Contemporary patient surveillance systems have streamlined central surveillance into the electronic health record interface. They are able to process the sheer volume of patient data by adopting machine learning approaches. However, these systems are not suitable for implementation in many hospitals, mostly in developing countries, with limited human, financial, and technological resources. Through conducting thorough research on intensive care facilities, we designed a novel central patient monitoring system and in this paper, we describe the working prototype of our system. The proposed prototype comprises inexpensive peripherals and a simplistic user interface. Our central patient monitoring system implements Kernel-based On-line Anomaly Detection (KOAD) algorithm for emergency event signaling. By evaluating continuous patient data, we show that the system is able to detect critical events in real-time reliably and has a low false alarm rate.