An Intelligent Optimization Strategy Based on Deep Reinforcement Learning for Step Counting

Table 1

Characteristic description of experimental data set.

Data set	Cambridge University data set
Acquisition frequency	100 Hz
Number of people collected	27
Sex ratio	2 (M) : 1 (F)
Age range	15–29
Height range	150–189 (cm)
Acquisition location	Hands, front pant pockets, back pant pockets, hand typing, handbags, and backpacks
Change in walking speed	Random walking ⟶ accelerating walking ⟶ decelerating walking
Data marker	Total number of steps and the start and end times of walking
Ratio of noise duration	47%