**2. System Model**

In this work, a data communications and activity classification framework is presented, as shown in Figure 1. The proposed data communication and machine-learning-based activity classification framework lays out a communication infrastructure capable of sampling and relaying patients' sensory data (vitals + activity-related data such as accelerometer readings etc.) over the internet to accumulate data from virtually infinite number of patients. It also proposes a machine-learning-based activity classification framework to process the accumulated patients' data on the cloud and translate the sensory data into physical activities, thus maintaining the exercise/activity logs for each patient.

**Figure 1.** Post-surgery patient's sensory data communications and activity classification framework.
