Mechanical arm obstacle avoidance method based on EIT touch sensor

Abstract

The invention discloses a mechanical arm obstacle avoidance method based on an EIT tactile sensor, and belongs to the technical field of robot tactile perception. Firstly, tactile information is obtained through an EIT tactile sensor, and a tactile sensing image is reconstructed; then, the visual module starts to detect the target object and the box, if the visual module does not recognize the target object, the target object is fed back to the control module, and the tail end of the mechanical arm is controlled to move to the next area until the whole workbench is traversed; if the target object and the box are detected, obtaining a target position and a placement position through coordinate transformation; after the tail end of the mechanical arm moves to a target position, a target object is clamped through a clamping jaw, and then the mechanical arm moves to a placement position; in the moving process of the mechanical arm, an obstacle touches the mechanical arm, the EIT tactile sensor detects the change of the conductivity of a sensing area, and the position of a change center point is converted to the position below a mechanical arm base system and transmitted to the control module to re-plan a route; and finally, the mechanical arm touches the obstacle for multiple times during movement, and the control module enables the mechanical arm to bypass the obstacle and reach the target object placement position through multiple route plans. According to the method, the obstacle model can be actively sensed and established, and a safer and more reliable obstacle avoidance route can be planned.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot perception, and in particular relates to a robot arm obstacle avoidance method based on an EIT tactile sensor. Background Art

[0002] Integrating a tactile sensor on a robotic arm can give it the ability to perceive touch. After receiving the tactile information, the robotic arm processes the tactile information and replans its route based on the processed information, allowing the robotic arm to avoid obstacles, or understand a person's intentions and respond accordingly by analyzing the person's touch position and method.

[0003] The EIT tactile sensor works based on the principle of electrical impedance tomography (EIT). Multiple electrodes are evenly distributed on the surface of the sensor. By applying a constant current to certain electrodes and measuring the voltage changes on other electrodes to which no current is applied, the conductivity changes in the contact area can be reflected. By performing numerical inversion calculations on these measured voltage changes, the conductivity distribution image of the contact area, i.e., the tactile sensing image, can be reconstructed. Based on this tactile sensing image, the position, size, shape and other information of the contact area can be identified, providing a good basis for the robot to avoid obstacles. Therefore, the present invention proposes a robot obstacle avoidance method based on the EIT tactile sensor. Summary of the invention

[0004] In view of the deficiencies in the prior art, the technical problem that the present invention intends to solve is to propose a robot arm obstacle avoidance method based on EIT tactile sensors. The EIT tactile sensor is used to give the robot arm the ability of tactile perception, so that the robot arm can determine the position of the obstacle based on the tactile information obtained when it contacts the obstacle. During the exploration process, the robot arm obtains the overall position information of the obstacle through tactile perception, and plans a safe obstacle avoidance route based on this, and finally arrives at the target point safely. Compared with the original artificial potential field method, the improvement of the present invention is that the robot arm no longer relies solely on pre-set obstacle information, but can actively perceive and establish an obstacle model, thereby planning a safer and more reliable obstacle avoidance route. This enables the robot arm to flexibly respond to and successfully complete tasks in complex and changing environments.

[0005] The present invention solves the technical problem by adopting the following technical solution:

[0006] A robot arm obstacle avoidance method based on an EIT tactile sensor, characterized in that it comprises the following steps:

[0007] Step 1: Place the target object on the workbench, use the EIT tactile sensor to obtain tactile information, and reconstruct the tactile sensing image based on the tactile information;

[0008] Step 2: Input the object to be grasped at the terminal of the control module, and the vision module starts to detect the target object and the box used to place the target object. If the vision module does not recognize the target object, it will feedback to the control module, and the control module will control the end of the robot arm to move to search the next area until the entire workbench is traversed; if the target object is not detected on the entire workbench, it will output that the target object is not detected; if the target object and the box are detected, the center coordinates of the detected target object and the box are converted to the robot arm base coordinate system, and the target position coordinates and the placement position coordinates are obtained and fed back to the control module;

[0009] Step 3: The control module controls the end of the robotic arm to move toward the target position. After reaching the target position, the control module controls the gripper to grab the target object, and then the robotic arm starts to move toward the placement position.

[0010] Step 4: When the robot moves to the placement position, an obstacle on the path touches the robot. At this time, the EIT tactile sensor on the robot detects the change in conductivity of the sensing area, converts the position of the center point of the change to the robot base coordinate system and transmits it to the control module;

[0011] Step 5: After receiving the location coordinates of the obstacle, the control module replans the route and bypasses the obstacle to continue moving to the placement location;

[0012] Step 6: The robot arm will touch obstacles multiple times during its movement, and gradually improve the shape and position information of the obstacles, so that the control module can go through multiple route planning and finally bypass the obstacles to reach the target object placement location.

[0013] Furthermore, the EIT tactile sensor uses an FPC flexible circuit board as a substrate, the electrodes are evenly distributed on the circuit board in a 4x4 arrangement, a layer of graphene conductive paint is sprayed on the surface of the circuit board, all electrodes are connected to form a conductive plane, and the conductive plane is covered with multiple discrete high-conductivity fabrics.

[0014] Furthermore, the high-conductivity fabric is made of nano-silver fibers.

[0015] Compared with the prior art, the present invention has the following beneficial effects:

[0016] The present invention adopts a tactile sensor based on the EIT principle, which gives the robot arm the ability to sense the surrounding environment and respond. This EIT tactile sensor adopts a sparse electrode layout, which can not only significantly expand the sensing range, but also has a flexible substrate design, which is easy to stick on the surface of the robot arm. At the same time, the shape and size of the sensor can be customized according to actual needs to meet the needs of different robot applications. Compared with traditional dot matrix tactile sensors, this EIT tactile sensor does not require complex wiring and a large number of components, and can achieve large-area dynamic tactile perception and real-time imaging.

[0017] Through the EIT tactile sensor in the tactile module, the robot can sense and understand objects in the surrounding environment, as well as the user's touch and gesture movements. The control module is responsible for processing and analyzing these tactile and visual information, accurately identifying the user's operating intentions, and the robot makes corresponding action feedback based on the analysis results to achieve a high degree of interaction between man and machine. The human-machine interaction method based on tactile perception breaks the limitations of traditional remote control operations. The robot can not only actively sense obstacles in the environment, but also understand the user's operating needs and make flexible adjustments, greatly improving the intelligence and safety of human-machine collaboration. At the same time, users can also directly control and interact with the robotic arm through intuitive touch and gestures. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is the structural diagram of the EIT tactile sensor;

[0019] Figure 2 It is the workflow diagram of the obstacle avoidance system;

[0020] Figure 3 It is an obstacle avoidance flowchart. DETAILED DESCRIPTION

[0021] Specific embodiments are given below in conjunction with the accompanying drawings. The specific embodiments are only used to introduce the technical solutions of the present invention in detail and are not intended to limit the protection scope of the present application.

[0022] The EIT tactile sensor of the present invention is as follows Figure 1 As shown, the sensor uses a flexible substrate and can be covered on the surface of the robot as the robot skin to provide the robot with tactile perception capabilities. The preparation process of the sensor is as follows:

[0023] Using FPC flexible circuit board as the substrate, the electrodes are evenly distributed on the circuit board in a 4x4 arrangement. FPC flexible circuit board has good foldability and can be attached to the surface of the robot; spray a layer of graphene conductive spray paint on the surface of the circuit board, connect the above 16 electrodes to form a conductive plane; paste 24×24 discrete high-conductivity fabrics on the surface of chloroprene rubber and cover them on the conductive plane to form an EIT tactile sensor, cover the sensor on the robot arm, and connect the data acquisition system at the same time.

[0024] The high-conductivity fabric used in this embodiment is nano silver fiber. As a non-conductive soft material, chloroprene rubber can play a role in buffering force, making the sensor more evenly stressed.

[0025] The specific process of using the EIT tactile sensor to obtain voltage data and image is as follows:

[0026] The data acquisition card and the multiplexer board form a data acquisition system. The 16 electrodes of the EIT tactile sensor are connected to two analog multiplexers (MAX 306) installed on the multiplexer board. The multiplexer board is connected to the current driver and the data acquisition device. The data acquisition card is connected to the host computer through a network port. During the operation of the sensing system, the microcontroller STM32F107 controls the multiplexer to select different electrodes. The electrodes in the EIT tactile sensor are numbered E1, E2, E3, ..., E16 in sequence, and two adjacent electrodes are respectively formed into a corresponding electrode pair to form 16 electrode pairs E1E2, E2E3, E3E4, ..., E16E1. The excitation current is injected into the electrode pairs in the EIT tactile sensor in sequence. Among them, one electrode pair is selected each time to apply an excitation current to it, and the voltage data of 16 electrode pairs are obtained at the same time. After the excitation signal is applied to all 16 electrode pairs, all the obtained voltage data are taken as a group of measured voltage data, totaling 16×16; repeat the above steps to obtain multiple groups of voltage data.

[0027] The obtained voltage data is input into the computer and imaged using the imaging algorithm. When an external object applies pressure to the sensor, the conductive fabric in the pressed area will stick to the conductive plane, causing the conductivity of the area to change. This change in conductivity will be reflected in the collected surface electrode voltage data. The change in voltage data is input into the imaging algorithm to obtain the conductivity distribution map of the area, which is the sensor image.

[0028] The present invention also provides a robot arm obstacle avoidance system, comprising a control module, a vision module and a tactile module.

[0029] The control module is written in Python on the PC and uses Socket communication to send commands to the robot arm and gripper or receive information from the robot arm and gripper. At the same time, this module can also receive information from the tactile module and the visual module, convert the received information into task instructions and send them to the robot arm and gripper. This module is the central module of the system and is responsible for connecting various modules and coordinating and controlling the operation of various modules. The workflow is shown in the attached figure. Figure 2 shown.

[0030] The visual module uses a RealSense D435 depth camera to collect visual data. The camera is installed on the gripper at the end of the robotic arm and connected to the computer via USB. The visual module uses YOLOv5 as the target detection model to extract the target object in the visual scene. When the control module issues a command, the visual module will turn on the camera and load the YOLOv5 model. Once the target object is detected, the visual module converts its position coordinates to the base coordinate system of the robotic arm, and then feeds the converted coordinates back to the control module. The control module then controls the robotic arm and gripper to move and grab the target object based on this target position information.

[0031] The tactile module uses EIT tactile sensors to collect tactile information, and integrates the EIT tactile sensor array on the connecting rod surface of the UR5 robot arm to provide the robot arm with tactile perception capabilities. These sensors measure the voltage change data on their own surfaces in real time and convert it into a conductivity image through an imaging algorithm. When the connecting rod of the robot arm collides with an obstacle, the conductivity of the area covered by the sensor changes, and the collision position is displayed on the image. The tactile module converts the detected collision position coordinates to the robot arm base coordinate system through coordinate transformation, and then feeds this information back to the control module. Based on this, the control module replans the movement path of the robot arm to bypass obstacles, so that the robot arm has tactile perception and response capabilities.

[0032] To obtain the transformation matrix from the conductivity change position to the robot base coordinate system, we first need to use the conductivity image obtained by the imaging algorithm, select the coordinate origin and the coordinate axis direction on the obtained conductivity image to determine the coordinates of the sensor sensing area; then we need to calculate the coordinates of the robot covering the sensor area in the base coordinate system. The UR5 robot used in this example needs to use the DH parameters of the robot, and obtain the robot's forward motion matrix through forward kinematics. Then, according to the general matrix transformation formula from coordinate system i to coordinate system i-1, we get the transformation matrix covering the sensor position, and convert the coordinates of the sensor sensing area to the robot base coordinate system according to this transformation matrix.

[0033] As attached Figure 3 As shown, the obstacle avoidance method of the present invention comprises the following steps:

[0034] Step 1: First, place a variety of objects such as oranges, apples, etc. on the workbench and place a box for grabbing the target object and placing it there. Run the tactile module to obtain tactile information and display the reconstructed conductivity image in real time on the computer. Run the visual module to obtain visual information and display the image captured by the camera in real time on the computer.

[0035] Step 2: Input the object to be grasped, such as "apple", at the terminal of the control module. After the command is input, the vision module starts to detect the apple and the box used to place the target object. If the camera does not recognize the target object at present, it will feedback to the control module. The control module controls the end of the robotic arm to move to search the next area until the entire workbench is traversed. If the target object is not detected on the entire workbench, it will output "No target object detected!"; if the target object and the box are detected, the center coordinates of the detected target object and the box are converted to the base coordinate system of the robotic arm to obtain the target position coordinates and the placement position coordinates and feedback to the control module.

[0036] Step 3: After receiving the target position coordinates, the control module controls the end of the robotic arm to move toward the target position. When reaching the target position, the control module controls the gripper to grab the target object. After acquiring the target object, the robotic arm starts to move toward the placement position coordinates.

[0037] Step 4: When the robot moves to the placement position coordinates, obstacles can be placed on the path or the user can touch the robot arm link with his hand. At this time, the EIT tactile sensor on the robot arm link detects the change in conductivity of the sensing area and displays it on the computer. At the same time, the tactile module converts the position of the center point of the change to the robot arm base coordinate system and transmits it to the control module.

[0038] Step 5: After receiving the position coordinates of the obstacle, the control module uses the path planning algorithm to re-plan the route to bypass the obstacle and continue to move to the placement position coordinates. The path planning algorithm used in the embodiment of the present invention is the artificial potential field method, which is a robot motion planning method based on an artificial potential field. This method designs the movement of the robot in the surrounding environment as a movement in an abstract artificial gravitational field.

[0039] Step 6: During the movement of the robotic arm, simple touch detection cannot build a complete obstacle model. In most cases, it cannot explore the correct obstacle avoidance path in one go. Therefore, during the movement, the robotic arm will contact the obstacle multiple times to gradually improve the shape and position of the obstacle. The control module will eventually bypass the obstacle and reach the target object placement position after multiple route planning.

[0040] Any matters not described in the present invention are applicable to the prior art.

Claims

1. A robot arm obstacle avoidance method based on EIT tactile sensor, characterized in that: The steps include: Step 1: Place the target object on the workbench, use the EIT tactile sensor to obtain tactile information, and reconstruct the tactile sensing image based on the tactile information; Step 2: Input the object to be grasped at the terminal of the control module, and the vision module starts to detect the target object and the box used to place the target object. If the vision module fails to recognize the target object, it will feedback to the control module, and the control module will control the end of the robot arm to move to search the next area until the entire workbench is traversed; if the target object is not detected on the entire workbench, it will output that the target object is not detected; If the target object and the box are detected, the center coordinates of the detected target object and the box are converted to the robot base coordinate system, the target position coordinates and the placement position coordinates are obtained and fed back to the control module; Step 3: The control module controls the end of the robotic arm to move toward the target position. After reaching the target position, the control module controls the gripper to grab the target object, and then the robotic arm starts to move toward the placement position. Step 4: When the robot moves to the placement position, an obstacle on the path touches the robot. At this time, the EIT tactile sensor on the robot detects the change in conductivity of the sensing area, converts the position of the center point of the change to the robot base coordinate system and transmits it to the control module; Step 5: After receiving the location coordinates of the obstacle, the control module replans the route and bypasses the obstacle to continue moving to the placement location; Step 6: The robot arm will touch obstacles multiple times during its movement, and gradually improve the shape and position information of the obstacles, so that the control module can go through multiple route planning and finally bypass the obstacles to reach the target object placement location.

2. The robot arm obstacle avoidance method based on EIT tactile sensor according to claim 1, characterized in that: The EIT tactile sensor uses an FPC flexible circuit board as a substrate, and the electrodes are evenly distributed on the circuit board in a 4x4 arrangement. A layer of graphene conductive paint is sprayed on the surface of the circuit board, and all electrodes are connected to form a conductive plane, which is covered with multiple discrete high-conductivity fabrics.

3. The robot arm obstacle avoidance method based on EIT tactile sensor according to claim 2, characterized in that: The high-conductivity fabric is made of nano-silver fibers.

Images