CN113459106A

CN113459106A - Software robot control method

Info

Publication number: CN113459106A
Application number: CN202110867823.4A
Authority: CN
Inventors: 张成红; 徐卫华; 徐建
Original assignee: Suzhou Fulixun Intelligent Equipment Technology Co ltd
Current assignee: Suzhou Fulixun Intelligent Equipment Technology Co ltd
Priority date: 2021-07-22
Filing date: 2021-07-22
Publication date: 2021-10-01

Abstract

A soft robot control method comprises a central pattern generator CPG, wherein the biological mechanism of the central pattern generator CPG is that a movement signal required by an organism is generated through mutual coupling with a physical environment, and meanwhile, the movement is reasonably regulated by using a neural central control signal and sensor feedback information so as to generate a regular rhythm signal, the central pattern generator CPG is an oscillation unit consisting of a series of intermediate neurons, the whole central pattern generator CPG control network is a complex distributed neural network consisting of a neural oscillator and a multiple reflection feedback loop system in a coupling mode, the distributed neural network can be reconstructed, the boundary of the central pattern generator CPG control network is flexible, the intermediate neurons can be converted into another network, and a plurality of small networks can also form a large network with new characteristics. The invention can effectively control the soft robot and scientifically research the software science.

Description

Software robot control method

Technical Field

The invention relates to the field of soft robots, in particular to a central pattern generator CPG (coherent population generator) consisting of oscillation units consisting of a series of intermediate neurons, namely a soft robot control method for controlling a complex distributed neural network consisting of a neural oscillator and a multi-reflection feedback loop system in a coupling mode.

Background

The soft robot relates to multiple disciplines of materials, chemistry, machinery, control and the like, and although the application of ionic EAP material in the soft robot has achieved great results, before the soft robot is designed by using ionic EAP drivers, many existing challenges need to be solved. A software robot needs to solve many problems such as material and process, structural design, software control, etc.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provides a central pattern generator CPG consisting of oscillation units consisting of a series of intermediate neurons, namely a software robot control method for controlling a complex distributed neural network consisting of a neural oscillator and a multiple-reflection feedback loop system in a coupling mode.

The technical scheme adopted by the invention for solving the technical problems is as follows: a soft robot control method comprises a central pattern generator CPG, wherein the biological mechanism of the central pattern generator CPG is that a movement signal required by an organism is generated through mutual coupling with a physical environment, and meanwhile, the movement is reasonably regulated by using a neural central control signal and sensor feedback information so as to generate a regular rhythm signal, the central pattern generator CPG is an oscillation unit consisting of a series of intermediate neurons, the whole central pattern generator CPG control network is a complex distributed neural network consisting of a neural oscillator and a multiple reflection feedback loop system in a coupling mode, the distributed neural network can be reconstructed, the boundary of the central pattern generator CPG control network is flexible, the intermediate neurons can be converted into another network, and a plurality of small networks can also form a large network with new characteristics.

The invention has the advantages that the invention can realize the effective control of the soft robot and the scientific research of the software science, including the scientific research of the prototype and the mechanism of the organism, the model of the organism is expressed by a mathematical method, and the real prototype which can be tested in the engineering technology is manufactured according to the mathematical model. The research of the soft robot enters a life-like system stage integrating biological characteristics and a structure from an original exploration stage, a macro profiling and moving soft body stage and a biological performance and electromechanical system fusion stage.

Detailed Description

The software robot control method comprises a central pattern generator CPG, the biological mechanism of the central pattern generator CPG is that the biological mechanism generates the movement signal required by the organism through the mutual coupling with the physical environment, meanwhile, the neural control signal and the sensor feedback information are utilized to reasonably regulate the movement, therefore, regular rhythm signals are generated, the central pattern generator CPG is an oscillation unit consisting of a series of intermediate neurons, the whole central pattern generator CPG control network is a complex distributed neural network formed by coupling a neural oscillator and a multiple reflection feedback loop system, the distributed neural network can be reconstructed, the boundary of the central pattern generator CPG control network is flexible, the intermediate neurons can be switched from one network to another network, and a plurality of small networks can also form a large network with new characteristics.

The invention discloses a soft robot control method, wherein a distributed neural network is divided into two network models of a chain network and a mesh network, the behavior of a central pattern generator CPG is the collective behavior of the networks, each oscillator in the central pattern generator CPG has own oscillation frequency, the central pattern generator CPG network shows the integral behavior and can oscillate at the same frequency, the control mechanism of the central pattern generator CPG comprises neurophysiology and biomechanics, the central pattern generator CPG network exists in an animal body, the central pattern generator CPG network is formed by a plurality of related nerve units, one or more units can be used for controlling the same degree of freedom, and the flexibility of the integral movement of the animal is coordinated; the central pattern generator CPG network can generate the rhythm control signals individually, and the output of the central pattern generator CPG network can be adjusted by feedback information of the central nerves and the environment.

The basic principle of the central mode generator CPG is that the output signal of the central mode generator CPG unit provides control signals of the position, moment or speed and the like of the joints of the soft robot, the joints of the soft robot can be coordinated and controlled through a central mode generator CPG network consisting of a plurality of central mode generator CPG units, the central mode generator CPG network generates stable and natural phase relation through phase locking and coordinates multi-joint movement, so that different movement modes of the soft robot are realized, the soft robot and environmental information are accessed into the central mode generator CPG control network through various sensor signals, the environmental adaptability of a control strategy is improved, and the central mode generator CPG movement control mainly comprises biology, Computational neurology, neurophysiology, molluscs, robotics, central pattern generator CPG provides new theories and methods for soft robot motion control, enabling the revealing of animal coordinated motion.

The control network of the central mode generator CPG utilizes a drain integrator model to establish the segments, bodies and limbs of the soft robot, and adopts a genetic algorithm to adjust network control parameters, the control system of the central mode generator CPG can carry out two motion modes of swimming and crawling of a salamander, and the speed, direction and gait of the soft robot can be realized by adjusting a direct current input signal.

In the implementation of the soft robot control method, the central pattern generator CPG can be designed into a snake-shaped robot, the snake-shaped robot adopts a distributed control system to coordinate and control a plurality of degrees of freedom, the snake-shaped robot adopts circulation inhibition, the snake-shaped robot can realize three-dimensional motion, the snake-shaped robot adopts bidirectional circulation inhibition to simulate a snake rhythm motion mechanism, the joint of the snake-shaped robot is controlled, and three typical motion patterns of the snake-shaped robot are realized: the control mechanism of the central pattern generator CPG can be applied to the soft robot and the snake-shaped robot and becomes the motion control theory of the soft robot and the snake-shaped robot.

The invention relates to a soft robot control method, wherein a central pattern generator CPG can be designed into an octopus robot, the motion process of the octopus robot is dominated by peripheral nervous systems distributed on each arm, various complex motions can be realized, the central nervous system plays a role in secondary guidance and realizes environment interaction, a control strategy in the crawling process of the octopus robot is that a visual system collects the environment covering 360 degrees around, video information is preprocessed and then sends signals to a CNS, the CNS selects and determines the motion direction, trigger signals and initial values are respectively sent to PNS of corresponding arms, and after the PNS receives the control signals, muscles corresponding to the arms are selected to carry out corresponding control, so that the crawling motion of the octopus robot is realized.

The invention relates to a soft robot control method, wherein the output signal of a central pattern generator CPG is used for controlling the angle or the moment of the joint of the soft robot, the sine or sine-like output signal can meet the requirements for the winding crawling of the snake robot with low motion requirements, and for the walking control of the soft robot with high motion requirements, under the condition that the sine signal can not completely meet the motion control of the soft robot, the central pattern generator CPG unit can generate a specific control signal or the phase locking information of the central pattern generator CPG is combined with the environment information to generate a control signal with specific requirements for controlling the walking of the soft robot.

The implementation of the control method of the soft robot of the invention can develop a novel active soft driving material and a processing method thereof, and the development of the novel soft driving material and the processing and manufacturing method thereof are needed to realize the soft robot with the behavior characteristics similar to mollusks. The current materials are mainly EAP and rubber, etc., which have defects in controllability, stress, strain, response speed and stability. Therefore, the development of high-performance soft intelligent materials is the key of the development of soft robots. Because the soft material has the characteristic of easy deformation, the traditional hard material processing technology cannot be applied. The processing method for the soft material mainly comprises a shape deposition method, a fused deposition forming method and a laser ablation method based on a rapid prototyping technology, and the methods have the disadvantages of complex process, high cost and difficulty in realizing mass production.

The control method of the soft robot can improve the structural design and the mechanical property of the ionic EAP soft robot, most of the existing ionic EAP soft robots use the bending deformation similar to a cantilever beam, and under the form, the movement of an ionic EAP driving program is limited. The software robot control method of the invention has better possibility of large-scale deformation and simulation of flexible biological motion than a rigid mechanical robot, and has great development potential. The software robot control method provided by the invention can be used for refreshing Braille display for IPMC driver and sensor integrated design and touch concave-convex array construction, and can improve structural design to realize complex surface deformation, nonlinear deformation or rotation and improve drive output. The control method of the soft robot provided by the invention aims at the novel structural design and the mechanical property improvement of the soft robot, and plays a key role in the development of the ionic EAP soft robot.

The software robot control method can further research a driving mechanism and an electromechanical coupling model, and although the preparation, control and application of the IPMC material have been the focus of research for more than 20 years, the electromechanical coupling mechanism in the material is still not completely clear, so that the experimental research lacks of a theoretical basis. For example, the interpretation of actuator bending deformation under the action of an electric field includes at least five mechanisms: swelling caused by migration of hydrated cations, imbalance of electrostatic force caused by cation transmission, electrostatic interaction between electrode particles and ionic polymer, dipole-dipole interaction of ion cluster polarization and electrochemical reaction of surface electrode. Although a rich research model is built based on these mechanisms, none fully reveals the multiple physical processes that occur in the material. Therefore, determining the internal driving mechanism and electromechanical coupling model of ionic EAP is one of the key issues for practical application.

The invention is a soft robot control method that can study the interaction between the liquid environment and the driver, and the fact that the electric driving of ionic EAP relies on the transfer of ions and solvent is one of its unique advantages because it can be used in underwater environments. When ionic EAP drivers are exposed to air, continuous operation can result in reduced driver performance. If one wants to advance the development of these underwater robots, one needs to study the effect of the interaction between the liquid environment and the internal solution environment of the material on the kinematic stability of the ionic EAP. Possible factors to be investigated include the liquid pressure, the liquid flow rate, the ionic concentration in the liquid environment and the presence of microorganisms.

The software robot control method can realize effective control of the software robot, can carry out scientific research on software science, comprises the scientific research on the prototype and mechanism of organisms, expresses the model of the organisms by a mathematical method, and manufactures a real object prototype which can be tested in engineering technology according to the mathematical model. The research of the soft robot enters a life-like system stage integrating biological characteristics and a structure from an original exploration stage, a macro profiling and moving soft body stage and a biological performance and electromechanical system fusion stage.

Claims

1. A soft robot control method comprises a central pattern generator CPG, which is characterized in that, the biological mechanism of the central pattern generator CPG is to generate the movement signals required by the organism through the mutual coupling with the physical environment, meanwhile, the neural control signal and the sensor feedback information are utilized to reasonably regulate the movement, therefore, regular rhythm signals are generated, the central pattern generator CPG is an oscillation unit consisting of a series of intermediate neurons, the whole central pattern generator CPG control network is a complex distributed neural network formed by coupling a neural oscillator and a multiple reflection feedback loop system, the distributed neural network can be reconstructed, the boundary of the central pattern generator CPG control network is flexible, the intermediate neurons can be switched from one network to another network, and a plurality of small networks can also form a large network with new characteristics.

2. The soft robot control method of claim 1, wherein the distributed neural network is divided into two network models of a chain network and a mesh network, the behavior of the central pattern generator CPG is an aggregate behavior of the network, each oscillator in the central pattern generator CPG has its own oscillation frequency, the central pattern generator CPG network exhibits an integrated behavior and can oscillate at the same frequency, the control mechanism of the central pattern generator CPG comprises neurophysiology and biomechanics, the central pattern generator CPG network exists inside the body of the animal, the central pattern generator CPG network is composed of a plurality of associated neural units, one or more units can be used to control the same degree of freedom, and the flexibility of the overall movement of the animal is coordinated; the central pattern generator CPG network can generate the rhythm control signals individually, and the output of the central pattern generator CPG network can be adjusted by feedback information of the central nerves and the environment.

3. The method as claimed in claim 1, wherein the motion control mode of the central pattern generator CPG is different from the conventional control method, the central pattern generator CPG is based on the principle that the output signal of the central pattern generator CPG unit provides the position, moment or speed control signals of the joints of the soft robot, the joints of the soft robot can be coordinated and controlled by the central pattern generator CPG network composed of a plurality of central pattern generator CPG units, the central pattern generator CPG network generates stable and natural phase relationship by phase locking to coordinate the movement of the joints, thereby realizing different motion modes of the soft robot, the environment information of the soft robot is accessed to the central pattern generator CPG control network by various sensor signals to improve the environment adaptability of the control strategy, the central pattern generator CPG motion control mainly comprises biology, computational neurology, neurophysiology, molluscs and robotics, and the central pattern generator CPG provides a new theory and method for controlling the motion of the soft robot and can reveal the coordinated motion of animals.

4. The method as claimed in claim 1, wherein the central pattern generator CPG comprises a biomechanical model and a neural control model, and can implement two kinds of motions of water and land, the control network of the central pattern generator CPG uses a drain integrator model to establish the segments, bodies and limbs of the soft robot, and uses a genetic algorithm to adjust the network control parameters, the control system of the central pattern generator CPG can perform two kinds of motions of swimming and crawling of a salamander, and the speed, direction and gait of the motion of the soft robot can be implemented by adjusting the dc input signal.

5. The soft robot control method of claim 1, wherein the central pattern generator CPG is designed as a snake robot, the snake robot uses a distributed control system to coordinate and control multiple degrees of freedom, the snake robot uses cyclic suppression, the snake robot can realize three-dimensional motion, the snake robot uses bidirectional cyclic suppression to simulate snake rhythm motion mechanism, so as to control the joints of the snake robot, and realize three typical motion patterns of the snake robot: the control mechanism of the central pattern generator CPG can be applied to the soft robot and the snake-shaped robot and becomes the motion control theory of the soft robot and the snake-shaped robot.

6. The soft robot control method of claim 1, wherein the central pattern generator CPG is designed as an octopus robot, the octopus robot has a motion process that is dominated by peripheral nervous systems distributed in each arm, and can realize various complex motions, the central nervous system plays a secondary role, and realizes environment interaction, the control strategy in the octopus robot crawling process is that a visual system collects an environment covering 360 ° around, video information is preprocessed and then sends a signal to the CNS, the CNS selects and determines a motion direction, a trigger signal and an initial value are respectively sent to the PNS of the corresponding arm, and after receiving a control signal, the PNS selects a muscle corresponding to the arm to perform corresponding control, thereby realizing crawling motion of the octopus robot.

7. The soft robot control method according to claim 1, wherein the output signal of the central pattern generator CPG is used to control the angle or torque of the joints of the soft robot, the sinusoidal or sine-like output signal can meet the requirements for the snaking crawling of the snake-shaped robot with low motion requirements, and for the walking control of the soft robot with high motion requirements, the central pattern generator CPG unit can generate a specific control signal or the phase locking information of the central pattern generator CPG is combined with the environmental information to generate a specific control signal to control the walking of the soft robot under the condition that the sinusoidal signal cannot completely satisfy the motion control of the soft robot.