Bipede robot : from an open loop to a closed loop control system





The behaviour of a bipede walking robot is simulated here with a program dedicated to this particular analysis. The moving picture above was built with the program that you can download here, rename the file with the extension .arj instead of .gz, uncompress it with ARJ.EXE. Written in C language, this program runs in DOS mode. My interest in the walking robot pushed me to developp this program. For now, the motion of the robot is only a simulation. It reacts only to values that have been computed, in order to reproduce a correct walk. The robot actually reproduces motions which have been stored in files. However, the gravitation is also simulated, so that the recorded values for the trajectories have to be properly made, otherwise, you will attend to the fall of the robot. The ultimate goal is to create a robot that computes itself the values for the trajectories. In this way, there are two mains things to observe :
-The aim of the robot (go ahead, turn left, right, stop, go back, etc...)
-The perfect control of the balance.


To understand physically these elementary aspects, it is necessary to study control system. What is control system ? Basically, control system is dedicated to the optimisation of the command for technical plants. The user of a plant applies an order. With an adapted control system, the order doesn't affect the plant directly. There is a corrector which interprets the difference between the order and the value measured with a captor (connected to the output of the plant). The tuning of a corrector is quite hard to solve. It is necessary to compute transfert functions, which depend on the number of parameters in the plant. For a walking robot, the tranfert function is not easy to calculate.

But before to manage these problems, there are mechanical aspects which have to be studied.
First of all, the efforts supported by the actuators are of two kinds :
-Statics efforts (without motion, the actuators works to maintain the static balance of the robot).
-The dynamical efforts (quite strong, especially in running mode).

Once the efforts have been evaluated, it becomes possible to compute the parameters for the actuators of the robot.