Plenary - PRISM: checking the correctness of measurements for the Internet of things

Prism signal processing is a new technique for filtering and tracking real-time signals. It combines the benefits of finite impulse response (FIR) methods – linear phase and numerical stability – with the computational speed usually associated with recursive, infinite impulse response (IIR) methods. A further significant advantage is low design cost, so that it is possible to create new signal processing networks of arbitrary specification in real time, for example in response to changing conditions. The characteristics of Prism are well-suited to the problems of fault detection, condition monitoring, non-destructive testing, and equipment validation. This talk will give an outline of the latest developments in Prism Signal Processing, and give two examples of its application in research at SUSU: condition monitoring of rotating machinery, and fault detection in a pressure sensor.