A cell scratch device

Abstract

The present invention provides a cell scratch device, which includes a magnetic drive turntable, a composite bracket, a robotic arm and a control system. The magnetic drive turntable is embedded with a honeycomb electromagnetic array, which is dynamically adsorbed to the ferromagnetic layer at the bottom of the multi-well plate to achieve precise positioning; the composite bracket integrates a pneumatic telescopic positioning rod and a composite positioning module to ensure the stability of the scratch and monitor the cell state in real time; the end of the robotic arm is connected to a dynamic balance actuator, and a flat straight shovel-shaped needle is assembled to achieve high-precision scratch operation; the control system coordinates the work of each part through a multi-level control module, and has functions of automatic scratching, real-time monitoring and feedback control.

Description

Technical Field

[0001] The invention relates to the fields of biomedicine and tissue engineering, and in particular to a cell scratching device. Background Art

[0002] In biomedical research, the scratch test is a commonly used method for studying cell migration and wound healing. Traditional scratch tests are usually performed manually, which has disadvantages such as uneven scratches, uncontrollable depth, and inability to monitor in real time. In recent years, with the development of automation technology and precision manufacturing technology, a variety of automatic scratch devices have emerged, but these devices still have some shortcomings, such as insufficient scratch accuracy, inability to adapt to multi-well plates of different specifications, and lack of real-time monitoring and feedback control functions. Therefore, the development of a high-precision, intelligent scratch device is of great practical significance. Summary of the invention

[0003] The purpose of the present invention is to provide a high-precision, intelligent scratching device, which can realize accurate scratching operations on biological samples such as cells and has real-time monitoring and control functions to solve the shortcomings of the prior art.

[0004] The present invention is a scratching device, comprising:

[0005] The magnetic drive turntable has a honeycomb electromagnetic array embedded on its surface, which forms a dynamic adsorption with the ferromagnetic layer at the bottom of the porous plate. An equally divided scale ring and a grating sensor are set on the edge to output the turntable angle signal in real time.

[0006] The composite bracket has an integrated pneumatic telescopic positioning rod in the vertical part of the bracket, which cooperates with the positioning hole of the turntable in an air pressure seal. The horizontal part is provided with a composite positioning module, which includes a coaxial industrial camera and a microscope objective lens to collect bright field images and a laser Doppler vibrometer to generate a vibration velocity map.

[0007] The end of the robot arm is connected to a dynamic balance actuator through a ball-jointed flexible joint. The actuator has a built-in dual piezoelectric chip driver and a six-dimensional force sensor, and the output end is equipped with a replaceable flat straight spade-shaped needle.

[0008] The flat straight spade-shaped needle has a 30° acute angle at the front of the blade, a 15° wedge-shaped reinforcement portion in the middle, and a bionic microgroove on the surface. The groove depth h and the blade thickness u satisfy h=0.2u±0.05u, where h is in μm and u is in μm.

[0009] Preferably, the honeycomb electromagnetic array is composed of 128 independent electromagnetic units, each unit is arranged in a regular hexagon with a side length of 5.17 mm or 2.83 mm;

[0010] Each unit is equipped with a PID temperature control module, and the working temperature is maintained at 25 ± 0.5 °C with a temperature control accuracy of ±0.1 °C;

[0011] The distance between adjacent electromagnetic units is 1 / 3 of the center distance of a single hole of the porous plate to be measured. The porous plate is a 6-well plate or a 12-well plate, and the center distance of a single hole is 15.5 mm or 8.5 mm respectively.

[0012] Preferably, the laser Doppler vibrometers are distributed in an equilateral triangle. The side length of the triangle is 0.7 times the diameter of a single hole of the porous plate, and the deviation of the center point of the triangle from the optical axis of the microscope objective is < 10 μm.

[0013] Preferably, the driving signal of the bimorph actuator is a linearly swept frequency waveform modulated by a Hanning window, with a swept frequency range of 20 - 30 kHz and a swept frequency rate of 500 Hz / ms.

[0014] Preferably, the silicone damping layer of the ball - hinge flexible joint has a gradient Young's modulus distribution:

[0015] It is 5 ± 0.2 MPa near the actuator end and 0.5 ± 0.05 MPa near the robotic arm end;

[0016] The rate of change of the damping coefficient with temperature is < 0.001 / °C, and the working temperature range is 20 - 30 °C.

[0017] Preferably, the turntable positioning control module: is electrically connected to the grating sensor, configured to receive the turntable angle signal and calculate the rotational speed n, and according to dynamically adjust the adsorption force F of the electromagnetic array, and transmit the rotational speed n to the trajectory planning module through the CAN bus, where the unit of n is rpm and the unit of F is N;

[0018] The vision - mechanics detection module: is integrated into the composite positioning module, and includes: an image processing unit that performs Sobel edge detection on the bright - field images continuously collected by the industrial camera, extracts the cell density distribution and calculates the cell migration speed; a frequency - domain analysis unit that performs Fourier transform on the vibration velocity map of the laser Doppler vibrometer and extracts the stiffness characteristics; a data synchronization unit that uses FPGA hardware to align the timestamps of the image and vibration velocity data;

[0019] The trajectory execution control module: forms a closed - loop with the robotic arm and the bimorph actuator, and is configured to:

[0020] Generate a third - order Bezier curve path based on the 3D - Unet neural network fusing cell density and stiffness data;

[0021] Convert the path into a robotic arm joint angular velocity command through inverse kinematics solution;

[0022] Receive the contact force data Fz of the six - dimensional force sensor in real - time and the transmitted cell migration speed v. The unit of Fz is mN, and the unit of v is μm / h;

[0023] Calculate the needle - angle compensation amount θc through the Lyapunov stability controller,

[0024] ;

[0025] where the unit of θc is rad;

[0026] dFz / dt is the contact - force change rate, and the unit is mN / ms;

[0027] Kp is the proportional - gain coefficient, and the unit is rad·h / μm;

[0028] Kd is the differential - gain coefficient, and the unit is rad·h² / μm;

[0029] Drive the bimorph actuator to adjust the needle - cutting angle.

[0030] Preferably, the input layer of the 3D - Unet convolutional neural network includes a parallel bright - field image branch and a laser - vibration - velocity - map branch;

[0031] The output layer generates a multi - order Bessel - curve path with probability weights.

[0032] Preferably, the sampling frequency of the six - dimensional force sensor is 10 kHz, and the Z - axis resolution is 0.1 mN;

[0033] Configure a fast Fourier transform module to extract the stiffness frequency - domain features in the 0.1 - 100 Hz frequency band, and the frequency resolution is 0.1 Hz.

[0034] Preferably, the sampling frequency of the six - dimensional force sensor is 10 kHz, and the Z - axis resolution is 0.1 mN;

[0035] Configure a filter, and the contact - force change rate is calculated as follows: , where t is the current time point, and the unit is s; Δ t is the sampling time interval, and the unit is s; Fz is the contact - force data, and the unit is mN. Specific implementation mode

[0036] The scratch device of the present invention includes the following main components: The magnetic - drive turntable is used to fix and drive the porous plate to rotate, achieving precise angle control; the composite bracket is used to support and position the porous plate to ensure the stability of the scratch operation; the robotic arm is used to control the movement of the scratch needle to achieve high - precision scratch operation; the control system is used to coordinate the work of each part to achieve automated scratch operation and real - time monitoring.

[0037] The magnetic drive turntable is one of the key components of the present invention. A honeycomb electromagnetic array is embedded on its surface, forming a dynamic adsorption with the ferromagnetic layer at the bottom of the porous plate. The electromagnetic array is composed of 128 independent electromagnetic units, and each unit is arranged in a regular hexagon with a side length of 5.17 mm or 2.83 mm. Each electromagnetic unit is equipped with a PID temperature control module, and the working temperature is maintained at 25 ± 0.5 °C with a temperature control accuracy of ± 0.1 °C. The distance between adjacent electromagnetic units is 1 / 3 of the center distance of a single hole of the porous plate to be measured. The porous plate is a 6-well plate or a 12-well plate, and the center distance of its single hole is 15.5 mm or 8.5 mm respectively.

[0038] The magnetic drive turntable forms a dynamic adsorption with the ferromagnetic layer at the bottom of the porous plate through the honeycomb electromagnetic array, realizing the precise positioning and stable fixation of the porous plate. The PID temperature control module ensures the stable working temperature of the electromagnetic unit, avoiding the influence of temperature changes on the experimental results. The grating sensor outputs the turntable angle signal in real time, providing precise angle control for subsequent scratching operations. The design of the magnetic drive turntable enables the porous plate to remain stable during the scratching process, avoiding scratching errors caused by vibration or displacement. At the same time, the dynamic adsorption function of the electromagnetic array can quickly replace the porous plate according to needs, improving the experimental efficiency.

[0039] The composite bracket includes a vertical part and a horizontal part. The vertical part integrates a pneumatic telescopic positioning rod, which is pneumatically and hermetically matched with the positioning hole of the turntable to realize the rapid positioning and fixation of the porous plate. The horizontal part is provided with a composite positioning module, including a coaxial industrial camera and a microscopic objective lens for collecting bright-field images, as well as a laser Doppler vibrometer. The laser Doppler vibrometer generates a vibration velocity map. The composite bracket realizes the rapid positioning and fixation of the porous plate through the cooperation of the pneumatic telescopic positioning rod and the positioning hole of the turntable. The coaxial industrial camera and the microscopic objective lens in the composite positioning module collect bright-field images for observing the morphology and distribution of cells. The laser Doppler vibrometer generates a vibration velocity map for monitoring the vibration of cells, providing a reference for subsequent scratching operations. The design of the composite bracket enables the porous plate to remain stable during the scratching process, avoiding scratching errors caused by vibration or displacement. At the same time, the multi-modal detection function of the composite positioning module can monitor the state of cells in real time, providing precise guidance for scratching operations.

[0040] The end of the robotic arm is connected to the dynamic balance actuator through a ball-joint flexible joint. The actuator is built with a bimorph driver and a six-axis force sensor, and its output end is equipped with a replaceable flat straight shovel-shaped needle. The front section of the blade of the flat straight shovel-shaped needle is a 30° acute thin layer, the middle section transitions to a 15° wedge-shaped reinforcement part, and the surface is provided with bionic micro-grooves. The groove depth h and the blade edge thickness t satisfy h = 0.2t ± 0.05t. The robotic arm is connected to the dynamic balance actuator through a ball-joint flexible joint to achieve high-precision motion control. The bimorph driver provides precise driving force, and the six-axis force sensor detects the contact force in real time to ensure the stability and accuracy of the scratching operation. The design of the flat straight shovel-shaped needle makes the scratching operation more precise, and the design of the bionic micro-grooves can reduce the resistance during scratching and improve the scratching quality.

[0041] The present invention includes a control system, which includes a multi-level control module, and constructs the control relationship between components in the following order:

[0042] Turntable positioning control module: electrically connected to the grating sensor, configured to receive the turntable angle signal and calculate the rotational speed n, and according to dynamically adjust the adsorption force of the electromagnetic array, and transmit the rotational speed n to the trajectory planning module through the CAN bus, where the unit of n is rpm and the unit of F is N;

[0043] Vision-mechanics detection module: integrated in the composite positioning module, including: an image processing unit that performs Sobel edge detection on the bright-field images continuously collected by the industrial camera, extracts the cell density distribution and calculates the cell migration speed; a frequency-domain analysis unit that performs Fourier transform on the vibration velocity map of the laser Doppler vibrometer and extracts the stiffness characteristics; a data synchronization unit that uses FPGA hardware to align the timestamps of the image and vibration velocity data.

[0044] Trajectory execution control module: forms a closed loop with the robotic arm and the bimorph driver, and is configured to:

[0045] Based on the 3D-Unet neural network to fuse the cell density and stiffness data, generate a third-order Bezier curve path;

[0046] Convert the path into a robotic arm joint angular velocity command through inverse kinematics solution;

[0047] Real-time receive the contact force data Fz of the six-axis force sensor and the transmitted cell migration speed v, where the unit of Fz is mN and the unit of v is μm / h;

[0048] Calculate the needle angle compensation amount θc through the Lyapunov stability controller,

[0049] = +

[0050] Wherein: , ,

[0051] dFz / dt is the contact force change rate, with the unit of mN / ms; Kp is the proportional gain coefficient, with the unit of rad·h / μm;

[0052] Kd is the differential gain coefficient, with the unit of rad·h² / μm;

[0053] Drive the bimorph actuator to adjust the needle cutting angle.

[0054] The control system realizes the precise control of the scratching device through multi-level control modules. The first-level control module receives the turntable angle signal of the grating sensor, calculates the turntable speed, and dynamically adjusts the adsorption force of the electromagnetic array according to the turntable speed to ensure the stable fixation of the multi-well plate. The second-level control module obtains the bright-field image and the vibration velocity map through the composite positioning module, extracts the cell density distribution and stiffness characteristics, and provides a reference for the subsequent scratching operation. The third-level control module generates a multi-order Bessel curve path based on the 3D-Unet convolutional neural network by fusing the bright-field image and the vibration velocity data, converts the path instruction into a manipulator joint angular velocity instruction, and drives the manipulator to move. The six-axis force sensor detects the contact force in real time, calculates the needle angle compensation amount through the Lyapunov stability controller, drives the bimorph actuator to adjust the needle cutting angle, forms a closed-loop control, and ensures the accuracy and stability of the scratching operation. The design of the multi-level control module of the control system enables the scratching device to achieve high-precision automated operation, and the real-time monitoring and feedback control functions can ensure the accuracy and consistency of the scratching operation, improving the reliability and repeatability of the experiment.

[0055] During the scratching operation, place the multi-well plate on the magnetically driven turntable, form dynamic adsorption through the electromagnetic array and the ferromagnetic layer at the bottom of the multi-well plate, and realize the precise positioning and stable fixation of the multi-well plate. The grating sensor outputs the turntable angle signal in real time, and the control system calculates the turntable speed according to the signal and dynamically adjusts the adsorption force of the electromagnetic array according to the turntable speed to ensure the stability of the multi-well plate during the scratching process. The coaxial industrial camera and the microscopic objective lens in the composite positioning module collect the bright-field image, and the laser Doppler vibrometer generates the vibration velocity map. The data is sent to the control system after timestamp alignment. The control system generates a multi-order Bessel curve path based on the 3D-Unet convolutional neural network by fusing the bright-field image and the vibration velocity data. Convert the path instruction into a manipulator joint angular velocity instruction to drive the seven-degree-of-freedom manipulator to move. The six-axis force sensor detects the contact force in real time, calculates the needle angle compensation amount through the Lyapunov stability controller, drives the bimorph actuator to adjust the needle cutting angle, forms a closed-loop control, and ensures the accuracy and stability of the scratching operation.

[0056] The entire scratching device realizes high-precision scratching operations on biological samples such as cells through the coordinated work of a magnetically driven turntable, a composite bracket, a robotic arm, and a control system. The magnetically driven turntable and the composite bracket ensure the stable fixation and precise positioning of the multi-well plate. The robotic arm realizes high-precision scratching operations. The control system achieves automated scratching operations and real-time monitoring through multi-level control modules, ensuring the accuracy and consistency of the scratching operations.

[0057] The scratching device of the present invention realizes high-precision scratching operations on biological samples such as cells through the coordinated work of a magnetically driven turntable, a composite bracket, a robotic arm, and a control system, and has real-time monitoring and control functions. This device has the advantages of simple operation, high precision, and strong adaptability, can meet the high requirements for scratching experiments in biomedical research, and has broad application prospects.

Claims

1. A cell scratch device, characterized in that include: The magnetic drive turntable has a honeycomb electromagnetic array embedded on its surface, which forms a dynamic adsorption with the ferromagnetic layer at the bottom of the porous plate. An equally divided scale ring and a grating sensor are set on the edge to output the turntable angle signal in real time. The composite bracket has an integrated pneumatic telescopic positioning rod in the vertical part of the bracket, which cooperates with the positioning hole of the turntable in an air pressure seal. The horizontal part is provided with a composite positioning module, which includes a coaxial industrial camera and a microscope objective lens to collect bright field images and a laser Doppler vibrometer to generate a vibration velocity map. The end of the robot arm is connected to a dynamic balance actuator through a ball-jointed flexible joint. The actuator has a built-in dual piezoelectric chip driver and a six-dimensional force sensor, and the output end is equipped with a replaceable flat straight spade-shaped needle. The flat straight spade-shaped needle tip has a 30° acute angle at the front of the blade, a 15° wedge-shaped reinforcement portion at the middle, and a bionic micro-groove on the surface. The groove depth h and the blade thickness u satisfy h = 0.2u ± 0.05u, where the unit of h is μm and the unit of u is μm; And integrate the following control subsystems: Turntable positioning control module: Electrically connected to the grating sensor, configured to receive the turntable angle signal and calculate the rotational speed n, and dynamically adjust the adsorption force F of the electromagnetic array according to F = 0.1n 2 +5, transmit the rotational speed n to the trajectory planning module through the CAN bus, where the unit of n is rpm and the unit of F is N; Visual-mechanical detection module: integrated in the composite positioning module, including: image processing unit, which performs Sobel edge detection on the bright field images continuously acquired by the industrial camera, extracts the cell density distribution and calculates the cell migration speed; frequency domain analysis unit, which performs Fourier transform on the vibration velocity map of the laser Doppler vibrometer to extract the stiffness characteristics; The data synchronization unit uses FPGA hardware to achieve the timestamp alignment of image and vibration data; Trajectory execution control module: forms a closed loop with the robot arm and the dual piezoelectric chip driver, and is configured as follows: The cell density and stiffness data were fused based on the 3D-Unet neural network to generate a third-order Bezier curve path; The path is converted into the angular velocity command of the robot joint through inverse kinematics solution; Receive the contact force data Fz and the cell migration speed v sent by the six-dimensional force sensor in real time. The unit of Fz is mN and the unit of v is μm / h. The needle angle compensation θc is calculated by the Lyapunov stability controller, Where: K P = 3.2V 0.5 , K d = 1.8V -0.3 ; Where, the unit of θc is rad; dFz / dt is the contact force change rate, in mN / ms; Kp is the proportional gain coefficient, in rad·h / μm; Kd is the differential gain coefficient, with the unit of rad·h 2 / μm; The bimorph actuator is driven to adjust the needle cutting angle.

2. The device according to claim 1, characterized in that: The honeycomb electromagnetic array consists of 128 independent electromagnetic units, each of which is arranged in a regular hexagon with a side length of 5.17mm or 2.83mm; Each unit is equipped with a PID temperature control module, the operating temperature is maintained at 25±0.5℃, and the temperature control accuracy is ±0.1℃; The spacing between adjacent electromagnetic units is 1 / 3 of the center distance of a single hole of the multi-hole plate to be tested. The multi-hole plate is a 6-hole plate or a 12-hole plate, and the center distance of a single hole is 15.5 mm or 8.5 mm, respectively.

3. The device according to claim 1, characterized in that: The laser Doppler vibrometers are distributed in an equilateral triangle, the side length of the triangle is 0.7 times the diameter of a single hole in the porous plate, and the deviation between the center point of the triangle and the optical axis of the microscope objective is less than 10μm.

4. The device according to claim 1, characterized in that: The driving signal of the bimorph actuator is a linearly swept waveform modulated by a Hanning window, with a sweep range of 20 - 30 kHz and a sweep rate of 500 Hz / ms.

5. The device according to claim 1, wherein: The silicone damping layer of the ball hinge flexible joint has a gradient Young's modulus distribution: It is 5 ± 0.2 MPa near the actuator end and 0.5 ± 0.05 MPa near the robotic arm end; The temperature change rate of the damping coefficient < 0.001 / °C, and the temperature working range is 20 - 30 °C.

6. The device according to claim 1, wherein: The input layer of the 3D-Unet convolutional neural network includes a bright-field image branch and a laser vibration velocity map branch in parallel; The output layer generates a multi-order Bessel curve path with probability weights.

7. The device according to claim 1, wherein: The sampling frequency of the six-axis force sensor is 10 kHz, and the Z-axis resolution is 0.1 mN; A fast Fourier transform module is configured to extract the stiffness frequency domain characteristics in the frequency band of 0.1 - 100 Hz, and the frequency resolution is 0.1 Hz.

8. The device according to claim 1, wherein: The sampling frequency of the six-axis force sensor is 10 kHz, and the resolution of the Z-axis is 0.1 mN. A filter is configured, and the central difference method is used for calculating the contact force change rate: △t = 0.1 mS. Among them, t is the current time point, with the unit of s; Δt is the sampling time interval, with the unit of s; Fz is the contact force data, with the unit of mN.

Images