A testbed evaluation of MAC layer protocols for smart home remote monitoring of the elderly mobility pattern

Abstract

Unobtrusive data collection in a smart home environment can help with performing automated health monitoring and assessment. For instance, changes in mobility patterns which can be related to changes in cognitive abilities or an early sign of depression. In this paper, a system, using paired passive infrared (PIR) sensors at each entry of the house, observes the collective activities of a single resident individual, to model his or her mobility patterns. This system is designed in a manner that, while ensuring reliability and accuracy, reduces possible impacts on the individual’s privacy and does not add additional complications to the installation phase. Based on a mobility pattern data collection over a period of three weeks in our smart home testbed, analysis of wellknown medium access control (MAC) protocols applicable to the scenario under study is reported. While most of the researches have simulated the existing MAC protocols, we focus on using a real system and traffic pattern. In this research, we set up a testbed which employs PIR sensors to accurately record movements and the activities of our case study for 21 days. Furthermore, to emphasize on the inaccuracy of simulation analysis, we compare our collected data to three suitable MAC protocols, namely X-MAC, RI-MAC and A-MAC, with respect to duty cycle and packet delivery ratio. The acquired results indicate that, by applying real data, the performance of MAC layer protocols is similar. While the simulation analysis of A-MAC shows a higher duty cycle rate to RI-MAC and X-MAC protocols by 16.8 and 23.5 percent, respectively, our real data does not show a similar performance of the three MAC protocols. Additionally, we apply the well-known CASAS project dataset to run further analysis, which shows a similar output to our reported results.