Digital Position Control for Analog Servos

RC-servos are widely used to build robots with many degrees of freedom, because they are small, inexpensive, and easy to interface. Their main advantage, though, is their good weight-to-torque ratio. Many robots in the RoboCup Humanoid League, for example, almost entirely consist of servos. RC-servos come in two flavors: analog and digital. In analog servos, the internal controller is synchronized to the pulse train that encodes the target position. A typical problem of analog servos is that they suffer from large position tracking errors and variations of the zero-position which are caused by changes in temperature and/or supply voltage. To overcome these problems, digital servos have been introduced in the last years. Digital servos are, however, about twice as expensive as analog servos. Furthermore, all large servos are still analog. We constructed the legs of a 120cm humanoid robot from large analog servos. To improve position control, we modified the servos by replacing in the internal circuit the potentiometer with two equivalent resistors. The potentiometer is now interfaced to a microcontroller that measures position, implements digital position control, and sends motion commands to the analog servo controller. This approach makes it possible to use more advanced control techniques, compared to the original controller. For example, long-term position errors can be avoided by integrating the short-term error. Similar to other intelligent actuators, the parameters of the controller can be changed on a fast time scale, e.g. depending on the gait phase. Furthermore, the digital actuator is able to generate feedback for higher control levels. As the microcontroller is shared between multiple servos and was already in place prior to the modification, the additional hardware needed is negligible. We evaluate the proposed approach using systematic test signals. The results indicate that position control is more precise, more flexible, and more stable than before.