning as formula (4), is a constant and .

[0074] Formulas (8) and (10) show that the influence of the subjective potential field on the robot is not only related to the current head position, but also related to the future head position. Under the comprehensive influence of the current position and the future position, the robot will try to avoid the movement trajectory of the human head as much as possible to ensure the comfort of the human body during the movement. ​​

[0075] The effect of the human subjective potential field is shown in Fig. 2. The arrows radiating from the center of the human head represent the repulsive effect of the potential field on the robot. The closer to the center, the stronger the repulsive effect. Figure 7

[0076] Step three, fusion of the human body's near-body space potential field in microgravity environment and trajectory planning of the robot.

[0077] Combining formula (6) and (9), the collision avoidance potential field and the subjective comfort potential field in the human body's near-body space are synthesized. The robot is equivalent to a particle moving in the potential field, and the resultant force acting on the robot is

[0078] (11)

[0079] wherein represents the attractive force of the target point on the robot. It is used to guide the robot to the target position. In the traditional artificial potential field method (APF), there are various expressions of attractive force. The most commonly used is the combination of parabolic function and conical function. Given the target position , the target radius is defined as When the Euclidean distance between the robot and the target position is greater than , the conical function is used, and its expression is

[0080] (12)

[0081] wherein is the attractive force coefficient, is the Euclidean distance between the position of the robot and the global state . When the Euclidean distance between the robot and the target position is less than , the parabolic function is used, and its expression is

[0082] (13)

[0083] The conical function and the parabolic function are equal outside the circle with the target position as the center and the target radius as the radius. From this, it can be seen that the expression of the attractive force is a piecewise function. When the robot is too far from the target, the conical function is used, which reduces the value of the attractive force , which can effectively avoid the problem of excessive attractive force.

[0084] The iterative calculation formula of the position at adjacent time in path planning is

[0085] (14)

[0086] wherein with are the robot positions at adjacent time instants, is the time step of the iterative computation. The value of determines the planning time and the planning accuracy. The smaller the value of is, the higher the planning accuracy is, but the computation amount will increase and the planning time will increase accordingly.

[0087] Figure 8 The path planning effect of the human body's near-body space in a microgravity environment is shown. The robot successfully reaches the target point after passing through the human body's near-body space. As can be seen from the figure, under the combined action of the collision avoidance potential field and the subjective comfort potential field, the robot effectively avoids the human's subjective perception area and the space that the limbs may occupy in the future, ensuring the safety and comfort during the movement process.

[0088] Embodiment:

[0089] Fig. 9 is a comparison of the planning effect of the near-body space potential field and the artificial potential field in a microgravity environment. (a) is the planning effect of the near-body space potential field, and (b) is the planning effect of the artificial potential field. Path planning is performed from nine directions around the human body. In the two figures, the red number represents the starting point, and the green number represents the end point. The starting point and the end point with the same number represent the same planning. The invention selects nine random directions with the same distance around the human body as the starting point and selects the positions with the same distance on the opposite side as the corresponding end point. The line connecting each starting point and end point passes through the center position of the human body's chest. As can be seen from the figure, under the influence of the near-body space potential field, the planning route of Fig. 9 (a) obviously avoids the area with high occupancy probability around the human body. In comparison, the route planned by the traditional artificial potential field method without predicting the future of the human body is avoided only when it approaches the human body. This experiment can qualitatively illustrate the planning effect of the human body's near-body space potential field. Under the influence of the potential field, the robot will avoid the space that will be occupied by the human body from a long distance, thereby improving the safety of path planning.

[0090] The method described in the invention is verified by simulation experiments, and the technical effects claimed by the invention are verified. The method proposed in the invention solves the technical problems proposed in the invention, and the method is verified by actual application, and the technical effects and practicality claimed by the invention are verified.

[0091] It should be understood that the various forms of flow shown above can be reordered, added to, or deleted from. For example, the steps described in the present application can be executed in parallel, in sequence, or in a different order, as long as the desired results of the technical solutions disclosed in the present application can be achieved, which are within the protection scope of the present application.

Claims

1. A microgravity close-range space planning method for service robots inside a space station, characterized in that, The method is used for path planning in the near-human space under microgravity conditions, enabling the robot to smoothly reach the target point after passing through the near-human space from the starting point. It effectively avoids the area of ​​human perception and the space that the robot's limbs may occupy in the future. It includes: Step 1: Kinematic modeling of human joints under microgravity environment, and dynamic collision avoidance potential field of human near space generated based on the modeling results; Step 2: Based on people's subjective comfort evaluation within the perception range, design the subjective comfort potential field of the human body's close space; Step 3: Combining the dynamic collision avoidance potential field and the subjective comfort potential field of the human body's near-body space, complete the robot's path planning within the human body's near-body space in a microgravity environment.

2. The microgravity close-range space planning method for a space station in-cabin service robot according to claim 1, characterized in that, Step 1: Kinematic modeling of human joints under microgravity environment, and generation of dynamic collision avoidance potential field in the near-body space based on the modeling results, specifically: The human body structure is identified and subdivided into multiple parts using image algorithms. The relationship between the limbs can be seen as a radial structure that decreases in size from the torso outwards. The torso is considered as the first-level limb, and the head, upper arm, and thigh connected to the first-level limb are the second-level limbs; similarly, the forearm and lower leg are the third-level limbs. A coordinate system for the first-level limbs is constructed with the center of gravity of the human body as the origin, and coordinate systems for the second-level and third-level limbs are constructed with each joint as the origin. All of the above coordinate systems are right-handed. The movement of each level of limb includes consistent translational movement and rotational movement centered on the connection point with the previous level. The translational movement of the human body is calculated through the trajectory of the center of gravity, and the rotational movement of the limbs can be regarded as the rotation of a rigid body. Based on the above, the kinematic modeling of human joints in a microgravity environment is performed. The motion of the human body in a microgravity environment is expanded according to the limb contour. Considering the motion volume of the service robot in the space station cabin, in order to avoid collisions between the robot and the human body, the above prediction results are expanded based on the electric potential field generation method to obtain the human body dynamic collision avoidance potential field. Step 2: Based on a person's subjective comfort evaluation within their perception range, design the subjective comfort potential field of the human body's near-person space, specifically as follows: Hypothesis: Based on the compensation of auditory perception for visual perception, the perceptual range of the human body in close space is approximately a sphere, with the center of the sphere coinciding with the center of the human head; Based on the above assumption, a spherical subjective perception range centered on the human head is established to create a subjective comfort potential field in the human body's near space. The subjective comfort potential field is used to ensure that the robot will experience a subjective repulsive force when it is in the potential field, so that the planned route avoids the human head. Step 3: Combining the dynamic collision avoidance potential field and the subjective comfort potential field of the human body's near-human space, complete the robot's path planning within the human body's near-human space in a microgravity environment, specifically: The collision avoidance potential field in the near-human space is synthesized with the subjective comfort potential field. The robot is equivalent to a point mass moving in the potential field, and the net force on the robot is obtained. When the robot is too far from the target, a cone function is used to reduce the value of gravity to effectively avoid excessive gravity. When the robot is close to the target, a parabolic function is used to ensure that the robot can reach the target point smoothly. Finally, the planned path is obtained by iterating the position of adjacent time steps.

3. The microgravity close-range space planning method for a space station in-cabin service robot according to claim 2, characterized in that, The method for determining whether the robot is too far or too close to the target is as follows: Define a global distance Rg; distances greater than Rg are calculated using a cone function, and distances less than or equal to Rg are calculated using a parabola; or, The target distance should be between 1m and 1.5m, or adjusted within this range based on the actual effect.

4. A microgravity close-range space planning method for a space station in-cabin service robot according to claim 2, characterized in that, Step one involves the kinematic modeling of human joints and the construction of the near-field collision avoidance potential field in a microgravity environment: When modeling human movement, we first combine and simplify it, and then use image algorithms to identify and subdivide the human structure into multiple parts: the human body is divided into six parts: head, trunk, upper arm, thigh, forearm (including hand), and lower leg (including foot). The relationship between the limbs of the human body can be regarded as a radial structure that decreases in level from the trunk outward. If the trunk is regarded as the first-level limb, then the head, upper arm, and thigh connected to the first-level limb are the second-level limbs; similarly, the forearm and lower leg are the third-level limbs. Determine the position of the human body's center of gravity, with S as the world frame, and construct a first-level limb coordinate system L with the human body's center of gravity as the origin and the direction in front of the body as the xt axis. t Construct a secondary limb coordinate system L with each joint as the origin. s1 L s2 L n L h1 L h2 and the three-level limb coordinate system L e1 L e2 L k1 L k2 The coordinate systems mentioned above are all right-handed. The x-axis of both the secondary and tertiary limb systems points along the central axis of the limb towards the next level joint. The movements of limbs at all levels include consistent translational movements and rotational movements centered on the connection point with the previous level. The translational movements of the human body can be calculated from the trajectory of the center of gravity. The rotational movements of the limbs can be regarded as the rotation of rigid bodies. The commonly used calculation methods are Euler's method and quaternion method. Since Euler's method has multiple solutions when calculating the coupled movements of multi-level rigid bodies, quaternion method is used to represent the rotational movements of limbs. The human body is simplified into simple geometric shapes: the torso is simplified to a cuboid, and the remaining limbs are simplified to cylinders; in quaternion operations, the axis vectors of the limbs are extracted to represent their motion; the initial position of the limbs is represented by the axis vectors. This indicates that after one sampling time, the limb movement reached... Position; where translational motion is used Indicates rotational motion This indicates that the axis of rotation is , This is the result of rotation only; Pick and The unit vector of the angle bisector is ,but （1）； Where s n1 With s n2 Let s1 and s2 represent the unit vectors respectively. Then the rotation quaternions of s1 and s2 are: （2）； This allows us to obtain the rotation quaternion of each limb at each sampling time; the motion of the human body in the microgravity environment is expanded according to the limb contour, and the spatial occupancy probability of different areas of the human body at future times is given. Considering the robot's own moving volume, to avoid collisions with the human body, the above prediction results are expanded using the method for generating the electric potential field to obtain the dynamic collision avoidance potential field for the human body; assuming... The robot's current location is point R, and the coordinates of R in the world system are... ;right Time to The space occupied by the human body at any given time is calculated to obtain a point set, denoted as Г. The coordinates of any point W in the point set in the world system are: Let the collision avoidance repulsion force vector be... Then the robot experiences a collision avoidance repulsive force potential function W at point R. for （3）； in, This represents the Euclidean distance between points W and R. This represents the range of the repulsive potential at point W. Represents the collision avoidance repulsion coefficient, which is related to Related, the expression is （4）； in, Represents the increase in collision avoidance repulsion force and , The sampling time at point W represents the location of the sampling point. is a constant and The corresponding collision avoidance repulsion force can be expressed as: （5）； Formula (4) represents the distance at which point W is located during the sampling time. The farther away, the more The smaller the value, the more appropriate the formula (5) is. Formula (5) indicates that when calculating the magnitude of the collision avoidance repulsion force on a point R in the calculation space, with R as the center of the sphere, The diameter of the sphere is only contained within the sphere. Points in R will have a repulsive effect on R, and its expression is: （6）； The collision avoidance potential field can completely wrap around the robot along the contour of the occupied space, which can effectively prevent the robot from colliding with the human body when it moves.

5. A microgravity close-range space planning method for a space station in-cabin service robot according to claim 4, characterized in that, Step two, the process of constructing the potential field for subjective comfort in the near-human space is as follows: When the in-cabin flying robot enters the human's perception zone, its speed, direction of movement, and distance from the human will all affect the human's perception and judgment. Based on this, a subjective potential field in the near-human space is constructed. Hypothesis: Based on the compensation of auditory perception for visual perception, the perceptual range of the human body in close space is approximately a sphere, with the center of the sphere coinciding with the center of the human head; Based on the above assumptions, a spherical subjective perception range centered on the human head is established to create a subjective comfort potential field in the near-human space. The effect of the subjective comfort potential field is similar to that of the collision avoidance potential field. When the robot is in the potential field, it will be subject to subjective repulsive force, causing the planned route to avoid the human head. The center of the human head at the current moment is The robot's current position is point R in front of the human body; the predicted trajectory of the center of the human head at future moments is a curve. It will pass through points in sequence. , , The robot is affected at point R. Subjective repulsive potential function for （7）； in Represents the subjective repulsive force constant. Indicates R and The distance vector, and represent The Euclidean distance between point R and point R; Represents the scope of subjective perception, that is The range of the repulsive potential at a point, and the corresponding subjective repulsive force can be expressed as: （8）； Similarly, the predicted trajectory can be calculated. Subjective repulsive force generated by each point on the robot ... Sum them up by weights to get Its expression is （9）； in, This indicates the position of the center of the human head in the predicted trajectory. The weighting coefficient is expressed as follows: （10）； in Represents the weighted gain constant. and The representation is the same as in formula (4). is a constant and ; Formulas (8) and (10) indicate that the influence of the subjective potential field on the robot is not only related to the current head position, but also to the future head position. Under the combined influence of the current position and the future position, the robot will try its best to avoid the movement trajectory of the human head and ensure the comfort of the human body during the movement.

6. A microgravity close-range space planning method for a space station in-cabin service robot according to claim 5, characterized in that, In step three, the specific process of fusing the potential field of the human body in the microgravity environment and planning the robot's trajectory is as follows: Combining formulas (6) and (9), the collision avoidance potential field and the subjective comfort potential field in the near-human space are synthesized. The robot is equivalent to a point mass moving in the potential field. Then the net force on the robot is: （11）； in, This represents the gravitational force exerted by the target point on the robot. This method guides a robot to a target location, using a combination of parabolic and conical functions to represent gravity in the Artificial Potential Field (APF) method. The target location is known to be... Define the target radius When the Euclidean distance between the robot and the target position is greater than When using a cone function, its expression is: （12）； in The gravitational coefficient, The robot's position and global state The Euclidean distance between the robot and the target position; when the Euclidean distance between the robot and the target position is less than or equal to... When using a parabolic function, its expression is: （13）； Conical functions and parabolic functions at the target position The center is the target radius. The outer circumferences of the circles are equal; from this, we can see that gravity... The expression is a piecewise function. When the robot is too far from the target, a cone function is used to reduce the gravitational pull. The value of should be chosen to avoid excessive gravitational force; The formula for calculating the position iteration of adjacent time steps in path planning is as follows: （14）； in and These represent the robot positions at adjacent time points. This refers to the time step size for iterative calculations. The value of will determine the planning time and planning accuracy; The smaller the value of , the higher the planning accuracy, but the amount of computation will increase, and the planning time will increase accordingly.

7. A microgravity close-range space planning method for a space station in-cabin service robot according to claim 4, characterized in that, The center of gravity of a person is set at 56% of their height.

8. A recommendation system for ice and snow tourism products based on neural collaborative filtering and linear confidence upper bound, characterized in that: The system has a program module corresponding to the steps of any one of claims 1-7, and executes the steps in the microgravity close-range space planning method for a space station in-cabin service robot when it is run.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program configured to, when invoked by a processor, implement the steps of the microgravity close-range space planning method for a space station in-cabin service robot according to any one of claims 1-7.

Images