In this project, we designed a system for the detection of children forgotten in cars. This system uses measurements from a piezoelectric sensor installed in the car’s seat. The generated signals are processed by a machine learning algorithm which includes the following blocks: preprocessing, feature extraction, classification and result filtering. This algorithm achieves performance of a single in more than 10,000 signal instances, on our acquired dataset. We implemented the above algorithm on a real time micro-controller, to produce a fully operational standalone device.