Algorithm 1: Gait cycle detection algorithm.

Input: Gait acceleration x-axis signal  Output: Start and end points of each gait cycle 1 Use Equation (1) to obtain all the minimal values of the input signal $P1$ 2 Use Equation (2) to initial screening $P1$ to obtain $P2$ after remove the pseudo-minimal values 3 Use Equation (3) to obtain the estimated step size of the gait signal L. 4 $\alpha =\frac{1}{4},\beta =\frac{3}{4},\gamma =\frac{1}{6},p=2$ 5 while $p<length\left(P2\right)$ do 6    if $position\_of\left(p\right)-position\_of(p-1)<\alpha L$ then 7    if $acceleration\_of\left(p\right)<acceleration\_of(p-1)$ then 8       remove(p) 9    else 10       remove($p-1$ 11    end 12    end 13    else if $position\_of\left(p\right)-position\_of(p-1)<\beta L$ then 14     if $position\_of(p+1)-position\_of\left(p\right)<\gamma L$ then 15      remove(p) 16     else 17      remove($p+1$) 18     end 19    end 20    $p=p+1$ 21 end