AI-Powered In-Vehicle Passenger Monitoring Using Low-Cost mm-Wave Radar

Abstract

We propose a novel algorithm to identify occupied seats, i.e., the number of occupants and their positions, using a frequency modulated continuous wave radar. Instead of using a high-resolution radar, which increases the cost and area, and performing complex signal processing with several variables to be tuned for each scenario, we integrate machine learning algorithms with a low-cost radar system. Based on heat maps obtained from the Capon beamformer, we train a machine classifier to predict the number of occupants and their positions in a vehicle. We follow two different classification methods: multiclass classification and binary classification. We compare three classifiers, support vector machine (SVM), K-Nearest Neighbors (KNN) and Random Forest (RF), in terms of accuracy and computational complexity for both testing and training sets. Our proposed system using an SVM classifier achieved an overall accuracy of 97% in finding the defined scenarios in both multiclass classification and binary classification methods. In addition, to show the effectiveness of our proposed in-vehicle occupancy detection method, we provide the results of a common group tracking and people counting method of occupancy detection. Compared to the common method, the effectiveness, robustness, and accuracy of our proposed in-vehicle occupancy detection method are shown.

Hallway Gait Monitoring Using Novel Radar Signal Processing and Unsupervised Learning

“…The number of chirps per frame 256 Bandwidth (MHz) The difference between the maximum and the minimum frequency a Capon beamformer algorithm to create a range-azimuth heatmap of the environment [42], [43]. This method not only paves the way for future versions of the algorithm that can track multiple subjects but provides more information on the environment to distinguish between the reflections from the walking subject and their multipath effects.…”

“…This method not only paves the way for future versions of the algorithm that can track multiple subjects but provides more information on the environment to distinguish between the reflections from the walking subject and their multipath effects. We refer the reader to our previous publications for details regarding how we applied the Capon beamformer algorithm to create rangeazimuth heatmaps of the environment [42]- [44].…”

AI-Powered Noncontact In-Home Gait Monitoring and Activity Recognition System Based on mm-Wave FMCW Radar and Cloud Computing

“…To overcome the dependency of the relative angle, one possible solution is to obtain the walking speed through the changes in a subject's position over time (i.e., velocity = position/time). A multiple inputs multiple outputs (MIMO) frequency modulated continuous wave (FMCW) radar [23], [33], [39]- [42] can provide the position of targets in addition to the micro-Doppler information, which makes it a good candidate for in-home gait monitoring assessment and activity recognition application [10].…”

Hallway Gait Monitoring System Using an In-Package Integrated Dielectric Lens Paired with a mm-Wave Radar

“…To use the same fabricated system for multiple people monitoring in our future work. Note that for multiple people monitoring, the information of all transmitters and receivers should be used, as demonstrated by our work in [ 28 ].…”