The utility of walking parameters such as stride length, cadence and gait velocity for monitoring motor functions of patients suffering from brain injury, Parkinson's disease and obesity is well established. The application of sensor networks in this context has also been actively researched however; most of the research has focused either on construction of formal models of walking or design of wearable monitors. Unfortunately these approaches are not always practical for real-life monitoring, since they either require users to continuously wear monitoring equipment or rely on mathematical models which can be susceptible to significant prediction errors. In this paper we propose distributed mechanisms which utilize the concept of phenomenon detection and tracking for monitoring walking parameters. Our mechanisms do not require patients to be encumbered with monitoring devices and can track a subject's walk in real-time, in an energy efficient manner without a priori knowledge of a fixed mathematical model, thereby making it suitable for practical deployments.