CN106891326B - A robot teaching method - Google Patents

Abstract

The invention discloses a kind of robot teaching methods, are suitable for robot (correlation) technical field.This method is a kind of teaching method based on gesture identification, gyroscope/accelerometer module is fixed on robotic user the back of the hand first, as the gesture of user changes, three-dimensional linear acceleration and three-dimensional angular velocity are obtained using gyroscope/accelerometer module, angle signal, signal is handled by single-chip microcontroller, micro or Large Amplitude Motion signal and joint space or the motor message of operating space are selected by control switch, signal is passed into driving motor and mechanical arm is driven to complete corresponding actions, the six-dimensional space movement of industrial robot end effector can also accordingly be controlled.The present invention may be implemented mechanical arm according to the teaching movement of operation user and make corresponding actions, human-computer interaction function can be better achieved, keep teaching operation safer, intelligent and user friendly by above-mentioned teaching method.

Description

A robot teaching method

technical field

The invention belongs to the technical field of robots, and in particular relates to a teaching method for robots.

Background technique

In modern industry, with the continuous improvement of intelligence and automation, industrial robots are more and more widely used in all walks of life, among which the teaching system is an important part of the robot control system. Traditional robot operation is usually carried out by mouse, keyboard or teaching box. The teaching process has a series of problems such as poor safety, low intelligence, complicated operation and unfriendly human-computer interaction.

Robot teaching has a relatively long history of research and development in the field of robotics at present, but gesture teaching research started late in China, and the technology is not yet mature. The present invention is significantly different from the previous patents. In the patent with the application number of 201610459874.2, the user obtains the gesture signal based on the position data of the gesture and the posture data of the gesture obtained by the Leap Motion sensor. The gesture is responsible for coarse adjustment, and the voice is responsible for fine adjustment. This teaching method needs to combine gesture and voice together. ,Indispensable. The patent of application number 201610159248.1 selects the gesture change posture according to the degree of agreement between the user's gesture motion feature points and the reserved posture in the motion database, and makes corresponding actions. This method requires a preset sample template and database. The teaching process described in the patent application No. 201510138234.7 is that the teach pendant controls the movement of the manipulator in different directions by detecting sensors arranged on the fingers of the manipulator. The teaching process is cumbersome, the operation is more complicated, and the degree of intelligence is low. The collected gesture signal described in the patent application No. 201610363903.5 needs to be matched with a plurality of preset signals in the preset signal library to generate the operation program. In this method, some gesture signals must be preset and stored in the preset signal library, and the preparation work before teaching is more complicated. The teaching method described in the patent application No. 201310183427.5 requires the user to operate the teaching tool to specify the teaching position. This method also needs to use a certain teaching tool, which is inconvenient to operate, and this method is limited to position teaching instead of Action teaching.

To sum up, a new technical solution is required to solve the above problems.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a teaching system that is simple, reliable, convenient and flexible in operation and more convenient for human-computer interaction, aims to solve the problems of poor safety and complicated operation in the traditional teaching process, and helps to improve human-computer interaction experience.

To achieve the above object, the present invention can adopt the following technical solutions:

A robot teaching method, the system used in the teaching method includes a sensing module, a single-chip microcomputer, a control switch, and a mechanical arm, and the sensing module includes a gyroscope and a three-axis accelerometer; the following steps are included:

(1), fix the gyroscope and the three-axis accelerometer module on the indicating object that gives the operation instruction, and the indicating object is movable;

(2) The gyroscope outputs the angular velocity and angular motion change data indicating the object, and the three-axis accelerometer outputs the three-dimensional acceleration data indicating the object;

(3), output the angular velocity, angular motion change, and three-dimensional acceleration data collected in step (1) to the single-chip microcomputer, and process the interference data such as gesture jitter and sensor clutter in the teaching process through Kalman filtering algorithm processing;

(4) Based on the angular velocity, angular motion change, and three-dimensional acceleration data, the single-chip microcomputer integrates the acceleration data twice to obtain the moving distance data that the robotic arm should perform, and directly outputs the angular motion data as the rotation angle that the robotic arm should perform. data;

(5), process the movement distance and rotation angle data that should be performed by the robot arm obtained in step (4), and discard the movement distance and/or rotation angle data that exceed the movement range of the robot arm; distance and/or rotation angle data, then go to the next step;

(6) For the movement distance and/or rotation angle data obtained in step (5) that conform to the motion range of the manipulator, the single-chip microcomputer transmits the signal to the manipulator motor through the I/O port, and drives the manipulator to complete the specified action.

Beneficial effects:

1. Through this teaching method of the robot, the user can freely choose to teach gestures in the joint space or operation space of the robot, which is more user-friendly and easier to fit the user's natural teaching method.

2. The user can teach the robot by selecting two modes: micro-motion or large-scale motion, which is more convenient for users to operate, avoids teaching errors caused by improper user operation, and greatly improves the teaching process of micro-motion teaching accuracy. By selecting the micro-movement method, the safety of the robot arm working under the condition of small movement variation can be greatly improved, and the accuracy of the movement of the robot arm can be further increased. Greater operational changes through gestures can also improve the The human-machine friendliness of the device; the choice of large-scale movement mode can reduce the actual change of gestures under the condition that the amount of movement of the mechanical arm is large, making the gesture operation more flexible and simple.

3. The movement boundary of the robot arm saved in advance in the microcontroller can prevent the movement of the robot arm from exceeding the feasible range due to improper operation by the user. On the one hand, it avoids the damage of the mechanical arm, and on the other hand, it also provides a certain guarantee for the safety of users.

4. This robot teaching method is carried out in real time, and there is no need to save the gesture change database in advance, so that the changes of the robot arm are more diverse and meet the actual production requirements, and it is also easy to maintain, and the improper teaching posture can be corrected in time. Make the changes of the robotic arm closer to the expected effect.

Further, the pointing object is the operator's hand.

Further, the single-chip microcomputer in the step (2) has the ability to perform Kalman filtering, that is, the linear system state equation can be used.

Further, when the single-chip microcomputer in the step (2) processes the signal through the Kalman filter algorithm, it needs to establish a state space model of the system:

Where Φ is the state transition matrix, Γ is the noise transition matrix, w(k) is the dynamic noise, C is the measurement matrix, v(k) is the measurement noise, X _k+1 , y(k) are the state and measurement at the time of use value;

The Kalman filtering process corresponding to the above equation is as follows:

P _k|k-1 =ΦP _k-1 Φ ^T +ΓQΓ ^T

K _k =P _k/k-1 C ^T (CP _k/k-1 C ^T +R] ^-1

P _k =(1-K _k C)P _k|k-1

in, In order to use the results predicted by the previous state, is the optimal result of the previous state, and P _k|k-1 is The corresponding error variance, P _k-1 is The corresponding error variance, Q is the equation matrix of the system dynamic white noise, and R is the equation matrix of the system observation noise.

Further, in step (4), the process of calculating the displacement by integrating the acceleration twice is as follows:

Let the acceleration measured by the accelerometer be: a(k)

Integrating the acceleration once gives the velocity:

Integrating the velocity signal once gives the displacement:

Among them, a(k) is the acceleration signal, v(k) is the velocity signal, s(k) is the displacement signal, a _k is the acceleration sampling value at time k, v _k is the velocity value at time k, a ₀ =0, v ₀ =0, Δt is the time difference between two samples at time k and time k-1.

Further, the control switch for selecting the motion signal should include two control buttons, each button contains two working states, correspondingly determines the output signal type is large motion signal or micro motion signal, joint space motion. Or manipulate spatial motion.

Description of drawings

Figure 1 is a schematic diagram of a robot teaching method.

Detailed ways

In order to make the purpose and technical solutions of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings of the embodiments of the present invention.

In order to illustrate the technical solutions of the present invention, the following descriptions are given in conjunction with the accompanying drawings and specific embodiments:

A robot teaching method, the operation steps of the method are as follows:

1) Fix the gyroscope/accelerometer module 1 on the back of the user's hand by means of Velcro; (process I in Figure 1)

2) Select the motion signal type and press the control switch button 2; (process II in Figure 1)

3) On the premise of ensuring that the worn gyroscope/accelerometer module does not change its attitude, the entire device is powered on, and the user performs gesture teaching. The gyroscope outputs the angular velocity and angular motion change data indicating the object, and the three-axis accelerometer outputs the 3D acceleration data indicating the object; the collected angular velocity, angular motion change, and 3D acceleration data are output to the single-chip microcomputer, and are processed by the Kalman filter algorithm to remove the teaching process. Gesture jitter and sensor clutter and other interference data in the robot; based on the angular velocity, angular motion change, and three-dimensional acceleration data, the single-chip microcomputer integrates the acceleration data twice to obtain the moving distance data that the robotic arm should perform, and outputs the angular motion data directly. It is the rotation angle data that the robot arm should execute.

Observe whether the single-chip microcomputer (component 3 in Figure 1) has an alarm, such as the alarm light flashing. If an alarm occurs, it means that the gesture teaching action performed by the user exceeds the motion range of the robotic arm (part 12 in the figure), the teaching signal data is discarded this time, and the robotic arm remains in its original state, and the user needs to make the correct indication again. Teaching operation; if it does not alarm, it means that the user's operation meets the requirements, and the robot arm will make corresponding teaching actions. (Process III in Figure 1).

When the single-chip microcomputer processes the signal through the Kalman filter algorithm, it needs to establish the state space model of the system:

Where Φ is the state transition matrix, Γ is the noise transition matrix, w(k) is the dynamic noise, C is the measurement matrix, v(k) is the measurement noise, X _k+1 , y(k) are the state and measurement at the time of use value;

The Kalman filtering process corresponding to the above equation is as follows:

P _k|k-1 =ΦP _k-1 Φ ^T +ΓQΓ ^T

K _k =P _k/k-1 C ^T (CP _k/k-1 C ^T +R] ^-1

P _k =(1-K ^k C)P _k|k-1

in, In order to use the results predicted by the previous state, is the optimal result of the previous state, and P _k|k-1 is The corresponding error variance, P _k-1 is The corresponding error variance, Q is the equation matrix of the system dynamic white noise, and R is the equation matrix of the system observation noise.

The process of calculating the displacement by integrating the acceleration twice as described above is as follows:

Let the acceleration measured by the accelerometer be: a(k)

Integrating the acceleration once gives the velocity:

Integrating the velocity signal once gives the displacement:

Among them, a(k) is the acceleration signal, v(k) is the velocity signal, s(k) is the displacement signal, a _k is the acceleration sampling value at time k, v _k is the velocity value at time k, a ₀ =0, v ₀ =0, Δt is the time difference between two samples at time k and time k-1.

The control switch button mentioned above determines the type of the output signal of the single-chip microcomputer, which in turn determines the different processing methods of the signal provided by the sensor in the single-chip microcomputer.

The switch contains two selection buttons 4 and 5. Each button has two states, which are divided into A, B and 1, 2 control states. The A and B buttons determine whether the output signal type is a large-scale motion signal or a micro-motion signal. , 1 and 2 buttons determine the output signal type as joint space motion signal or operation space motion signal.

If you choose A and 1, the single-chip microcomputer transmits a large-scale motion signal in the joint space to the robotic arm;

If you choose A and 2, the single-chip microcomputer transmits a large-scale motion signal in the operating space to the robotic arm;

If B and 1 are selected, the single-chip microcomputer transmits the micro-motion signal of the joint space to the robot arm;

If B, 2 are selected, that is, the single-chip microcomputer transmits the micro-motion signal of the operation space to the robot arm.

Among them, the control button 4 for determining the motion amplitude signal can set the ratio of the gesture change signal and the actual control signal according to the actual situation, and the ratio of the gesture change signal and the actual control signal can be set respectively, and an example will be described below.

In this embodiment, when the control switch determines that the single-chip microcomputer outputs a large-scale motion signal, the ratio of the user's gesture movement distance and angle amplitude to the mechanical arm movement distance and angle amplitude is 1:1. For example, if the user executes a straight line with a displacement of 20cm Teaching action, the robotic arm makes a 20cm linear action in the same direction;

When the control switch determines that the single-chip microcomputer outputs a trace motion signal, the ratio of the user's gesture movement distance and angle amplitude to the robot arm movement distance and angle amplitude is 20:1. If the user performs a linear teaching action with a displacement of 20cm, the robot arm will A 1cm linear motion in the same direction.

In one embodiment, the control switch determines that the output signal type is joint space or operation space motion signal depending on the actual operation requirements and scope. If the end of the robot arm is required to quickly reach a certain position far away, joint space control can be selected. button; if the end of the manipulator is required to perform an action more precisely, you can choose the operation space control button; you can also use the joint and operation space button in sequence according to your needs.

4) According to the movement mode determined by the control switch button, the single-chip microcomputer will perform different operations on the received gesture and posture change signals processed by Kalman filtering, and solve the driving motors 6 to 11 that drive the motion of each joint of the robotic arm. The range of motion that occurs, each drive motor sequentially corresponds to the six joints 13 to 18 on the robotic arm, so that the robotic arm can complete the entire gesture teaching process under the drive of the drive motor. (Process IV in the figure)

The teaching operation process of the method includes: coarse adjustment of the joint space, fine adjustment of the operation space, and robot teaching combined with the joint space and the operation space.

The rough adjustment of the joint space ensures that the end effector of the robotic arm can reach the working point in a relatively short time, which improves the work efficiency to a certain extent, saves the working time, and reduces the amount of calculation inside the microcontroller; in the operation space The fine-tuning is aimed at that when the end effector of the robot arm reaches the vicinity of the working point, it should reach the area to be processed (working) as safely and smoothly as possible to avoid damage such as collision between the end effector of the robot arm and the workpiece caused by improper operation. occurrence, enhances the user-friendliness of man-machine and improves the safety of work.

This combination of joint space and operation space to teach the robot can allow users to freely choose to teach gestures in the joint space or operation space of the robot, which is more humane and easier to meet the user's natural teaching needs; it is also carried out in real time. , there is no need to save the gesture change database in advance, so that the changes of the robot arm are more diverse and meet the actual production requirements, and it is also easy to maintain, and the improper teaching posture can be corrected in time, so that the changes of the robot arm are closer to the expected effect.

The above-mentioned embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited by the above-mentioned embodiments, and any other changes, modifications, substitutions, combinations, The simplification should be equivalent replacement manners, which are all included in the protection scope of the present invention.

Claims (6)

1. a robot teaching method, the system used by this teaching method comprises a sensing module, a single-chip microcomputer, a control switch, a mechanical arm, and the sensing module comprises a gyroscope and a three-axis accelerometer, and is characterized in that, comprising the following steps: (1), fix the gyroscope and the three-axis accelerometer module on the indicating object that gives the operation instruction, and the indicating object is movable; (2) The gyroscope outputs the angular velocity and angular motion change data indicating the object, and the three-axis accelerometer outputs the three-dimensional acceleration data indicating the object; (3), output the angular velocity, angular motion change, and three-dimensional acceleration data collected in step (1) to the single-chip microcomputer, and remove the gesture jitter and sensor clutter interference data in the teaching process through Kalman filtering algorithm processing; (4) Based on the angular velocity, angular motion change, and three-dimensional acceleration data, the single-chip microcomputer integrates the acceleration data twice to obtain the moving distance data that the robotic arm should perform, and directly outputs the angular motion data as the rotation angle that the robotic arm should perform. data; (5) Make a safety judgment on the movement distance and rotation angle data that the robot arm should execute obtained in step (4). When the angle that the motor is about to change exceeds the physically limited angle range of the robot arm, the program stops at the current position, and the robot The arm does not move, and the buzzer alarms, while discarding the movement distance and/or rotation angle data that exceeds the movement range of the robot arm; if the movement distance and/or rotation angle data that conform to the movement range of the robot arm are obtained, go to the next step; (6) For the movement distance and/or rotation angle data obtained in step (5) that conform to the motion range of the manipulator, the single-chip microcomputer transmits the signal to the manipulator motor through the I/O port, and drives the manipulator to complete the specified action. 2 . The robot teaching method according to claim 1 , wherein the pointing object is an operator's hand. 3 . 3 . The robot teaching method according to claim 1 , wherein the single-chip microcomputer in the step (3) has the ability to perform Kalman filtering, that is, a linear system state equation can be used. 4 . 4. robot teaching method according to claim 2, is characterized in that: When the single-chip microcomputer in the step (3) processes the signal through the Kalman filter algorithm, it needs to establish a state space model of the system: where Φ is the state transition matrix, Γ is the noise transition matrix, w(k) is the dynamic noise, C is the measurement matrix, v(k) is the measurement noise, X(k+1), y(k) are the states at the time of adoption and measured values; The Kalman filtering process corresponding to the above equation is as follows: P _k|k-1 =ΦP _k-1 Φ ^T +ΓQΓ ^T K _k =P _k/k-1 C ^T (CP _k/k-1 C ^T +R] ^-1 P _k =(1-K _k C)P _k|k-1 in, In order to use the results predicted by the previous state, is the optimal result of the previous state, and P _k|k-1 is The corresponding error variance, P _k-1 is The corresponding error variance, Q is the equation matrix of the system dynamic white noise, and R is the equation matrix of the system observation noise. 5. robot teaching method according to claim 1, is characterized in that: in step (4), the process that described acceleration is integrated twice to find displacement is as follows: Let the acceleration measured by the accelerometer be: a(k) Integrating the acceleration once gives the velocity: Integrating the velocity signal once gives the displacement: Among them, a(k) is the acceleration signal, v(k) is the velocity signal, s(k) is the displacement signal, a _k is the acceleration sampling value at time k, v _k is the velocity value at time k, a ₀ =0, v ₀ =0, Δt is the time difference between two samples at time k and time k-1. 6. The robot teaching method according to claim 1, characterized in that: the control switch for selecting the motion signal should include two control buttons, each button contains two working states, and the output is determined accordingly. The signal types are large-scale motion signals or micro-motion signals, joint space motion or manipulation space motion.