Algorithm 2 Data Reduction Algorithm Based on Adaptive Reduction Threshold

Input:  current threshold $e_{max}$, sensor reading $z_k$, data trends $d_{k-1}$, data cache cachedval   1:while True do   2:    if k = 1 then   3:        Insert $z_k$ into cachedval   4:        Send $z_k$ to remote processor   5:    else   6:        Insert $z_k$ into cachedval   7:        Calculate $d_k$   8:        Call Algorithm 1 to calculate adaptive reduction threshold   9:        if $e_{max}$ changes then 10:           send $d_k$ to edge server 11:        Call the Kalman filter to calculate estimated value $x_k$ 12:        $e_k$ = $z_k$ − $x_k$ 13:        if |$e_k$| > $e_{max}$ then 14:           send $z_k$ to remote processor