This thesis examines the application of a 77 GHz Multiple-Input Multiple-Output (MIMO) Frequency Modulated Continuous Wave (FMCW) radar, which features 31 virtual receive elements enabled by Time Division Multiplexing (TDM) access, for non-invasive sensing in two key areas: vital signs assessment and Hand Gesture Recognition (HGR). The first section of this thesis targets the measurement of Breathing Rate (BR) and Heart Rate (HR) using Principal Component Analysis (PCA) applied to Range-Angle Maps (RAMs) to filter out noise from phase data over time. When tested against the EQ02+ LifeMonitor wearable device, this method yields a Mean Absolute Percentage Error (MAPE) of 1.06% for HR and 3.27% for BR, demonstrating its efficacy as a contactless health monitoring tool. The second section introduces a dual-stream neural network that integrates Convolutional Neural Networks (CNNs) and Gated Recurrent Units (GRUs) to analyze Range-Doppler Maps (RDMs) and RAMs for HGR. Evaluated through Leave-One-Out Cross-Validation (LOOCV), the CNN-GRU achieves a mean precision of 93.14% ± 5.50%, surpassing the performance of its Long Short-Term Memory (LSTM) counterpart, the CNN-LSTM, which has a mean precision of 90.90%±7.90%. Challenges arise with gestures such as Swipe Right (SR), Pinch Open (PO), and Pinch Close (PC), which indicate a reduced sensitivity of the radar to these specific gestures, thereby complicating classification. This is evidenced by a 8.33% misclassification rate from SR to Clockwise (CW) with the GRU. The study underscores the suitability of the developed algorithms for the recognition of vital parameters and gestures and thus forms a solid basis for the development of future radar-based systems for monitoring vital parameters (e.g. in motor vehicles) as well as combined systems for gesture and vital parameter recognition in healthcare.     «

This thesis examines the application of a 77 GHz Multiple-Input Multiple-Output (MIMO) Frequency Modulated Continuous Wave (FMCW) radar, which features 31 virtual receive elements enabled by Time Division Multiplexing (TDM) access, for non-invasive sensing in two key areas: vital signs assessment and Hand Gesture Recognition (HGR). The first section of this thesis targets the measurement of Breathing Rate (BR) and Heart Rate (HR) using Principal Component Analysis (PCA) applied to Range-Angle Ma...     »