Abstract
The paper gives the answer to the question formulated in the title for the case of manipulator-level languages [1]. Theoretical considerations show that regardless of the type and number of robots and cooperating devices used in the system as well as irrespective of the number and kind of sensors included in the system, a general purpose language has to be extended by a single instruction with a rather complex semantics. Due to this complexity it is more convenient to introduce two, much simpler instructions. The paper also presents the method of implementing those instructions in a hierarchical distributed control system. The presented approach has been used in the implementation of MRROC/MRROC++ robot programming libraries/languages. The programmer uses predefined library modules to construct the controller structure solving a specific multi-robot task. The structure is fine-tuned to the task at hand by supplying adequate motion generators in a plug-in fashion. The practical validity of the formal approach followed in the implementation of MRROC and MRROC++ has been positively verified on diverse robotic tasks.

This publication has 11 references indexed in Scilit: