Development and Validation of a Spike Detection and Classification Algorithm Aimed at Implementation on Hardware Devices

Algorithm 1

Pseudocode for AdaBandFlt and Limada.

AdaBandFlt

number_of_samples=0;

number_of_windows=0;

F; // definition of sample frequency //

while number_of_windows < 100 do  O( 𝟏𝟎𝟎 )

   while number_of_samples < F*0.01 // define a 10 ms window // O(F* 𝟎.𝟎𝟏 )

      take a sample

      number_of_samples ← number_of_samples + 1;

   end

   compute the RMS and put it in a vector named “error”   O( 𝟏 )

   number_of_windows ← number_of_windows + 1;

end

BubbleSort(error) // sort the vector of 100 RMS // O( 𝟏𝟎𝟎²)

define the threshold initial value (25th percentile)  O( 𝟏 )

Limada

number_of_samples=0;

number_of_windows=0;

threshold_init=0;

F; // definition of sample frequency //

ok=false;

while ok == false do //ok is false until 100 clean windows are collected // O( 𝐍 )

   while number_of_samples < F*0.01 // define a 10 ms window //  O(F* 𝟎.𝟎𝟏 )

      take a sample

      number_of_samples ← number_of_samples + 1;

   end

   BubbleSort(samples) //sort the voltage values of a window //  O( 𝟏𝟎𝟎²)

   compute the 2nd (V _.02) and the 30th (V _.30) percentiles    O( 𝟏 )

   if number_of_windows < 100    O( 𝐍 )

      test= V _.02/V _.30;    O( 𝟏 )

      if test < 5 & |𝑉_.30| is significantly non zero O( 𝟏 )

         threshold_init ← 0.99* threshold_init + 0.01*|𝑉_.30|

         number_of_windows ← number_of_windows + 1;

      end

   else ok==true; // 100 clean windows are collected //

   end

end