You have time stamped performance indices of different machines on shop floor. These indices include $$noise index$=$ h$,$temperature$=$ t$,$vibrations$=$ g$,$performance$=$ r$.$ While modeling it as a multi$-$modal time series can you use CNN to detect future performance, breakdown probability $($at given time t$)$ and optimal maintenance schedule $($best next maintenance date$)?$ How? Model it as a constraint satisfaction problem $($CSP$)$ that handles time series data with time series forecasting methods. Provide diagram, pseudo code and highlight specific issues while detecting trends.