Supportive smart home systems show the potential to enable older adults to age-in-place. However, research has not considered the communication challenge accompanied by wide-scale use. This thesis provides insight into supportive home systems' network traffic, identifies the impact of network impairments on a mechanism aimed to reduce network traffic, and develops a solution to ensure robustness of the traffic reduction mechanism to network impairments. Network traffic for two smart home systems and bed sensors was analyzed for 57 days. Results indicated a 10-fold difference in traffic between similar systems and the predominance of small packets which consume the network. Dual Machine Learning was implemented to reduce network traffic and, under simulated network impairments, yielded inaccuracies in cloud-recorded data. A solution was developed to mitigate the impact of network impairments, whereby accuracy increased from 71.4% to 94.6% for latency, 64.1% to 90.3% for jitter, and 61.6% to 78.9% for packet loss.