Formal Verification and Development of Living Assistance System

Formal verification and validation of computer software has always been beneficial for the end user. It helps to tackle the future errors in the system which may cause the system to perform inadequately. Currently, the world is moving towards systems which use machine learning and deep learning concepts. A lot of work regarding formal verification is still required in this aspect. Formal verification of systems involving deep learning networks is relatively new and requires a hand load of efforts. In this paper, we verify and validate a living assistance system which detect human activities (i.e Falling and Choking) through machine learning techniques and informs the end user. We used IAR VisualState and UPPAAL to perform verification and validation of this system. By formally verifying the system, we were able to ensure the correctness of our system; making it trustworthy and reliable to be deployed in a real world environment.