Adaptable language understanding: Can robots understand context bypass grammar rules?

Language is the only way for us to communicate with robots. It can be programming language, language of commands or ordinary speech. To achieve adaptable autonomous functioning of a robot while understanding commands within different contexts is strongly desirable. We suggest a biomimetic approach to the problem of context understanding evoking the model of animal behaviour within a social environment, where context plays crucial role. We propose a quasi-semantic model for a context which allows to take into the account words values as well as word frequency.