CC BY
1895 Waltham Nick. «CCD And CMOS Sensors. CMOS» // [Электронный ресурс]. Режим доступа: http://ankaa.unibe.ch/ (дата обращения: 07.05.2018).
6 Bikarna Pokhael. « Machine vision and object sorting. Chapter III. Machine vision process. Output» // [Электронный ресурс]. Режим доступа: https://www.theseus.fi/ (дата обращения: 07.05.2018).
WALKING ROBOTS. ZMP TECHNOLOGY Gritsay I.P.1, Krasilo M.S.2
'Gritsay Irina Petrovna — Senior Lecturer, SCIENTIFIC TECHNICAL TRANSLATION DEPARTMENT; 2Krasilo Mikhail Sergeevich — Student, FACULTY AUTOMATION, MECHATRONICS AND ROBOTICS, DON STATE TECHNICAL UNIVERSITY, ROSTOV-ON-DON
Abstract: the ability to move robots in hard and difficult conditions is a priority for many industries where robotic mechanisms are used. For example, rescuers, geologists and the military require robots that can overcome such obstacles that can't be overcome by wheeled or caterpillar vehicles. That is why scientists in many countries are developing walking, on two or more limbs, robots.
Keywords: ZMP-technology, walking robots, the projection the center of gravity, humanoid robots, difficult to pass terrain.
Many people have seen of humanoid robots, which have two arms, two legs and one head in movies and the pictures. In films, these robots move faster and more accurately than humans, but few of the audience thought that moving the robot on two "legs" is one of the most difficult tasks of robotics. But if this is such a difficult task, then why create walking robots if there are wheels and caterpillars? And the answer is simple, wheeled and caterpillar robots can't travel everywhere, for example, in mountains, in buildings, and walking robots can be useful in the analysis of blockages. Therefore, scientists began to develop walking robots. The first publications devoted to the theoretical and practical issues of creating walking robots belong to the 70s-80s of the 20th century. Moving the robot using the "legs", as already mentioned, is a difficult task of dynamics. Now in many countries there is a development of robots moving on two extremities, but these robots can't yet achieve such a stable movement, as in humans.
Also, many mechanisms are created, moving on three, four, and so forth limbs. To begin with, we list the main technological solutions for this topic that are currently in use.
1. Servo-drive + hydro mechanical drive: early technology for the construction of walking robots, implemented in a number of models of experimental robots manufactured by General Electric in the 1960 s.
2. ZMP technology: ZMP (Zero Moment Point, "zero point") is an algorithm used in robots similar to ASIMO by Honda. The on-board computer controls the robot in such a way that the sum of all external forces acting on the robot is directed toward the surface along which the robot moves.
3. Leaping robots: in the 1980s, Professor Marc Raibert of the English Leg Laboratory of the Massachusetts Institute of Technology developed a robot capable of maintaining balance by jumping using only one leg.
The movement of the robot resembles the movements of a person on the simulator pogo-stick. And now authors of the paper pay attention to ZMP-technology.
Fig. 1. Stable position of chair Fig. 2. Unstable position of chair 26
Let's the ZMP technology as an example with a chair. In Fig. 1 we see chair in a stable position. It will be on a stable position, if projection of centre of the gravity is in shaded area. But if projection of centre of the gravity isn't in shaded area (Fig. 2), chair will fall. The same principle with walking robots, the acting forces moment should be equal to zero. And so, you can make a robot walk by calculating all this, and making a motion so that it is always true. The robot above could lower its raised legs, shift its centre of mass forwards to the in shaded area of those feet, and then lift the other feet and move them forwards in the air. The problem is that this is a very slow and inefficient way of walking. To make a better walking system, you need to consider the dynamics of the body. System must be "dynamically stable". For example you suddenly froze in the middle of a step, then you will fall, because our walking is not statically stable. We rely on the inertia of our body at the step to move on to the next phase. If we were to rely on a statically stable gait, we would have to move our legs along the ground, spending a lot of time, shifting our weight between both feet, set on the floor. The Zero moment point is a bit like a "dynamic centre of mass." It takes into account not just positions, but forces and inertias too. It is a lot more complicated to calculate, but it makes robots far more fast, efficient and versatile. We can see this in Fig. 3.
But there is another problem, the robot must see where he puts his foot. Otherwise it can fall and damage yourself or the objects surrounding it (smash equipment, harm a person, crash a blockage and etc.). Therefore, stepping robots need a machine 3D-vision system. Digital image processing involves extraction of information from an image, it is the backbone of machine vision systems. Digital images are the combinations of different light intensity level. Each point in a digital image is a representation of the pixel value that is solved that corresponds to x- and y- coordinates in the image plane; it determines the intensity at that point. Machine vision processes the image in real time, that is, processing occurs at the time the image is acquired.
Thus, it represents a real image in real time. It extracts the 3D information from the 2D image of the object.[1]Thus, one can conclude that the ZMP technology is relatively simple in the implementation of the walking robots method of movement. Of course, it is not complicated with respect to other methods. Also the advantage of such a method can be a higher speed of movement of robots, as well as this method is closest to the gait of a people.
1. Gorgadze L., Israelyan H. // Machine vision in industry, International Scientific Journal "Innovative Science". № 1, 2018.
2. Dale Thomas // Imagine taking a completely stationary robot and calculating its Center of mass. [Electronic resource]. URL: http://www.quora.com/ (date of access: 25.02.2018).
3. From the history of the development of robotics // Alexey Derevyankin. [Electronic resource]. URL: http://www.robot.ipksko.kz (date of access: 03.03.2018).
4. Gritsay I., Yakubov B. // Holistic design optimization in mechatronics, Technologies of the XXI century: problems and development prospects: a collection of articles of the International Scientific and Practical Conference. June, 2017.
Ce ofg
Fig. 3. Scheme of all forces acting on the robot
References
