Solution Space Modeling for Robotic Systems

Abstract

Motivated from the experience of developing a lidar based vehicle tracking system, a model-based approach for specifying the solution space of robotic system is proposed in this paper. Solution space modeling can expand the design space, help finding the best possible solution, identify variation points, and also permit to perform context based run-time adaptation of the system. Models based on the proposed Solution Space Modeling Language (SSML) can help in system level reasoning, making tradeoffs, documenting decisions, and comparing them based on functional and non-functional properties, and can act as an artifact for formally proving and validating the final implementation.