Simulation and sensitivity analysis of sensors network for cardiac monitoring

Yaël Kolasa (1, 2) , Thierry Bastogne (2, 1, 3) , Jean-Philippe Georges (2)

1 CYBERnano [Villers-lès-Nancy]
2 CRAN - Centre de Recherche en Automatique de Nancy
3 BIGS - Biology, genetics and statistics

This study's aim was to model and simulate a wireless sensor network and to propose a two-step sensitivity analysis for a targeted application related to home cardiac monitoring. After an initial phase of research to design the appropriate network simulator implemented in Omnet++, 13 simulation parameters have been selected to test their criticality. The sensitivity analysis relies on two consecutive steps carried out in Matlab: a screening phase (Plackett-Burman design) and a global sensitivity analysis (Space-fiiling design). Two output variables are considered: the number of packets received by the sink and the reception cache hit percentage. Four critical simulation parameters have been identified: the message length, bit rate, the background noise power and the energy detection of the radio receiver. In perspective, this sensitivity analysis will be included as a component of a Quality-by-Design approach of network development. This contribution is the early stage result obtained by Y. Kolasa during his MsC and PhD thesis (begun end of 2017) [2].

Mots clés: Simulation; Internet of Things; Sensitivity Anal- ysis; Cardiac Monitoring

Domaines: Réseaux et télécommunications [cs.NI]; Applications [stat.AP]; Cardiologie et système cardiovasculaire

Yaël Kolasa, Thierry Bastogne, Jean-Philippe Georges. Simulation and sensitivity analysis of sensors network for cardiac monitoring. 8th International Digital Health Conference, DH2018, Apr 2018, Lyon, France. ⟨hal-01925649⟩