Text | Slow Hardcore said
Edit | Slow Hardcore said
The design of artificial robots is mainly rigid. Based on existing rigid body dynamics, they can be precisely controlled and exhibit good performance in speed, strength and repetitive tasks.
Basic introduction
In recent years, robots have used soft materials to achieve safer human-computer interaction, better adaptability to complex terrain, and better self-protection in extreme environments.
Soft robots can simulate biological locomotion mechanisms in greater depth than rigid robots, opening up endless possibilities in the field of robotics.
At this stage, most soft robots are manufactured on a macroscopic scale, driven by a variety of novel drive mechanisms, which can be divided into three categories: variable-length tendons, fluid drives, or electroactive polymers.
When it comes to precise operation or movement in a limited space, soft robots need to be reduced to small dimensions and powered or driven externally.
In this case, magnetic control shows its unique advantages, contributing to the study of external forces, where many different external forces are used to drive small robots, including swimming microbes and contracted cells, gliding behavior through microorganisms, chemical reactions, temperature, light, pH and remotely transmitted magnetic fields.
Of all these external forces, magnetism is particularly important because it can provide wide-area direct control, allowing for a variety of programming methods.
In addition, when robots are applied in vivo, robot control driven by external forces such as microorganisms, chemical reactions, temperature, light, pH and so on is more complicated. Therefore, it is very suitable to choose a magnetic field as the driving method for small soft robots.
While electromagnetic systems are preferred over permanent magnet systems, permanent magnet systems generate more force at a lower cost and offer greater flexibility. Therefore, a permanent magnet system is used for this work.
In nature, many organisms evolved legs to move their bodies to cope with complex terrain and various conditions. Currently, most small robots can only move in a simple way, such as rolling and crawling.
In the field of magnetic control, the hard magnetic particles of NdFeB in the legs are programmed to realize a magnetic multi-legged robot crawling forward in a paddling manner.
Later, someone designed a starfish-inspired million-soft robot capable of all-round movement. However, the current design of miniature multi-legged robots still has many shortcomings.
This is because the legs of the microrobot can only achieve the movements of a single joint, and the movement of these joints is determined before manufacturing.
In this paper, we propose a multi-legged soft-bodied robot similar to an echinoderm. The inhomogeneity of the magnet's magnetic field is used to realize independent drive of different legs.
With the cooperation of these legs, the robot performs different movement patterns.
Design & Manufacturing
Legs and/or feet are common in many living animals. These structures are used to support their body weight and provide effective movement, thus allowing them to constantly move to a favorable environment.
In this work, we designed a novel type of decorative column array soft robot to mimic the superior capabilities of these biological systems.
With such a design, the robot can realize the comprehensive functions inherent in a single legged animal, such as the octopus's adaptability to various environments, and the caterpillar's excellent obstacle passing ability.
The design principle of the robot is that on the basis of existing equipment, the size of the robot should be as small as possible and the distribution of legs can be controlled.
As shown in Figure 1, the robot is designed as a soft body with predefined evenly distributed slender legs. The body is made of pure silicone gel, and the legs are made of a silicon-iron mixture, driven by an external magnetic field.
Because only the robot's legs have magnetic particles, the robot's ability to move depends entirely on the legs. In order to maintain the topography of the robot, a mold-based manufacturing method is adopted.
In addition, since uniformly saturated magnetized NdFeB cannot be obtained, we use iron powder as a magnetic actuation source, which is biocompatible and has unique properties different from hard magnetic materials.
Figure 2 depicts the entire process of manufacturing a tetherless multi-legged soft robot. They consist of the following steps: First, combined molds are manufactured through 3D printing technology.
The mold consists of two parts. The first part is a square plate of 3 mm with a length of 45 mm. A series of through-holes with a diameter of 0.7 mm are evenly distributed in the center of the square. The second part is a square ring with a thickness of 3.5 mm, which can be embedded with the square ring to form a complete mold.
After the mold is made, the magnetic legs are made with a combination of silicone and iron particles. First, mix the A and B bottles of silica gel and iron powder in a plastic cup according to the mass ratio of 1:1:2.
After stirring well, put the plastic cup in a vacuum to degas for 5-7 minutes to minimize foaming. Then, pour the ferrosilicon mixture into the square plate, making sure each vias are filled.
After scraping off the excess silicone, two pieces of cut acrylic (also coated with release agent) are clamped to the square sheet with tape. After waiting for about 45 minutes, we remove the square piece from the magnet, mix it with the square ring, and place the mixed pure dragon skin 20 on it.
After the upper layer of pure silica gel and the lower layer of magnetic doped silicone are fully cured and firmly bonded together, the silicone is carefully removed from the mold with tweezers.
Finally, we check the mobility of the legs. The last step is crucial because if we waste too much time on one of the intermediate steps, the iron-silicon mixture will semi-solidify and the iron particles inside will not be able to rearrange to the optimal orientation.
Motion analysis
3.1. Magnetic field distribution of rectangular permanent magnets
In order to understand the magnetic field more intuitively, the magnetic field was simulated using finite element software. Figure 4 shows the distribution of the magnetic field at the interface-sub-interface.
The parameters of the magnets are set to the saturation state, so they are larger than the actual value. However, its magnetic field distribution in space gives us a good reference.
3.2. Principle of magnetic leg drive
Assuming that the diameter of the pure iron particles we use is approximately equal to the critical diameter, the domain structure of each particle can be considered single. Mix the iron powder with silica gel at a 1:1 mass ratio and put the mixture into a 3D printing mold, which is sprayed with mold release agent before use.
During the iron-silica powder curing process, an external magnetic field is applied along the long axis of the outrigger, so that the easy axis of the internal iron particles tends to align with the long axis of the outrigger, resulting in anisotropy of magnetization caused by manual intervention. To quantitatively study the leg's drive ability, we measured the magnetization of the leg under different magnetic fields with a vibrating sample magnetometer.
3.3. Motion analysis of multi-legged robots
Set the forwarding policy of the multi-legged robot to a series of repeating loops. The core of robot motion is to control the reciprocating motion of magnets associated with the robot.
As shown in Figure 5(a)-(d), at the beginning of a new cycle, the robot is stationary due to the balance of magnetic drag force and static friction.
Then, due to the deviation of the magnet, the tilted magnetic field acting on the front leg of the robot creates a magnetic moment that flips the front half of the robot upwards. The extremum of this deformation depends on the vertical distance between the magnet and the robot, as does the maximum deviation distance.
After this, the magnet returns and drags the robot forward until all of the robot's legs are back on the ground. It can be seen that during the entire movement, the drag force generated between the magnet and the magnetic leg plays a decisive role in the progress of the robot.
However, this does not mean that randomly placing magnets will allow the robot to advance in a steady pattern. In most cases, the robot is pulled to the minimum point of magnetic potential energy, its attitude is distorted, and it eventually loses controllability.
In order to maintain the stability of the robot's attitude, the relative position of the magnet and the robot should not be too close or too far away. Therefore, the movement of a multi-legged robot is very similar to a puppet show, with magnetism as invisible lines and magnetic legs as joints for puppets.
experiment
The robot's response to a magnet was tested, and the robot was placed on a flat acrylic plate, which was placed under the acrylic plate and kept at a distance of about 5 to 7 centimeters from the acrylic plate.
When the magnet at the bottom is moved in a certain way, the shape of the robot changes with its relative distance from the magnet. The study found that only a small part of the legs were consistently affected by the magnetic field during each exercise cycle, and the results are shown in Figure 6.
When the magnet approaches a distance threshold horizontally from the front of the robot, the front of the robot flexes for a short time and quickly flexes to a peak.
It's a delicate balance, and the key to keeping it is that the acceleration at which the magnet moves forward must be greater than the acceleration at which the robot moves forward.
However, each iteration of the magnet must reach a position that allows the robot to react. This makes it necessary to know the exact distribution of the magnetic field around the magnet.
To illustrate how the magnetic field changes, we measured the precise distribution of the magnetic field in the workspace, as shown in Figure 7.
Finally, the robot's comprehensive locomotion ability in the multi-terrain maze was tested. As shown in Figure 8, there are multiple routes in the maze with different terrain settings, including speed bumps, shallow pits, slopes, humps, and narrow corners.
conclusion
Legs play a vital role in animal movement, and they can provide adequate physical support, greater motor flexibility and better ability to cross obstacles.
In this paper, a multi-legged miniature soft robot is designed. With an improved magnetic powder-assisted molding method, the robot can move forward, backward, turn and cross obstacles under the control of a magnetic field.
In addition, the magnetic field distribution of the permanent magnet and the gait of the robot are analyzed from the force level, which lays a foundation for the subsequent magnetic field control.
Through experiments, we can know that our robot can walk out of the maze in 300 seconds, and the total length of the path exceeds 30 robot lengths.
bibliography
[1] S. Chiaverini, B. Siciliano, O. Egeland, A Robotic Arm Control Method Based on Least Squares Inverse Kinematics, Journal of Mechanical Engineering. Control system. Technology 2(2)(1994): 123-134.
[2] L. Prez, I. Rodríguez, N. Rodríguez, R. Usamentiaga, D.F. García, Robotic Navigation Based on Machine Vision Technology: A Comparative Review, Sensors 16(3) (2016) 335.
[3] S. Sheikholeslami, A. Moon, E.A. Croft, Research on the Application of Robotic Gestures in Human-Computer Interaction, Journal of Human-Computer Interaction. J. Robots. Res. 36 (5-7) (2017) 699-720
[4] S. Hirose, Y. Umetani, Development of flexible manipulators, mechanical engineering. Mach Theories 13(3)(1978): 351-359.
[5] D. Rus, M.T. Tolley, Design, Fabrication and Control of Soft Robots, Natural Sciences, 521(7553) (2015): 467-475.