Movement detection of people in indoor spaces with radar sensors

"Movement detection of people in indoor spaces with radar" Growing enthusiasm for the development of smart homes and intelligent building management systems highlights an important and developing market that focuses on boosting energy efficiency, strengthening security measures, and preserving privacy. This pattern underscores the urgent need to develop trustworthy methods for human detection and sensing. These technologies are essential for energy conservation because they automate and optimize the functioning of heating, ventilation, and air conditioning (HVAC) systems. They also greatly advance the development of security systems. These systems need to incorporate privacy-preserving methods to fully realize the benefits of automation and security within intelligent applications. Radar technology is crucial for maintaining human privacy in intelligent buildings. Radar systems differ from camera and microphone-based systems in that they utilize reflected electromagnetic waves to detect presence without the need to capture images or audio. This allows for strong security measures while maintaining privacy in an indoor setting. The objective of this thesis work is to investigate state-of-the-art sensor-based indoor human movement detection techniques and aim to implement a detection system with a focus on FMCW radar sensors. Detecting indoor human movement without relying on tags attached to objects or wearable devices presents a significant research challenge. The experimental portion of this thesis used a BGT60TR13C radar operating at a frequency of 60 GHz to locate and detect human movement. During the controlled indoor experiments, I performed two trajectories (linear and U-shape) in the radar's field of view. Gathered ground-truth data to assess the effectiveness of the suggested strategy. We used point cloud data and the Density-Based Spatial Clustering of Application with Noise (DBSCAN) method during the experimental process to improve robustness. The challenges we encounter in this work using FMCW radar sensing are the processing of radar-generated point clouds, which pose certain obstacles. The presence of excessive noise and clutter in these points of cloud data complicates the task of distinguishing between genuine and misleading targets. To address this problem, the DBSCAN algorithm and the modified threshold sensitivity can exclusively detect human movement while disregarding other immobile objects, including little insects, which are of no interest to us. This approach allows for clear differentiation between mobile individuals, stationary objects, and noises in the environment. After successful detection, we record object data into a comma-separated value (CSV) file to enable more thorough analysis later. This CSV file is the basis for computing performance measures like Mean Error and Root Mean Square Error (RMSE). The graphic visualization of both the measured values and the ground truth data establishes a reference system and improves clarity. The growing importance of human movement detection using FMCW radar offers numerous advantages. This thesis work employs a systematic methodology that aims to enhance the precision of human movement detection compared to conventional techniques. Smart buildings, security systems, and environments focusing on energy conservation could potentially use this system. The BGT60TR13C FMCW radar is used to detect motion and estimate distance, which makes it ideal for room presence detection.