Machine Learning-Based Non-Contact Elevator Control System

Abstract

This invention discloses a machine learning-based non-contact elevator control system, relating to the field of elevator control. It includes: an external call panel, a first non-contact capacitive sensor, a second non-contact capacitive sensor, a third non-contact capacitive sensor, a data acquisition module, a data processing module, and an elevator control module. The external call panel is located outside the elevator. The first, second, and third non-contact capacitive sensors are all electrically connected to the data acquisition module. The data processing module is electrically connected to the data acquisition module. The elevator control module is connected to the data processing module. This invention uses a non-contact method to control the elevator, combining advantages such as low cost, high recognition accuracy, prevention of accidental touches, low susceptibility to environmental influences, and high convenience.

Description

Technical Field

[0001] This invention relates to the field of elevator control, and more specifically to a non-contact elevator control system based on machine learning. Background Technology

[0002] In related technologies, elevator control for upward or downward movement often employs methods such as contact button operation, vision-based gesture control, infrared-based gesture control, or radio frequency-based gesture control. However, contact button operation has the following drawbacks: First, mechanical buttons are prone to wear and tear after prolonged use, leading to sluggish response or malfunction. Second, in high-traffic public places, the surface and crevices of mechanical buttons easily accumulate dirt and bacteria, increasing the risk of cross-infection, and the dirt in the crevices is also difficult to clean. Vision-based gesture control, on the other hand, heavily relies on the performance and stability of external cameras and image processing algorithms, undoubtedly increasing system complexity. It is also highly susceptible to factors such as lighting, obstructions, and the complexity of hand movements, leading to decreased recognition accuracy. Furthermore, visual recognition requires capturing and analyzing user hand movements, which may raise concerns about user privacy and security. Gesture control based on infrared technology not only requires more complex hardware, leading to higher costs, but infrared light is also susceptible to interference from ambient light, and direct sunlight can cause misjudgments. Furthermore, because infrared technology cannot detect hands or other objects blocking infrared light, accidental touches are inevitable. Gesture control based on radio frequency (RF) technology requires users to carry RFID-enabled cards while riding elevators, which is inconvenient; moreover, contactless elevator buttons implemented with RF technology may experience interference in buildings with many metal structures.

[0003] Therefore, there is an urgent need for a non-contact elevator control system that is low in cost, has high recognition accuracy, can prevent accidental touches, is less affected by the environment, and is highly convenient. Summary of the Invention

[0004] In view of this, the present invention provides a non-contact elevator control system based on machine learning to solve the technical problems existing in related technologies.

[0005] This invention provides a machine learning-based non-contact elevator control system, characterized in that it includes:

[0006] An external call panel is located outside the elevator; one side of the external call panel is provided with a fingerprint recognition button, an up button, and a down button; the fingerprint recognition button, the up button, and the down button are all integrally formed with the external call panel.

[0007] A first non-contact capacitive sensor is located on the other side of the external call panel and corresponds to the position of the fingerprint recognition button, used to detect the user's intention to approach the fingerprint recognition button without physical contact.

[0008] A second non-contact capacitive sensor is located on the other side of the external call panel and corresponds to the position of the up button. It is used to detect the user's intention to approach the up button without physical contact.

[0009] A third non-contact capacitive sensor is located on the other side of the external call panel and corresponds to the position of the down button. It is used to detect the user's intention to approach the down button without physical contact.

[0010] The data acquisition module is electrically connected to a first non-contact capacitive sensor, a second non-contact capacitive sensor, and a third non-contact capacitive sensor, respectively. It is used to acquire a first sensing signal output by the first non-contact capacitive sensor and convert the first sensing signal into a first electrical signal; acquire a second sensing signal output by the second non-contact capacitive sensor and convert the second sensing signal into a second electrical signal; and acquire a third sensing signal output by the third non-contact capacitive sensor and convert the third sensing signal into a third electrical signal.

[0011] The data processing module, electrically connected to the data acquisition module, is used to preprocess the first sensing signal to generate an image signal; then extract fingerprint features from the image signal, compare the extracted fingerprint features with existing fingerprint features in the database to verify the user's identity; when the comparison results match, activate the second and third non-contact capacitive sensors; and receive a second electrical signal when the second non-contact capacitive sensor is activated, compare the second electrical signal with a first preset threshold, and output an uplink command when the signal is greater than the first preset threshold; and receive a third electrical signal when the third non-contact capacitive sensor is activated, compare the third electrical signal with a second preset threshold, and output a downlink command when the signal is greater than the second preset threshold.

[0012] The elevator control module, connected to the data processing module, is used to control the elevator's actuators to perform corresponding operations based on upward or downward commands.

[0013] In one alternative implementation, the system further includes:

[0014] An internal call panel is located inside the elevator; one side of the internal call panel is provided with an elevator door open button, an elevator door close button, and multiple floor buttons; the elevator door open button, elevator door close button, and multiple floor buttons are all integrally formed with the internal call panel;

[0015] A first non-contact capacitive sensor array is located on the other side of the inner call panel to detect the user's intention to approach the elevator open button, elevator close button, or floor button without physical contact.

[0016] The data acquisition module is electrically connected to the first non-contact capacitive sensor array and is also used to acquire the fourth sensing signal output by the first non-contact capacitive sensor array and convert the fourth sensing signal into a fourth electrical signal.

[0017] The data processing module is also used to receive a fourth electrical signal and use a pre-trained machine learning algorithm to parse and identify the fourth electrical signal to generate the user's elevator control instructions; the elevator control instructions include door opening instructions, door closing instructions, and instructions to reach the target floor.

[0018] The elevator control module is also used to control the elevator's actuators to perform corresponding operations based on the user's elevator control commands.

[0019] In one alternative implementation, the system further includes:

[0020] A gesture recognition device is installed inside the elevator; one side of the gesture recognition device is provided with a gesture recognition panel and a confirmation button; the gesture recognition panel and the confirmation button are both integrally formed with the gesture recognition device;

[0021] A second non-contact capacitive sensor array is located on the other side of the gesture recognition device, used to detect the user's hand gestures and intention to approach the confirmation button without physical contact; the hand gestures include swiping up, swiping down, swiping left, and swiping right.

[0022] The data acquisition module is electrically connected to the second non-contact capacitive sensor array and is also used to acquire the fifth sensing signal output by the second non-contact capacitive sensor and convert the fifth sensing signal into a fifth electrical signal.

[0023] The data processing module is also used to preprocess the fifth electrical signal and use a pre-trained machine learning algorithm to extract the temporal features of the preprocessed fifth electrical signal; then, the extracted temporal features are compared with the template gestures in the database; when the comparison results are consistent, the floor is selected until the target floor is displayed and the user is identified as approaching the confirm button, and then an instruction to reach the target floor is generated.

[0024] The elevator control module is also used to control the elevator's actuators to perform corresponding operations based on the instruction to reach the target floor.

[0025] In one optional embodiment, the dielectric in the first non-contact capacitive sensor, the second non-contact capacitive sensor, the third non-contact capacitive sensor, the first non-contact capacitive sensor array, and the second non-contact capacitive sensor array all adopt a cylindrical microstructure with grooves on the upper surface, and a protrusion array is formed at the grooves.

[0026] In one optional embodiment, the protrusion array is a 3×3 array of cube pillars, a cone array, or a cone-like array; the cone-like array is formed by stacking multiple cylinders with progressively increasing radii from top to bottom.

[0027] In one optional embodiment, the dielectric in the first non-contact capacitive sensor, the second non-contact capacitive sensor, the third non-contact capacitive sensor, the first non-contact capacitive sensor array, and the second non-contact capacitive sensor array is made of at least one of PDMS material, or a composite material formed by PDMS and carbon nanotubes, or a composite material formed by PDMS, carbon nanotubes, and silver nanowires.

[0028] In one optional implementation, the inner call panel, outer call panel, fingerprint recognition device, fingerprint recognition button, up button, down button, elevator door open button, elevator door close button, multiple floor buttons, gesture recognition panel and confirmation button all adopt a groove structure.

[0029] In one optional implementation, both the inner call panel and the outer call panel are equipped with LED displays 4 to display the current floor of the elevator.

[0030] In one alternative implementation, the data processing module is integrated into a Raspberry Pi or an FPGA.

[0031] In one alternative embodiment, the second non-contact capacitive sensor array includes 36 non-contact capacitive sensors and 12 analog switches; all non-contact capacitive sensors are arranged in a 6×6 array, with each row and each column of non-contact capacitive sensors connected to an analog switch.

[0032] The present invention has the following beneficial effects:

[0033] 1. Elevator buttons based on non-contact capacitive sensors use a proximity method to control the elevator, which can avoid the risk of cross-infection among elevator passengers and also prevent button wear and dirt accumulation.

[0034] 2. The non-contact capacitive sensor used in this embodiment of the invention has low material cost, is easy to integrate and deploy, has low maintenance cost, high flexibility and scalability, and high convenience.

[0035] 3. By combining machine learning, the changes in capacitance values ​​caused by different substances approaching the non-contact capacitive sensor can be distinguished, effectively realizing the function of preventing accidental touch.

[0036] 4. Using PDMS, PDMS-carbon nanotubes, and PDMS-carbon nanotube-silver nanowires as the dielectric of the sensor and fabricating a dielectric with a microstructure can enhance the electric field distribution, thereby improving the sensitivity and recognition accuracy of the non-contact capacitive sensor. At the same time, it can also optimize the electric field distribution, reduce the influence of external electromagnetic interference on the sensor, and improve the sensor's anti-interference ability. Attached Figure Description

[0037] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0038] Figure 1 This is a structural framework diagram of a non-contact elevator control system based on machine learning according to an embodiment of the present invention;

[0039] Figure 2 This is a schematic diagram of the external call panel according to an embodiment of the present invention;

[0040] Figure 3 This is a schematic diagram of the structure of a contactless capacitive sensor according to an embodiment of the present invention;

[0041] Figure 4 This is a schematic diagram of the structure of a non-contact capacitive sensor according to an embodiment of the present invention;

[0042] Figure 5 This is a schematic diagram of the internal breathing plate according to an embodiment of the present invention;

[0043] Figure 6 This is a schematic diagram of the structure of a gesture recognition device according to an embodiment of the present invention;

[0044] Figure 7 This is a circuit schematic diagram of a second non-contact capacitive sensor array according to an embodiment of the present invention;

[0045] Figure 8 This is a schematic diagram of a microstructure with a 3×3 cubic columnar array according to an embodiment of the present invention;

[0046] Figure 9 This is a schematic diagram of a microstructure with a 3×3 cone-like array according to an embodiment of the present invention;

[0047] Figure 10 This is a schematic diagram of a microstructure with a 3×3 conical array according to an embodiment of the present invention;

[0048] Figure 11 This is a schematic diagram of a microstructure formed by molding from three materials according to an embodiment of the present invention.

[0049] Figure label:

[0050] 1. External call panel; 11. Fingerprint recognition button; 12. Up button; 13. Down button; 2. Internal call panel; 21. Elevator door open button; 22. Elevator door close button; 23. Floor button; 3. Gesture recognition device; 31. Gesture recognition panel; 32. Confirm button; 4. LED display screen. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] In the description of this invention, it should be noted that the terms "upper," "lower," "left," "right," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the purpose of simplifying the description of this utility model and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this utility model. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0053] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can also refer to the internal connection of two components; and they can refer to a wireless connection or a wired connection. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0054] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0055] With the rapid development of artificial intelligence technology and the widespread adoption of intelligent devices, human-computer interaction is gradually shifting from contact-based operation to contactless and natural interaction. This shift not only enhances the user experience but also demonstrates enormous potential in areas such as public health and safety, and assistance for special groups. Among these advancements, contactless capacitive sensors and gesture recognition based on them, as crucial components of contactless human-computer interaction, are gradually becoming a new bridge connecting people and intelligent devices. Furthermore, gesture recognition technology, based on the cross-disciplinary integration of computer science, sensor technology, and artificial intelligence, captures, analyzes, and recognizes human gestures to enable interaction with devices. This technology is not only applicable to smart devices used by ordinary consumers but also plays a significant role in fields such as healthcare, education, and industrial control.

[0056] In related technologies, elevator control for upward or downward movement often employs methods such as contact button operation, vision-based gesture control, infrared-based gesture control, or radio frequency-based gesture control. However, contact button operation has the following drawbacks: First, mechanical buttons are prone to wear and tear after prolonged use, leading to sluggish response or malfunction. Second, in high-traffic public places, the surface and crevices of mechanical buttons easily accumulate dirt and bacteria, increasing the risk of cross-infection, and the dirt in the crevices is also difficult to clean. Vision-based gesture control, on the other hand, heavily relies on the performance and stability of external cameras and image processing algorithms, undoubtedly increasing system complexity. It is also highly susceptible to factors such as lighting, obstructions, and the complexity of hand movements, leading to decreased recognition accuracy. Furthermore, visual recognition requires capturing and analyzing user hand movements, which may raise concerns about user privacy and security.

[0057] Furthermore, infrared gesture control primarily utilizes the transmission and reception of infrared rays to achieve contactless operation. Infrared transmitters and receivers are installed on elevator buttons. When a finger or other object approaches the button, it blocks the infrared beam, causing a change in the received signal. By detecting this change, the elevator system determines the selected floor. However, this method not only requires more complex hardware, leading to higher costs, but infrared light is also susceptible to interference from ambient light; direct sunlight can cause misjudgments. Moreover, because infrared technology cannot detect hands or other objects blocking infrared light, accidental touches are inevitable.

[0058] Secondly, gesture control based on radio frequency (RFID) technology transmits signals via radio waves. Typically, each button is equipped with an RFID tag. When a user holds an RFID-enabled card or mobile device near the button, the device emits radio waves to communicate with the RFID tag on the button. By verifying the user's identity and permissions, the elevator system can identify the floor the user intends to select and automatically execute the corresponding action. This method requires users to carry an RFID-enabled card while riding the elevator, which is inconvenient. Furthermore, contactless elevator buttons implemented with RFID technology may experience interference in buildings with a lot of metal structures.

[0059] Based on this, embodiments of the present invention provide a non-contact elevator control system based on machine learning, which is low in cost, high in recognition accuracy, prevents accidental touch, is less affected by the environment, and is highly convenient.

[0060] Figure 1 The diagram shows a structural schematic of a non-contact elevator control system based on machine learning according to an embodiment of the present invention, including: an external call panel 1, a first non-contact capacitive sensor, a second non-contact capacitive sensor, a third non-contact capacitive sensor, a data acquisition module, a data processing module, and an elevator control module.

[0061] Specifically, such as Figure 2 As shown, the external call panel 1 is located outside the elevator; one side of the external call panel 1 is provided with a fingerprint recognition button 11, an up button 12 and a down button 13; the fingerprint recognition button 11, the up button 12 and the down button 13 are all integrally formed with the external call panel 1.

[0062] A first non-contact capacitive sensor is located on the other side of the call panel 1 and corresponds to the position of the fingerprint recognition button 11. The first non-contact capacitive sensor is used to detect the user's intention to approach the fingerprint recognition button 11 without physical contact. A second non-contact capacitive sensor is located on the other side of the call panel 1 and corresponds to the position of the up button 12. It is used to detect the user's intention to approach the up button 12 without physical contact. A third non-contact capacitive sensor is located on the other side of the call panel 1 and corresponds to the position of the down button 13. It is used to detect the user's intention to approach the down button 13 without physical contact.

[0063] The data acquisition module is electrically connected to the first non-contact capacitive sensor, the second non-contact capacitive sensor, and the third non-contact capacitive sensor, respectively, for acquiring the first sensing signal output by the first non-contact capacitive sensor and converting the first sensing signal into a first electrical signal; for acquiring the second sensing signal output by the second non-contact capacitive sensor and converting the second sensing signal into a second electrical signal; and for acquiring the third sensing signal output by the third non-contact capacitive sensor and converting the third sensing signal into a third electrical signal.

[0064] The data processing module is electrically connected to the data acquisition module. The data processing module is used to preprocess the first electrical signal to generate an image signal; then, it extracts fingerprint features from the image signal and compares the extracted fingerprint features with existing fingerprint features in the database to verify the user's identity; when the comparison results match, it activates the second and third non-contact capacitive sensors; and it is used to receive the second electrical signal when the second non-contact capacitive sensor is activated, and compare the second electrical signal with a first set threshold. When the second non-contact capacitive sensor is activated, it outputs an uplink command; and it is used to receive the third electrical signal when the third non-contact capacitive sensor is activated, and compare the third electrical signal with a second set threshold. When the third non-contact capacitive sensor is activated, it outputs a downlink command.

[0065] The elevator control module is connected to the data processing module and is used to control the elevator's actuators to perform corresponding operations based on upward or downward commands.

[0066] Before describing the embodiments of the present invention, the sensor principle of the contactless capacitive sensor will first be explained. For example... Figure 3 and Figure 4 As shown, the bipolar contactless capacitive sensor includes two metal plates and a dielectric material (not shown) between them. When energized, an electric field exists between the two parallel plates. When an external object approaches the plates, it changes the electric field distribution between the plates and between the plates and the object, thereby causing a change in capacitance.

[0067] The formula for calculating the capacitance value of a contactless capacitive sensor is as follows:

[0068]

[0069] Where ε0 is the relative permittivity of vacuum, k is the electrostatic constant, C0 is the initial capacitance of the sensor, and ε r0 Let d0 be the initial relative permittivity of the sensor dielectric, d0 be the initial electrode spacing, A0 be the initial effective electrode area, C be the changed sensor capacitance, and ε0 be the initial relative permittivity. rLet A be the relative permittivity of the sensor dielectric after the change, d be the initial effective area of ​​the electrode, d be the initial electrode spacing, and ΔC / C0 be the relative capacitance change.

[0070] For example, when a finger approaches the fingerprint recognition button 11, a capacitive coupling is formed between the skin of the finger and the electrode of the first non-contact capacitive sensing, causing a change in the electric field distribution around the electrode and thus a change in capacitance. Since fingerprints are microscopically uneven, the distance and contact area between the skin and the electrode vary in different areas. Furthermore, each person's fingerprint has unique ridges and valleys, which affect the distribution and changes in the electric field. Therefore, by measuring and analyzing these changes in the electric field, fingerprint information can be extracted.

[0071] At this time, the data acquisition module collects the sensing signal output by the first non-contact capacitive sensor and converts it into an electrical signal. The electrical signal is then sent to the data processing module for amplification, filtering, and other preprocessing before being converted into an image signal. Fingerprint features are then extracted from the image signal and compared with existing fingerprint features in the database to achieve fingerprint recognition. This operation verifies the user's identity. If the comparison results are inconsistent, authentication fails, and the elevator control system does not respond. If the comparison results are consistent, authentication is successful, activating the second and third non-contact capacitive sensors, i.e., unlocking the up button 12 and the down button 13.

[0072] When a finger approaches the up button 12 and the down button 13, a specific range of capacitance values ​​are generated between the finger and the plates of the second or third non-contact capacitive sensor, respectively. At this time, the data acquisition module collects the sensing signal output by the first or second non-contact capacitive sensor and converts the sensing signal into an electrical signal. The electrical signal is then sent to the data processing module and compared with the corresponding first or second set threshold. If the value is greater than the first or second set threshold, it is considered that the up button 12 or the down button 13 has been pressed. At this time, an up command or a down command is sent to the elevator control module, and the LED inside the up command or down command will light up accordingly.

[0073] Finally, the elevator control module can perform corresponding operations based on the up or down command. In one example, assuming the user is on the third floor and the elevator is on the 23rd floor, when the user verifies their identity outside the elevator on the third floor and presses the down button 13, the elevator control module will control the corresponding actuator to descend the elevator to the third floor and open the elevator door.

[0074] In one alternative implementation, the machine learning-based non-contact elevator control system further includes an internal call panel 2 and a first non-contact capacitive sensor array.

[0075] Specifically, such as Figure 5 As shown, the inner call panel 2 is located inside the elevator; one side of the inner call panel 2 is provided with an elevator door opening button 21, an elevator door closing button 22 and multiple floor buttons 23; the elevator door opening button 21, the elevator door closing button 22 and the multiple floor buttons 23 are all integrally formed with the inner call panel 2.

[0076] It should be noted that the position, size, shape, and number of the elevator door opening button 21, elevator door closing button 22, and multiple floor buttons 23 on the inner call panel 2 can be set according to actual needs, and no specific limitation is made here.

[0077] The first non-contact capacitive sensor array is located on the other side of the inner call panel 2, and is used to detect the user's intention to approach the elevator open button 21, elevator close button 22 or floor button 23 without physical contact.

[0078] The data acquisition module is electrically connected to the first non-contact capacitive sensor array and is also used to acquire the fourth sensing signal output by the first non-contact capacitive sensor array and convert the fourth sensing signal into a fourth electrical signal.

[0079] The first non-contact capacitive sensor array consists of multiple non-contact capacitive sensors arranged in a matrix, which can continuously measure and monitor the capacitance changes between the electrodes and ground. Therefore, touch operations can be detected and located by utilizing the capacitance changes formed between a conductor (such as a human finger) and the electrodes of the multiple non-contact capacitive sensors.

[0080] The data processing module is also used to receive the fourth electrical signal and use a pre-trained machine learning algorithm to parse and identify the fourth electrical signal to generate the user's elevator control instructions; the elevator control instructions include door opening instructions, door closing instructions, and instructions to reach the target floor.

[0081] The elevator control module is also used to control the elevator's actuators to perform corresponding operations based on the user's elevator control commands.

[0082] For example, when a finger approaches a floor button 23, that is, when it approaches an electrode of the first non-contact capacitive sensor array, capacitive coupling is formed with that electrode, causing a change in the electrode's capacitance value. At this time, the data acquisition module captures this capacitance change by acquiring the sensing signal output by the first non-contact capacitive sensor and converts the sensing signal into an electrical signal. Then, the electrical signal is sent to the data processing module, where a pre-trained machine learning algorithm analyzes and identifies the electrical signal to determine the location of the touch point and the corresponding button indication. Based on the button indication corresponding to the touch point location, the module generates the user's elevator control commands. The elevator control commands include door opening commands, door closing commands, and reaching the target floor commands. Finally, the elevator control module controls the elevator's actuators to perform the corresponding operations based on the user's elevator control commands.

[0083] In one alternative implementation, the machine learning-based non-contact elevator control system further includes: a gesture recognition device 3 and a second non-contact capacitive sensor array;

[0084] Specifically, such as Figure 6 As shown, the gesture recognition device 3 is installed inside the elevator; one side of the gesture recognition device 3 is provided with a gesture recognition panel 31 and a confirmation button 32; the gesture recognition panel 31 and the confirmation button 32 are both integrally formed with the gesture recognition device 3.

[0085] The second non-contact capacitive sensor array is located on the other side of the gesture recognition device 3, and is used to detect the user's gestures and intention to approach the confirmation button without physical contact; the gestures include swiping up, swiping down, swiping left and swiping right.

[0086] It should be noted that the second non-contact capacitive sensor array, like the first non-contact capacitive sensor array, consists of multiple non-contact capacitive sensors arranged in a matrix. The difference is that the second non-contact capacitive sensor array also includes multiple analog switches, which are connected to the non-contact capacitive sensors in each row and column of the array, respectively.

[0087] In one alternative implementation, such as Figure 7 As shown, the second non-contact capacitive sensor array includes 36 non-contact capacitive sensors (P-Cx series) and 12 analog switches (USx series); the non-contact capacitive sensors are arranged in a 6×6 array, and each row and each column of non-contact capacitive sensors is connected to an analog switch.

[0088] Specifically, 6×6 contactless capacitive sensors (P-C1 to P-C36) are connected in rows and columns via wires. Each row and column has an analog switch at its end, forming a time-division multiplexed selection circuit. For example, when selecting P-C1 in row COL1 and column ROW1, the main controller applies a high level to pin 4A of analog switches US1 and US7, connecting the lower plate of capacitor C1 to pin 1B2 of US1 and leading it to pin Pcap1. Simultaneously, the upper plate of capacitor C1 is connected to pin 1B2 of US7 and led to pin Pcap2. The main controller then applies a low level to pin 4A of the remaining analog switches, causing the upper and lower plates of the remaining contactless capacitive sensors to connect to pin 3B1 of the analog switches and grounded. Repeating these steps allows selection of any one contactless capacitive sensor in the sensor array, while simultaneously shielding the coupling and crosstalk between the remaining contactless capacitive sensors.

[0089] The data acquisition module is electrically connected to the second non-contact capacitive sensor array and is also used to acquire the fifth sensing signal output by the second non-contact capacitive sensor and convert the fifth sensing signal into a fifth electrical signal.

[0090] The data processing module is also used to preprocess the fifth electrical signal and use a pre-trained machine learning algorithm to extract the temporal features of the preprocessed fifth electrical signal. Then, the extracted temporal features are compared with the template gestures in the database. When the comparison results are consistent, the floor is selected until the target floor is displayed and the user is recognized to be approaching the confirmation button, and then the instruction to reach the target floor is generated.

[0091] It should be noted that the template gestures in the database are pre-defined gestures.

[0092] The elevator control module is also used to control the elevator's actuators to perform corresponding operations based on the instruction to reach the target floor.

[0093] For example, when a user's finger approaches the gesture recognition panel 31 and makes an up, down, left, or right swipe gesture, non-contact capacitive sensors at different positions will sequentially detect changes in capacitance. The temporal characteristics of these changes reflect the trajectory of the finger movement, thus recognizing the gesture. At this time, the data acquisition module captures this temporal capacitance change by acquiring the sensing signal output from the second non-contact capacitive sensor and converts the sensing signal into an electrical signal. Then, the electrical signal is sent to the data processing module for amplification, filtering, and other preprocessing. A pre-trained machine learning algorithm is used to extract the temporal features from the preprocessed electrical signal. The extracted temporal features are then compared with template gestures in the database. When the comparison results match, a floor is selected until the target floor is displayed and the user is detected approaching the confirmation button, generating an "arrive at target floor" command. The elevator control module then controls the elevator's actuators to perform corresponding operations based on the "arrive at target floor" command.

[0094] In one example, suppose a user enters the elevator from the current 3rd floor and wants to go to the 23rd floor. The user can increment the current floor number by swiping up multiple times until the number reaches 23. Then, swiping near the confirmation button indicates confirmation of arrival at the 23rd floor. At this point, the data acquisition module will generate a command indicating arrival at the 23rd floor. Subsequently, the elevator control module will control the corresponding actuator to raise the elevator from the current 3rd floor to the 23rd floor and open the elevator doors.

[0095] It should be noted that the correspondence between user gestures and elevator operations can be automatically set according to needs, and no specific limitations are made here. For example, an upward swipe gesture can represent adding 1, adding 2, subtracting 1, subtracting 2 floors, etc.

[0096] In one optional embodiment, the dielectric in the first non-contact capacitive sensor, the second non-contact capacitive sensor, the third non-contact capacitive sensor, the first non-contact capacitive sensor array, and the second non-contact capacitive sensor array all adopt a cylindrical microstructure with grooves on the upper surface, and a protrusion array is formed at the grooves.

[0097] In one alternative implementation, such as Figures 8-10 As shown, the protrusion array is a 3×3 cube column array, cone array, or quasi-cone array; the quasi-cone array is composed of multiple cylinders whose radii increase sequentially from top to bottom.

[0098] In one optional embodiment, the dielectric in the first non-contact capacitive sensor, the second non-contact capacitive sensor, the third non-contact capacitive sensor, the first non-contact capacitive sensor array, and the second non-contact capacitive sensor array is made of at least one of PDMS material, or a composite material formed by PDMS and carbon nanotubes, or a composite material formed by PDMS, carbon nanotubes, and silver nanowires.

[0099] The dielectric material in this embodiment of the invention can be formed by 3D printing a microstructure with a 3×3 cubic columnar array, a conical array, or a quasi-conical array, using PDMS material, or a composite material formed by PDMS and carbon nanotubes, or a composite material formed by PDMS, carbon nanotubes, and silver nanowires, after casting. Figure 11 As shown.

[0100] The dielectric of this invention, through the microstructure manufactured using the above-described process, can alter the distribution of the electric field, making it more sensitive to minute changes in capacitance. The electric field change becomes more pronounced when an external object approaches. The microstructure increases the amount of capacitance change, thereby improving sensitivity and recognition accuracy. Simultaneously, it optimizes the electric field distribution, reducing the impact of external electromagnetic interference on the sensor and enhancing its anti-interference capability. Through specific microstructure design, the sensor can better shield unwanted electric and magnetic fields, thereby improving its performance in complex electromagnetic environments.

[0101] In one alternative implementation, the inner call panel 2, the outer call panel 1, the fingerprint recognition device 3, the fingerprint recognition button 11, the up button 12, the down button 13, the elevator door open button 21, the elevator door close button 22, the multiple floor buttons 23, the gesture recognition panel 31, and the confirmation button 32 all adopt a groove structure.

[0102] In this embodiment of the invention, the inner call panel 2, the outer call panel 1, the fingerprint recognition device 3, and all the buttons and panels thereon are designed with a groove structure, which can prevent passengers from touching the buttons when they come into contact with the elevator and causing cross-infection.

[0103] In one optional implementation, both the inner call panel 2 and the outer call panel 1 are equipped with LED displays 4 to display the current floor of the elevator.

[0104] In one alternative implementation, the data processing module is integrated into a Raspberry Pi or an FPGA.

[0105] Specifically, the data processing module is mainly composed of a Raspberry Pi or FPGA. A pre-trained machine learning model is ported to the Raspberry Pi or FPGA. This model has collected a large amount of data on different materials before porting, such as fingers, fabric, plastic bags, and wood. Feature extraction is performed on this data. The feature peaks of fingers are different from those of other materials, so after extensive training on this data, the model can distinguish fingers from other materials. The Raspberry Pi receives data from various contactless capacitive sensors, processes the data, and outputs control signals.

[0106] The process of building a machine learning model includes:

[0107] S1. Data Collection and Preprocessing:

[0108] Capacitance data from non-contact capacitive sensor tests are collected and preprocessed. Preprocessing includes data cleaning (e.g., handling missing and outlier values), feature selection, and transformation to ensure data accuracy and consistency. Feature selection and transformation include operations such as feature scaling, feature dimensionality reduction, and feature combination to improve model performance.

[0109] S2. Model selection and training:

[0110] A suitable machine learning model is selected based on the characteristics of the collected capacitance data. Common machine learning models include linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), and neural networks. Then, the collected data is classified into training and test sets in a 7:3 ratio. The selected model is trained using the training data, and parameters are tuned to achieve better performance.

[0111] S3. Model Evaluation and Testing:

[0112] The trained model is tested using a test set to evaluate its performance, including accuracy, precision, recall, F1 score, confusion matrix, and ROC curve.

[0113] S4, Model Deployment

[0114] After the machine learning model has been tested, it is deployed on a Raspberry Pi or FPGA. After deployment, the test dataset is uploaded to the Raspberry Pi or FPGA for testing and optimization of the results.

[0115] The present invention has the following beneficial effects:

[0116] 1. Elevator buttons based on non-contact capacitive sensors use a proximity method to control the elevator, which can avoid the risk of cross-infection among elevator passengers and also prevent button wear and dirt accumulation.

[0117] 2. The non-contact capacitive sensor used in this embodiment of the invention has low material cost, is easy to integrate and deploy, has low maintenance cost, high flexibility and scalability, and high convenience.

[0118] 3. By combining machine learning, the changes in capacitance values ​​caused by different substances approaching the non-contact capacitive sensor can be distinguished, effectively realizing the function of preventing accidental touch.

[0119] 4. Using PDMS, PDMS-carbon nanotubes, and PDMS-carbon nanotube-silver nanowires as the dielectric of the sensor and fabricating a dielectric with a microstructure can enhance the electric field distribution, thereby improving the sensitivity and recognition accuracy of the non-contact capacitive sensor. At the same time, it can also optimize the electric field distribution, reduce the influence of external electromagnetic interference on the sensor, and improve the sensor's anti-interference ability.

[0120] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A non-contact elevator control system based on machine learning, characterized in that, include: External call panel, located outside the elevator; One side of the external call panel is provided with a fingerprint recognition button, an up button, and a down button; the fingerprint recognition button, the up button, and the down button are all integrally formed with the external call panel; A first non-contact capacitive sensor is located on the other side of the external call panel and corresponds to the position of the fingerprint recognition button, used to detect the user's intention to approach the fingerprint recognition button without physical contact. A second non-contact capacitive sensor is located on the other side of the external call panel and corresponds to the position of the up button. It is used to detect the user's intention to approach the up button without physical contact. A third non-contact capacitive sensor is located on the other side of the external call panel and corresponds to the position of the down button. It is used to detect the user's intention to approach the down button without physical contact. The data acquisition module is electrically connected to a first non-contact capacitive sensor, a second non-contact capacitive sensor, and a third non-contact capacitive sensor, respectively. It is used to acquire a first sensing signal output by the first non-contact capacitive sensor and convert the first sensing signal into a first electrical signal; acquire a second sensing signal output by the second non-contact capacitive sensor and convert the second sensing signal into a second electrical signal; and acquire a third sensing signal output by the third non-contact capacitive sensor and convert the third sensing signal into a third electrical signal. The data processing module, electrically connected to the data acquisition module, is used to preprocess the first sensing signal to generate an image signal; then extract fingerprint features from the image signal, compare the extracted fingerprint features with existing fingerprint features in the database to verify the user's identity; when the comparison results match, activate the second non-contact capacitive sensor and the third non-contact capacitive sensor; and receive a second electrical signal when the second non-contact capacitive sensor is activated, compare the second electrical signal with a first set threshold, and output an uplink command when the signal is greater than the first set threshold. And is used to receive a third electrical signal when the third non-contact capacitive sensor is activated, and compare the third electrical signal with a second set threshold. When the signal is greater than the second set threshold, a downlink command is output. The elevator control module, connected to the data processing module, is used to control the elevator's actuators to perform corresponding operations based on upward or downward commands. The system also includes: An internal call panel is located inside the elevator; one side of the internal call panel is provided with an elevator door open button, an elevator door close button, and multiple floor buttons; the elevator door open button, elevator door close button, and multiple floor buttons are all integrally formed with the internal call panel; A first non-contact capacitive sensor array is located on the other side of the inner call panel to detect the user's intention to approach the elevator open button, elevator close button, or floor button without physical contact. The data acquisition module is electrically connected to the first non-contact capacitive sensor array and is also used to acquire the fourth sensing signal output by the first non-contact capacitive sensor array and convert the fourth sensing signal into a fourth electrical signal. The data processing module is also used to receive a fourth electrical signal and use a pre-trained machine learning algorithm to parse and identify the fourth electrical signal to generate the user's elevator control instructions; the elevator control instructions include door opening instructions, door closing instructions, and instructions to reach the target floor. The elevator control module is also used to control the elevator's actuators to perform corresponding operations according to the user's elevator control commands; The system also includes: A gesture recognition device is installed inside the elevator; one side of the gesture recognition device is provided with a gesture recognition panel and a confirmation button; the gesture recognition panel and the confirmation button are both integrally formed with the gesture recognition device; A second non-contact capacitive sensor array is located on the other side of the gesture recognition device, used to detect the user's hand gestures and intention to approach the confirmation button without physical contact; the hand gestures include swiping up, swiping down, swiping left, and swiping right. The data acquisition module is electrically connected to the second non-contact capacitive sensor array and is also used to acquire the fifth sensing signal output by the second non-contact capacitive sensor and convert the fifth sensing signal into a fifth electrical signal. The data processing module is also used to preprocess the fifth electrical signal and use a pre-trained machine learning algorithm to extract the temporal features of the preprocessed fifth electrical signal; then, the extracted temporal features are compared with the template gestures in the database; when the comparison results are consistent, the floor is selected until the target floor is displayed and the user is identified as approaching the confirm button, and then an instruction to reach the target floor is generated. The elevator control module is also used to control the elevator's actuators to perform corresponding operations based on the instruction to reach the target floor. The dielectrics in the first non-contact capacitive sensor, the second non-contact capacitive sensor, the third non-contact capacitive sensor, the first non-contact capacitive sensor array, and the second non-contact capacitive sensor array are all made of composite materials formed from PDMS, carbon nanotubes, and silver nanowires.

2. The system according to claim 1, characterized in that, The dielectric in the first non-contact capacitive sensor, the second non-contact capacitive sensor, the third non-contact capacitive sensor, the first non-contact capacitive sensor array, and the second non-contact capacitive sensor array all adopt a cylindrical microstructure with grooves on the upper surface, and a protrusion array is formed at the grooves.

3. The system according to claim 2, characterized in that, The protrusion array is a 3×3 cube column array, cone array, or quasi-cone array; each quasi-cone in the quasi-cone array is composed of multiple cylinders whose radius increases sequentially from top to bottom.

4. The system according to claim 1, characterized in that, The internal call panel, external call panel, fingerprint recognition device, fingerprint recognition button, up button, down button, elevator door open button, elevator door close button, multiple floor buttons, gesture recognition panel and confirmation button all adopt a groove structure.

5. The system according to claim 1, characterized in that, Both the inner and outer call panels are equipped with LED displays to show the current floor of the elevator.

6. The system according to claim 1, characterized in that, The data processing module is integrated into a Raspberry Pi or FPGA.

7. The system according to claim 1, characterized in that, The second non-contact capacitive sensor array includes 36 non-contact capacitive sensors and 12 analog switches; all non-contact capacitive sensors are arranged in a 6×6 array, and each row and each column of non-contact capacitive sensors is connected to an analog switch.

Images