According to an article published by NewsUSA on the study conducted by the National Institute on Aging (NIA), more than one older adult falls each year and more than 80 percent of the falls happen in the bathroom. So there is a high need for systems that can detect fall and report it in real time especially inside bathrooms. All the systems that are currently available are either wearable devices or involve technologies that are either too expensive to install or intrude privacy. Also, dependency on wearable devices for automatic fall detection greatly limits the quality of life of older adults. In order to overcome these limitations and improve the quality of life of older adults, we present a new fall detection system that is wireless, cheap and efficient in detecting falls. The system we propose in this report is built using Micro-Doppler radar. We utilize the intensity captured by the Doppler sensor to determine the probability of fall. We developed a machine learning model that consumes encoding of captured intensities and determines the activity as fall and non fall. The model determines and reports a possible fall within 1 second. We also tested our encoding model approach on Android smartphone by capturing accelerometer and gyroscope data. The results obtained from both the experiments were very promising and encouraging.