Action representation for robots using MIRA
Bama Srinivasan, Ranjani Parthasarathi · 2017
Instructions given to a robot from the perspective of temporal logic have considered future states to activate movement and task execution. This paper looks at an alternate methodology using the formalism of MIRA (Mi̅ma̅ṁsa̅ Inspired Representation of Actions), where instructions to a robot are represented in a straight forward manner and the task execution is evaluated to one of the three values Satisfaction, Violation and Disinterest instead of true or false. Using this formalism, a framework is proposed and then compared with the existing methodology that uses temporal logic.