Checksum based error detection in linearized representations of non linear control systems

Abstract

While the past decades have seen major revolutions in the computing and communication industries, the next decade will experience a proliferation of intelligent autonomous systems capable of performing complex tasks. These systems will consist of a plethora of sensors communicating with each other to achieve specified system level objectives. The success and deployment of these systems in commercial arena depends on the reliability and security of the system design. In this work, we discuss error detection methodologies that can detect sensor failures and actuator malfunctions in a linearized representation of a non-linear control system. We demonstrate the proposed approaches on the linearized representation of the classical non-linear problem of an inverted pendulum which has a linear Kalman filter as its sensor and a linear servo motor as its actuator. Simulation results show the effectiveness of the proposed methodology.