Page 292 - Concise Encyclopedia of Robotics
P. 292

Scaling
                            to achieve data communication between two locations that are nearly at
                            opposite points on the surface of Earth. Station X can be considered the
                            location of the control operator, and station Y the location of a remotely
                            controlled robot.
                              One of the biggest challenges facing researchers in artificial intelligence
                            (AI) is how to link computers that are separated by vast distances. There
                            is no way to overcome the fact that the speed of light is slow on a large
                            scale, and when considered in terms of the time required for a computer
                            to execute a clock cycle.
                              See also MICROWAVE DATA TRANSMISSION.
                         SCALING
                            Scaling is a principle familiar to structural engineers and physicists. As
                            an object is made larger to an equal extent in all linear dimensions, its
                            structural integrity diminishes.
                              When things get larger but stay in the same relative proportions, me-
                            chanical strength increases as to the square (second power) of linear di-
                            mension—height, width, or depth. However, mass increases according to
                            the cube (third power) of the linear dimension.The illustration shows how
                            this works with cubes. The mass, and thus the weight in a constant gravita-
                            tional field, goes up faster than the linear dimension or cross-sectional
                            area increases. Eventually, if an object gets large enough, it becomes physi-
                            cally unstable or mechanically unworkable.
                              Consider a theoretical solid cube of variable size but perfectly homo-
                            geneous matter. In the illustration, the smaller cube has height = 1 unit,
                            width = 1 unit,and depth = 1 unit.The larger cube is double this size in each
                            linear dimension: height = 2 units, width = 2 units, and depth = 2 units.
                            The base (or cross-sectional) area of the smaller cube is 1 unit squared
                            (1   1); the volume of the smaller cube is 1 unit cubed (1   1   1). The
                            base (or cross-sectional) area of the larger cube is 4 units squared (2   2);
                            the volume is 8 units cubed (2   2   2). If the cubes are made of the
                            same homogeneous material, doubling the linear dimension also doubles
                            the weight per unit surface area at the base. As the cube keeps getting
                            larger, it will eventually fall through or sink into the surface, or collapse
                            under its own weight.
                              Imagine the situation with a humanoid robot. If its height suddenly
                            increases by a factor of 10, its cross-sectional structural area increases by
                                                                   3
                                      2
                            a factor of 10 = 100. However, its mass becomes 10 = 1000 times as
                            great. That is the equivalent of a 10-fold increase in gravitational acceler-
                            ation. A robot constructed of ordinary materials would have difficulty
                            maneuvering under these conditions, and would be unstable. Another



                                                   
   287   288   289   290   291   292   293   294   295   296   297