Ventilation Estimation from a Chest Strap: A Case Study Limitations

Researchers’ attempts to estimate the exercise energy expenditure (EE) in more efficient and less cumbersome manner are still undergoing. The relation between the EE and the ventilation (V E ) was proved over the last decades; V E is now considered an image of EE. The estimation of V E is mainly based on magnetic sensor belts placed around the chest to measure the agitation of the rib cage and abdomen. The chest strap multi-sensor device, Zephyr BioHarness 3, integrates a pressure sensor that can measure the breathing movement by tracking the chest displacements during respiration. By this mean, the Zephyr device delivers, among others, a Breathing waveform (BW) signal and a Breathing Rate (BR) variable. In our previous health applications studies, mainly focusing on activity recognition, we assessed the robustness of the acceleration output given by this device. Now, in order to fully benefit from this promising module, our current study is dedicated to explore the BW and BR outputs, in an attempt to better quantify the activity by estimating the related V E variable. To do so, a subject was asked to perform a controlled-breathing protocol with different breathing rates and in three conditions of different strap tensions and positions on the body. Our validation experiment proved the reliability of the BR variable derived from the BW signal when validated against the output of the reference gas exchange system, the Cosmed K4B2. We however demonstrated that one can hardly derive the V E parameter directly from the BW measure, since i) it is not delivered in real physical units and ii) depends on how tight the device is fastened and iii) at which position the strap is adjusted.