An intelligent nursing robot for intensive care units

Abstract

The present invention relates to an intelligent nursing robot for an intensive care unit, comprising a bed frame, a mattress and a bracket arranged at the foot of the bed frame, a host is arranged on the bracket, a slide rail is arranged above the bed frame, an intelligent mechanical arm is suspended on the slide rail, and the intelligent mechanical arm can move freely along the slide rail; it also comprises cameras located on both sides above the bed body; the intelligent mechanical arm and the camera are both communicatively connected to the host; the host comprises a display screen and an input device arranged outside the display screen, and the input device comprises handwriting input, audio input, and video input functions. The intelligent nursing robot can realize all-weather and multi-type monitoring and nursing of patients.

Description

Technical Field

[0001] The invention relates to a nursing robot, in particular to an intelligent nursing robot used in an intensive care unit. Background Art

[0002] Nursing robots have been a hot research topic in recent years, and many nursing robots are widely used in the fields of medical care and elderly care. Current nursing robots usually have conventional functions such as walking, answering questions, and routine measurements, but cannot achieve more complex functions.

[0003] Intensive Care Units, also known as ICUs, are used for critically ill patients. There are many devices in the ICU, and the patients are in serious condition, requiring all-round, all-weather monitoring and care to prevent various accidents. When accidents occur, doctors and nurses often fail to detect them in time.

[0004] In addition, when abnormal conditions occur in the data monitored by various instruments, they cannot be quickly locked down.

[0005] Therefore, there is a need for an intelligent nursing robot that can monitor and care for patients in various types around the clock. Summary of the invention

[0006] The purpose of the present invention is to provide an intelligent nursing robot for an intensive care unit, which can monitor the conditions of patients or instruments in the intensive care unit 24 hours a day, and can promptly discover hidden situations and deal with them or notify medical staff.

[0007] The technical solution adopted by the present invention is: an intelligent nursing robot for an intensive care unit, comprising a bed frame, a mattress and a bracket arranged at the end of the bed frame, a mainframe being arranged on the bracket, a slide rail being arranged above the bed frame, an intelligent mechanical arm being suspended on the slide rail, and the intelligent mechanical arm being able to move freely along the slide rail; further comprising cameras located on both sides above the bed body; the intelligent mechanical arm and the camera are both communicatively connected to the mainframe; the mainframe comprises a display screen and an input device arranged outside the display screen, and the input device comprises handwriting input, audio input, and video input functions.

[0008] Furthermore, the host also includes a camera, which is used to monitor the patient and the use of instruments around the patient.

[0009] Furthermore, the track includes a middle track located in the middle position above the bed frame and extending along the length direction of the bed body, the middle track is connected to the right upper track located at the head of the bed above the bed frame and extending along the width direction of the bed body through an arc track, the right upper track is connected to the right track located at the right position above the bed frame and extending along the length direction of the bed body through an arc track, the right track is connected to the lower track located at the foot of the bed above the bed frame and extending along the width direction of the bed body through an arc track, the lower track is connected to the left track located at the left position above the bed frame and extending along the length direction of the bed body through an arc track, the left track is connected to the left upper track located at the head of the bed above the bed frame and extending along the width direction of the bed body; the upper right track is not connected to the middle track

[0010] Furthermore, it also includes an intelligent care module arranged in the host, which includes a semantic processing module. The semantic processing module is used to accept inputs from the audio input module, the video input module, and the handwriting input module, and extract the input content based on a large language model and convert it into machine-recognizable commands.

[0011] Furthermore, the semantic processing module understands the user's input through the large language model, organizes it into an event set, and stores it in a register; the event set includes at least four events: time, target, object, and action.

[0012] Furthermore, it also includes a machine vision module, which is used to identify targets and objects, and accurately identifies the targets and objects therein based on the photos or videos taken by the camera. The machine vision module compares the identified targets and / or objects with the predetermined targets and / or objects to determine whether the targets and / or objects are abnormal; in one embodiment, the machine vision module compares the range of the target with the spatial range of the object to determine whether the ranges of the two are within the predetermined overlapping range, if so, there is no abnormality; if not, the intelligent mechanical arm is activated to move the target and / or object so that the range between the two overlaps within the predetermined range.

[0013] It also includes a calculation module, which is used to calculate the coordinate information of the target and the object; the calculation module calculates the distance between the target and the object by using a camera, and calculates the coordinates of the target and the object based on the distance; the coordinates of the target and the object are stored in a storage module, and the storage module also stores time events and action events, and inputs the data to the intelligent robotic arm through an output module; when the time event is triggered, the intelligent robotic arm first drives to a position close to the target, starts the robotic arm to grab the target, and after grabbing, sends the target to the object, releases and completes the action after the object picks up the object or reaches the object position, and returns to the standby position.

[0014] Furthermore, the two cameras located above the bed frame are both binocular cameras, thereby forming a binocular camera combination, and the binocular ranging technology is used to obtain the distance between the target or object and a single binocular camera; the binocular camera pre-identifies the target or object based on the target and object information sent by the machine vision module, and adjusts the rotation of the camera, focuses the camera on the target or object, records the rotation angle of the camera to obtain the pitch angle of the camera, and calculates the precise coordinates of the target and object based on the distance between the target or object and the single binocular camera and the pitch angles of the target or object and the single binocular camera respectively through the trigonometric function relationship.

[0015] The precise coordinates of the target and the object are stored in the storage module, which also stores time events and action events. The above data are input to the intelligent robotic arm 6 through the output module. When the time event is triggered, the intelligent robotic arm 6 first drives to a position close to the target, starts the robotic arm to grab the target, and after grabbing, sends the target to the object. When taking the object, it releases and completes the action, and returns to the standby position.

[0016] Further, the following steps are adopted to calculate the three-dimensional coordinates of any target or object: establish a three-dimensional coordinate system O-XYZ in the indoor space of the intensive care unit, and take the ground as the bottom surface of the three-dimensional coordinate system. Then, the two binocular cameras in this application are simplified to space points A and B, and the projection points of points A and B on the bottom surface of the three-dimensional coordinate system are D and E respectively; connect AD and BE; any target or object is simplified to space point C, and its projection point on the bottom surface of the three-dimensional coordinate system is K; connect CK; connect AC and BC, and make a straight line CF perpendicular to AC, and a straight line CG perpendicular to BE; then the pitch angle of camera A is ∠CAD, and the pitch angle of camera B is ∠CBE;

[0017] The height of target C from the ground CK = AD-AF or CK = BE-BG; AD is the height of camera A from the ground, which is a known value; BE is the height of camera B from the ground, which is also a known value; AF = AC*cos∠CAD or BG = BC*cos∠CBE; AC and BC values ​​are the distances between the binocular camera and the target, which are also known values.

[0018] CF=AC*sin∠CAD or CG=BC*sin∠CBE; the projection line of CF on the ground is KD, CF=KD; the projection line of CG on the ground is KE, CG=KE; DE is the line connecting the projections of the camera on the ground; a triangle KDE consisting of the projection points of the two cameras A, B and the target C on the ground is constructed; DE is the vertical distance between the two cameras, which is pre-stored in the calculation module, then KD, KE, and DE are all calculated; then ∠KDE can be calculated by trigonometric functions.

[0019] cos∠KDE=(DE2+DK2-KE2) / (2DE*DK);

[0020] Make projection points H and J of the projection point K of target C on the coordinate axis and calculate KH and KJ;

[0021] KH=KD*SIN∠KDE; then the X-axis coordinate of target C in the three-dimensional coordinate system can be obtained OJ=KH;

[0022] DH=KD*cos∠KDE; EH=DH-DE, DE is known, that is, the distance between the two cameras; then OH=OE-EH, where OE is a known value, that is, the Y-axis coordinate value of camera B;

[0023] In this way, the three-dimensional coordinate values ​​of the target C have been obtained; X=OJ, Y=OH, Z=CK.

[0024] Furthermore, the camera also provided on the host can monitor the display screen of the instrument used by the patient in real time, record the numbers or graphics on the display screen, and extract predetermined numbers or graphics based on the machine vision module.

[0025] Furthermore, the binocular cameras can be set to multiple, and the multiple binocular cameras are used to form a binocular camera combination in pairs to calculate the coordinate values ​​of multiple targets or objects, and these target values ​​are corrected to obtain accurate coordinate values.

[0026] The present invention adopts a track-type robot, which can make full use of the space in the intensive care unit without blocking the movement of medical staff, patients, and medical equipment. The use of a track with a specific configuration can enable the robot to move in a wider range and can achieve all movements within a 360-degree range around the bed. The machine vision module can monitor the status of a specific target after identifying the specific target. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a structural schematic diagram of an intelligent nursing robot for an intensive care unit according to the present invention;

[0028] Figure 2 This is a schematic diagram of the host structure of an intelligent nursing robot for an intensive care unit according to the present invention;

[0029] Figure 3 It is a track schematic diagram of an intelligent nursing robot for an intensive care unit according to the present invention;

[0030] Figure 4 It is a system schematic diagram of an intelligent nursing module of an intelligent nursing robot for an intensive care unit according to the present invention;

[0031] Figure 5 It is a schematic diagram of spatial geometry calculation of target coordinates of an intelligent nursing robot for an intensive care unit described in the present invention. DETAILED DESCRIPTION

[0032] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0033] like Figure 1 The intelligent nursing robot for the intensive care unit of the present invention comprises a bed frame 1, a mattress 2 and a bracket 3 arranged at the end of the bed frame 1, a host 4 is arranged on the bracket 3, and a slide rail 7 is arranged above the bed frame 1, and an intelligent mechanical arm 6 is suspended on the slide rail 7, and the intelligent mechanical arm 6 can move freely along the slide rail 7. It also includes cameras 5 located on both sides above the bed.

[0034] refer to Figure 2 The host 4 of the present invention includes a display screen and an input device 5 arranged outside the display screen, and the input device 5 includes handwriting input, audio input, and video input functions. The host also includes a camera, which is used to monitor the patient and the use of the equipment beside the patient.

[0035] The intelligent robotic arm 6 and the camera 5 are both communicatively connected to the host 4, and the connection method can be various methods such as a wifi router, near field communication, and remote server communication.

[0036] refer to Figure 3 , which shows the structure of the track 7 of the present application, wherein the track 7 includes a middle track 71 located in the middle position above the bed frame 1 and extending along the length direction of the bed body, the middle track 71 is connected to an upper right track 72 located at the head of the bed above the bed frame 1 and extending along the width direction of the bed body through an arc track, the upper right track 72 is connected to a right track 73 located at the right position above the bed frame 1 and extending along the length direction of the bed body through an arc track, the right track 73 is connected to a lower track 74 located at the foot of the bed above the bed frame 1 and extending along the width direction of the bed body through an arc track, the lower track 74 is connected to a left track 75 located at the left position above the bed frame 1 and extending along the length direction of the bed body through an arc track 77, the left track 75 is connected to an upper left track 76 located at the head of the bed above the bed frame 1 and extending along the width direction of the bed body, thus constructing a non-annular track system that can take into account the bed body and both sides of the bed body.

[0037] refer to Figure 4The intelligent nursing robot for the intensive care unit of the present invention further includes an intelligent nursing module 41, which includes a semantic processing module 41, and the semantic processing module 41 is used to accept the input of the audio input module 47, the video input module 48, and the handwriting input module 49, and extract the input content based on the large language model and convert it into a machine-recognizable command. The large language model can be, for example, chatgdp, Wenxinyiyan, Huawei Pangu and other large language models. After calling such large models, the machine can accurately understand human language.

[0038] The semantic processing module understands the user's input through the large language model, organizes it into an event set, and stores it in the register. The event set includes at least four events: time, target, object, and action. For example, when the voice input is: give me aspirin at 2 o'clock. The semantic processing module extracts it into time: 2 o'clock; target: aspirin; object: the person being cared for; action: deliver aspirin to the patient. In this way, the task for the intelligent robotic arm 6 can be obtained.

[0039] It also includes a machine vision module, which is used to identify targets and objects, and accurately identifies targets and objects based on the photos or videos taken by the camera 5. The machine vision module is based on a SAM model (Segment Anything Model), which can find and segment any object in an image or video. The machine vision module identifies and marks the targets and objects stored in the register in the semantic processing module.

[0040] The system further includes a calculation module, which is used to calculate the coordinate information of the target and the object. The calculation module calculates the distance between the target and the object by using the camera 5, and calculates the coordinates of the target and the object based on the distance. Figure 1 In the example shown in FIG. 5 , two binocular cameras 5 are shown, thereby forming a binocular camera combination. The distance between the target and the object and the camera can be easily obtained by using the binocular ranging technology, and then the precise coordinates of the target and the object can be calculated through the trigonometric function relationship.

[0041] The precise coordinates of the target and the object are stored in the storage module, which also stores time events and action events. The above data are input to the intelligent robotic arm 6 through the output module. When the time event is triggered, the intelligent robotic arm 6 first drives to a position close to the target, starts the robotic arm to grab the target, and after grabbing, sends the target to the object. When taking the object, it releases and completes the action, and returns to the standby position.

[0042] like Figure 4As shown, the present application now describes in detail how to calculate the three-dimensional coordinates of any target or object. A three-dimensional coordinate system O-XYZ is established in the indoor space of the intensive care unit, with the ground as the bottom surface of the three-dimensional coordinate system. The two binocular cameras 5 in the present application are simplified to spatial points A and B. The projection points of points A and B on the bottom surface of the three-dimensional coordinate system are D and E respectively; connect AD and BE; any target or object is simplified to spatial point C, and its projection point on the bottom surface of the three-dimensional coordinate system is K; connect CK. After using the machine vision module to identify the target C, adjust the direction of the binocular camera so that it focuses on the target C, connect AC and BC; and make a straight line CF perpendicular to AC, and a straight line CG perpendicular to BE; then the pitch angle ∠CAD of camera A and the pitch angle ∠CBE of camera B can be obtained; the pitch angle values ​​of cameras A and B are obtained by recording the rotation angles of the cameras. In this way, the present invention recognizes the target in advance and focuses the camera on the target, so the pitch angle of the camera can be obtained by adjusting the rotation angle of the camera, without complicated calculation, and only needs to record the pitch angle of the camera rotation. The pitch angle value is transmitted to the calculation module of the host 4.

[0043] In the computing module of the host 4, the coordinates of the cameras A and B are pre-recorded, and the coordinates of the target C are calculated based on the coordinates.

[0044] Then the height of target C from the ground is CK = AD-AF or CK = BE-BG; AD is the height of camera A from the ground, which is a known value; BE is the height of camera B from the ground, which is also a known value; AF = AC*cos∠CAD or BG = BC*cos∠CBE. The AC and BC values ​​are the distances between the binocular camera and the target, which are also known values. In this way, the height of target C from the ground can be obtained.

[0045] CF=AC*sin∠CAD or CG=BC*sin∠CBE; the projection line of CF on the ground is KD, CF=KD; the projection line of CG on the ground is KE, CG=KE; DE is the line connecting the projections of the cameras on the ground; a triangle KDE consisting of the projection points of the two cameras A, B and the target C on the ground is constructed; DE is the vertical distance between the two cameras, which is pre-stored in the calculation module, then KD, KE, and DE are all calculated; then ∠KDE can be calculated by trigonometric functions:

[0046] cos∠KDE=(DE2+DK2-KE2) / (2DE*DK);

[0047] Make projection points H and J of the projection point K of target C on the coordinate axis and calculate KH and KJ;

[0048] KH=KD*SIN∠KDE; then the X-axis coordinate of target C in the three-dimensional coordinate system can be obtained OJ=KH;

[0049] DH=KD*cos∠KDE; EH=DH-DE, DE is known, that is, the distance between the two cameras; then OH=OE-EH, where OE is a known value, that is, the Y-axis coordinate value of camera B;

[0050] In this way, the three-dimensional coordinate values ​​of the target C have been obtained; X=OJ, Y=OH, Z=CK.

[0051] The calculation module stores the calculated three-dimensional coordinate value of the target C in the storage module. Similarly, the three-dimensional coordinates of the object are calculated, and the coordinates of the target and the object are sent to the intelligent robot arm 6 through the output module 46. The intelligent robot arm 6 goes to the track closest to the coordinates according to the coordinate form, starts the gripper, grabs the target object, and then sends it to the object target. The binocular camera can be set to multiple, and the coordinate values ​​of multiple targets or objects are calculated by using a method of building a binocular camera combination between multiple binocular cameras, and these target values ​​are corrected to obtain accurate coordinate values. For example, multiple coordinate values ​​are corrected by weighting, interpolation, etc.

[0052] The present invention uses a machine vision model to quickly locate the target and object, and calculates the three-dimensional coordinates of the target and object through two binocular cameras, and sends the three-dimensional coordinates to the intelligent mechanical arm, which performs the action from the target to the object. The complex graphic calculation method is avoided, and the target positioning is obtained by the machine vision module, and the camera focus is adjusted to obtain the record of the pitch angle.

[0053] In one embodiment, the machine vision module identifies the target based on the photos and videos taken by the camera, and determines whether the target is in a predetermined position, or determines whether the coverage range of the target matches the predetermined range. If not, the target is moved to the predetermined position.

[0054] For example, when the patient's quilt slips off, the machine vision module recognizes that the overlap between the range of the quilt and the range of the mattress exceeds the predetermined range, and the host 4 sends an instruction to the intelligent robotic arm 6, and the intelligent robotic arm recognizes the feature point mark of the quilt (such as setting a label at the corner of the quilt, and the label is, for example, RFID), and moves the feature point mark to a predetermined position.

[0055] In one embodiment, the machine vision module is used to monitor whether the patient's breathing mask is in a position of slipping or falling. The machine vision module identifies the target (breathing mask) and the object (human face, especially the nose) from the photo or video taken by the camera, and determines whether the range occupied by the target covers the range occupied by the object. If not, it means that the breathing mask has slipped from the nose of the object, then the host 4 sends a command to the intelligent mechanical arm, and the intelligent mechanical arm will grab the mask and move it to a predetermined position. In summary, the machine vision module compares the identified target and / or object with the predetermined target and / or object to determine whether the target and / or object is abnormal; in one embodiment, the machine vision module compares the range of the target with the spatial range of the object to determine whether the ranges of the two are within the predetermined overlapping range. If yes, there is no abnormality; if not, the intelligent mechanical arm is started to move the target and / or object so that the range between the two overlaps within the predetermined range.

[0056] Patients in the intensive care unit are usually in a serious condition and often remove their tubes or oxygen masks by themselves, or are unconscious when this happens. If no one is monitoring them, a dangerous situation will occur. The present invention uses a machine vision module to identify a specific target and monitor the status of the specific target.

[0057] In one embodiment, the target is also provided with an electronic radio frequency tag, RFID tag. When the robotic arm approaches the target, the RFID tag on the target is first scanned by a scanning device provided on the intelligent robotic arm. If it matches the predetermined object, it is then grasped, thereby ensuring that the target is grasped more accurately.

[0058] The host 4 is also provided with a monitoring camera, which can monitor the display screen of the instrument used by the patient in real time, record the numbers or graphics on the display screen, and extract predetermined numbers or graphics. For example, the electrocardiogram photo is extracted using a machine vision model, and the photo is compared with a standard photo in a database. If an abnormality occurs, it is sent to a doctor or nurse. At present, the scope of intelligent medical care is limited to sending data such as blood pressure, blood oxygen, and heartbeat to doctors or nurses, but it is impossible to monitor the graphics in real time. The present invention can extract specific graphics from photos or images using a machine vision model, and compare them with standard graphics to monitor the patient's physical condition in real time, and when the condition is abnormal, the abnormal graphics are captured in time to assist doctors in diagnosis.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent nursing robot for an intensive care unit, characterized in that: The invention comprises a bed frame, a mattress and a bracket arranged at the foot of the bed frame, a mainframe is arranged on the bracket, a track is arranged above the bed frame, an intelligent mechanical arm is suspended on the track, and the intelligent mechanical arm can move freely along the track; it also comprises cameras located on both sides above the bed body; the intelligent mechanical arm and the camera are both communicatively connected to the mainframe; the mainframe comprises a display screen and an input device arranged outside the display screen, and the input device comprises handwriting input, audio input, and video input functions; It also includes a machine vision module, which is used to identify targets and objects, and accurately identify the targets and objects in the photos or videos taken by the camera; determine whether the targets and / or objects are within a predetermined range; The two cameras located above the bed frame are both binocular cameras, thus forming a binocular camera combination, and the binocular ranging technology is used to obtain the distance between the target or object and a single binocular camera; the binocular camera pre-identifies the target or object according to the target and object information sent by the machine vision module, and adjusts the rotation of the camera, focuses the camera on the target or object, records the rotation angle of the camera to obtain the pitch angle of the camera, and calculates the precise coordinates of the target and the object according to the distance between the target or object and the single binocular camera and the pitch angle between the target or object and the single binocular camera respectively through the trigonometric function relationship; The following steps are used to calculate the three-dimensional coordinates of any target or object: A three-dimensional coordinate system O-XYZ is established in the indoor space of the intensive care unit, and the ground is used as the bottom surface of the three-dimensional coordinate system. Then the two binocular cameras in this application are simplified to space points A and B, and the projection points of points A and B on the bottom surface of the three-dimensional coordinate system are D and E respectively; connect AD and BE; Any target or object is simplified to a spatial point C, whose projection point on the bottom of the three-dimensional coordinate system is K; connect CK; connect AC, BC, and make a straight line CF perpendicular to AC, and a straight line CG perpendicular to BE; then the pitch angle of camera A is ∠CAD, and the pitch angle of camera B is ∠CBE; The height of target C from the ground is CK=AD-AF or CK=BE-BG; AD is the height of camera A from the ground, which is a known value; BE is the height of camera B from the ground, which is also a known value; or ; AC and BC values ​​are the distances between the binocular camera and the target, which are also known values; or ; The projection line of CF on the ground is KD, CF=KD; The projection line of CG on the ground is KE, CG=KE; DE is the line connecting the projections of the cameras on the ground; A triangle KDE consisting of the projection points of the two cameras A, B and the target C on the ground is constructed; DE is the vertical distance between the two cameras, which is pre-stored in the calculation module, then KD, KE, and DE are all calculated; Then ∠KDE can be calculated by trigonometric functions; ； Make projection points H and J of the projection point K of target C on the coordinate axis and calculate KH and KJ; ; Then the X-axis coordinate of target C in the three-dimensional coordinate system can be obtained OJ=KH; ; EH = DH-DE, DE is known, that is, the distance between the two cameras; then OH = OE-EH, where OE is a known value, that is, the Y-axis coordinate value of camera B; In this way, the three-dimensional coordinate values ​​of the target C have been obtained; X=OJ, Y=OH, Z=CK.

2. The intelligent nursing robot for an intensive care unit according to claim 1, characterized in that: The host also includes a camera, which is used to monitor the patient and the use of the instruments around the patient.

3. The intelligent nursing robot for an intensive care unit according to claim 1, characterized in that: The track includes a middle track located in the middle position above the bed frame and extending along the length direction of the bed body; the middle track is connected to the right upper track located at the head position above the bed frame and extending along the width direction of the bed body through an arc track; the right upper track is connected to the right track located at the right position above the bed frame and extending along the length direction of the bed body through an arc track; the right track is connected to the lower track located at the foot position above the bed frame and extending along the width direction of the bed body through an arc track; the lower track is connected to the left track located at the left position above the bed frame and extending along the length direction of the bed body through an arc track; the left track is connected to the left upper track located at the head position above the bed frame and extending along the width direction of the bed body; the upper right track is not connected to the middle track.

4. The intelligent nursing robot for an intensive care unit according to claim 1, characterized in that: It also includes an intelligent care module set in the host, which includes a semantic processing module. The semantic processing module is used to accept inputs from the audio input module, the video input module, and the handwriting input module, and extract the input content based on a large language model and convert it into machine-recognizable commands.

5. The intelligent nursing robot for an intensive care unit according to claim 4, characterized in that: The semantic processing module understands the user's input through the large language model, organizes it into an event set, and stores it in a register; the event set includes at least four events: time, target, object, and action.

6. The intelligent nursing robot for an intensive care unit according to claim 5, characterized in that: It also includes a calculation module, which is used to calculate the coordinate information of the target and the object; the calculation module calculates the distance between the target and the object by using a camera, and calculates the coordinates of the target and the object based on the distance; the coordinates of the target and the object are stored in a storage module, and the storage module also stores time events and action events, and inputs the data to the intelligent robotic arm through an output module; when the time event is triggered, the intelligent robotic arm first drives to a position close to the target, starts the robotic arm to grab the target, and after grabbing, sends the target to the object, releases and completes the action after the object picks up the object or reaches the object position, and returns to the standby position.

7. An intelligent nursing robot for an intensive care unit according to any one of claims 1 to 6, characterized in that: The camera also arranged on the host can monitor the display screen of the instrument used by the patient in real time, record the numbers or graphics on the display screen, and extract the predetermined numbers or graphics based on the machine vision module.

8. The intelligent nursing robot for an intensive care unit according to claim 7, characterized in that: The binocular cameras are arranged in a plurality, and a binocular camera combination is constructed between the plurality of binocular cameras in pairs, so as to calculate the coordinate values ​​of a plurality of targets or objects, and these target values ​​are corrected to obtain accurate coordinate values.

Images