Robot called automatic device having a manipulator - a mechanical analogue of the human hand - and this manipulator control system. Both of these components can have a variety of devices - from very simple to extremely complex. Manipulator normally consists of hinged units as the human hand is made up of the bones of the joints, and ends with coverage, which is something of a hand of a human hand.
The links of the manipulator are movable relative to each other and can perform rotational and translational movements. Sometimes, instead of the gripper arm is the last link of any working tool, such as a drill, wrench, spray gun or torch.
Move the manipulator units provide a so-called drive - analogues of the muscles in the human hand. Usually, electric motors are used as such. Then, the actuator includes another gear (a system of gears that reduce the speed of the motor and torques increase) and an electrical control circuit that regulates the motor speed.
Also frequently used electric hydraulic drive. Its action is very simple. In a cylinder 1 in which a piston 2 is connected by a rod to the arm 3, enters the pressurized fluid, which moves the piston in one direction or another, and with it the "hand" of the robot. The direction of this motion is determined by which part of the cylinder (into the space above the piston or below) the currently flagged liquid. Hydraulic manipulator can inform and rotational motion. Similarly acting pneumatic actuator, but instead of liquid air is used here.
This is the outline of the manipulator device. With regard to the complexity of the tasks that may allow one or the other robot, they largely depend on the complexity and perfection of the control device. In general, to talk about three generations of robots: industrial, adaptive and robots with artificial intelligence.
The earliest examples of simple robots were created in 1962 in the United States. They were "Versatran" company "AMP Versatran" and "Yunimeyt" firm "Yunimeyshn Incorporated." These robots, as well as those that followed them, acting on a rigid, unchanging in the process of the program and were intended for the automation of simple operations at constant environmental conditions. As a control device for such robots could provide, for example, "programmable drum". He acted as follows: on a cylinder rotating motor, housed contacts the manipulator drives and around the drum - conductive metal plate that closes the contacts when they are touched. Contact arrangement is such that during rotation of the drum manipulator drives are included in the right time, and the robot to perform the programmed operations started in the correct sequence. Similarly, control may be performed using punch cards or tape.
It is obvious that even the slightest changes in the environment, the slightest glitch in the process, leading to a breach of the actions of the robot. However, they possess considerable advantages - they are cheap, simple, easily reprogrammed and should be able to replace human when performing heavy repetitive work. It is on this type of work, and were first used robots. They coped well with simple technological repeated operations: perform spot and arc welding, carried out loading and unloading, catered to the press and stamps. Robot "Yunimeyt", for example, was created to automate the resistance spot welding of bodies of cars, and the robot such as "the SMART" installed on the wheels of cars.
However, the impossibility of a stand-alone (without human intervention) the functioning of the first generation of robots very hindered their widespread introduction into production. Scientists and engineers persistently tried to eliminate this disadvantage. The result of their work was the creation of a much more complex adaptive robots of the second generation. A distinctive feature of these robots is that they can change their actions depending on the environment. So, if you change the parameters of the manipulation of the object (its angular orientation or location), as well as the environment (for example, the appearance of some obstacles in the way of arm movement), these robots can design your actions accordingly.
It is clear that by working in a changing environment, the robot must continually receive information about it, or he would not be able to navigate in the surrounding area. In this regard, adaptive robots have considerably more complex than the first-generation robots control system. This system is divided into two subsystems: 1) touch (or sensitization) - it includes those devices that collect information about the external environment and the location in space of the different parts of the robot; 2) a computer which analyzes the information and in accordance with a predetermined program, and it controls the movement of the robot and the manipulator.
To touch devices include tactile touch sensors, photometric sensors, ultrasonic, locating, and a variety of machine vision systems. The latter are particularly important. The main objective vision (actually "eyes" of the robot) is to transform the image of the environment of objects into an electrical signal, clear the computer. The general principle of vision systems is that by using a television camera into a computer to transmit information about the workspace. The computer compares it with the available memory in the "models" and selects the appropriate circumstances the program. In this way, one of the central issues in the creation of adaptive robots was to teach a machine to recognize patterns. Many of the objects the robot should highlight those that are necessary for it to perform some action. That is, he must be able to distinguish the signs of objects and classify objects according to these characteristics. This is due to the fact that the robot is in memory images of the prototypes necessary objects and compares them with those that fall within his field of vision. Usually, the problem of "recognition" of the desired object is divided into several simple tasks: robot searches the environment desired object by changing the orientation of his view, measures the range to the object of observation, automatically adjusts the sensitive video sensor to match the brightness of the subject, compares each object "model" stored in its memory, by several factors, i.e. allocates outline, texture, color and other characteristics. As a result of all this is a "recognition" of an object.
The next step of the adaptive robot are usually some action with this subject. A robot must approach it, grab and move to another place, and not haphazardly, but in a certain way. To perform all these complicated manipulations, some knowledge of the environment is not enough - the robot must precisely control every movement and how to "feel" themselves in the space. To this end, in addition to the sensory system that reflects the external environment, adaptive robot is equipped with a complex internal information system: internal sensors continuously transmit computer reports location of each manipulator link. They seem to give the car an "internal sense". As such internal sensors can be used, for example, high-precision potentiometers.
High-precision potentiometer is a device such as the well-known resistors, but characterized more precisely. It does not touch the rotating jumps from round to round, as in conventional offset handle rheostat, and follows the wire coils themselves. The potentiometer is mounted within the arm so that rotation of one link with respect to another movable contact is displaced too and, therefore, the resistance changes. Analyzing the value of change, the computer judges the location of each of the links of the manipulator. Movement speed of the manipulator is associated with the rotational speed of the drive motor. With all this information, the computer can measure the speed of movement of the manipulator and direct its movement.
How does a robot "plans" his behavior? In this capacity, there is nothing supernatural - "ingenuity" of the machine is entirely dependent on the complexity of the program prepared for it. In memory of a computer adaptive robot is usually incorporated as many different programs as may be of different situations. While the situation does not change, the robot operates on a basic program. When external sensors tell the computer about changing the situation, it analyzes it and choose the program that corresponds to a given situation. With the general program "behavior", the stock of programs for each individual situation, external information about the environment and internal information about the state of the manipulator, the computer directs all actions of the robot.
The first models of adaptive robots appeared virtually simultaneously with industrial robots. The prototype for them served automatically operating the manipulator, designed in 1961 by American engineer Ernst and later called "Ernst hand." This arm was gripping device equipped with various sensors - photovoltaic, and other tactile. With these sensors, as well as control computer, he found and took to him by randomly placed objects. In 1969, more sophisticated robot "neck" was created at Stanford University (USA). This machine also has the technical vision, be able to detect objects around and manipulate a given program.
The robot is driven by two stepper motors with independent drive to the wheels on each side of the trolley. In the upper part of the robot which can rotate about a vertical axis, are mounted a television camera and an optical rangefinder. The center has a control unit which distributes the commands from the computer to the mechanisms and devices that implement the appropriate actions. Along the perimeter of installed sensors to obtain information about the collision of the robot with obstacles. "Necks" can move along the shortest path to the specified location areas, thus calculating the path so as to avoid a collision (he perceived walls, doors, doorways). Computer because of its large size was separate from the robot. Communication between them is carried out by radio. The robot is able to select the desired objects and move them around "pushing" (arm he was not) in the right place.
Later, there were other models. For example, in 1977 the firm «Quasar Industries» Robot was created which was able to sweep floors, dust the furniture, work with a vacuum cleaner and remove the floor rasteksheysya water. In 1982, the company "Mitsubishi" announced the creation of a robot that was so clever that he could light a cigarette and lift the handset. But most remarkable was recognized as established in the same year the American robot which, through its mechanical fingers-eye camera and a computer-brain in less than four minutes, collecting the Rubik's cube. Serial production of the second generation of robots began in the late 70s. Most importantly, they can be successfully used in the assembly operations (for example, when assembling vacuum cleaners, alarm clocks and other simple household appliances) - this type of work is still very difficult resisted automation. Adaptive robots have become an important part of many flexible (agile on the issues of new products) automated production.
The third generation of robots - robots with artificial intelligence - yet only projected. Their main purpose - a purposeful behavior in a complex, poorly organized environment, moreover, in such circumstances, it is not possible to foresee all variants of its changes. Having some general problem, a robot should be to develop the program itself of its performance for each situation (we recall that the adaptive robot can only choose one of the proposed programs). If the operation failed, the robot with artificial intelligence will be able to analyze the failure, create a new program, and try again.