Skip to main content
Fig. 1 | Journal of NeuroEngineering and Rehabilitation

Fig. 1

From: Improving activity recognition using a wearable barometric pressure sensor in mobility-impaired stroke patients

Fig. 1

Block diagram of the activity recognition algorithm. Following the acquisition of the IMU and barometric pressure signals from the wearable device, the acquired signals are then preprocessed to extract key events (postural transitions, steps, lying periods). Then these events are combined into a hierarchical FIS to output the basic activities. The output of FIS II, i.e. the detected activities were fed into the decision tree for body elevation estimation and fed back into the FIS I for the detection of next activities

Back to article page