Car Occupancy Detection Using Ultra-Wideband Radar

Abstract

We present a data-driven car occupancy detection algorithm using ultra-wideband radar based on the ResNet architecture. The algorithm is trained on a dataset of channel impulse responses obtained from measurements at three different activity levels of the occupants (i.e. breathing, talking, moving). We compare the presented algorithm against a state-of-the-art car occupancy detection algorithm based on variational message passing (VMP). Our presented ResNet architecture is able to outperform the VMP algorithm in terms of the area under the receiver operating curve (AUC) at low signal-to-noise ratios (SNRs) for all three activity levels of the target. Specifically, for an SNR of −20 dB the VMP detector achieves an AUC of 0.87 while the ResNet architecture achieves an AUC of 0.91 if the target is sitting still and breathing naturally. The difference in performance for the other activities is similar. To facilitate the implementation in the onboard computer of a car we perform an ablation study to optimize the tradeoff between performance and computational complexity for several ResNet architectures. The dataset used to train and evaluate the algorithm is openly accessible. This facilitates an easy comparison in future works.

Variational Message Passing-Based Respiratory Motion Estimation and Detection Using Radar Signals

“…The threshold for all detectors was set such that a constant false alarm rate of p FA = 0.01 was achieved at each SNR value. As comparison, we evaluated the detection performance of an estimator-correlator, which models r as a Gaussian process to incorporate MPCs into the detection [4], and an FFT-based detector which computes the FFT over the rows of the matrix V R and compares the peak against a threshold. For the FFT-based approach the channel is transformed to the time domain to concentrate the signal energy in the delay bin corresponding to the LoS component for easier detection.…”

“…Future cars will be required to detect if they are occupied when the car is being locked to prevent the confinement of small children, which makes detecting the presence of people a safety-critical application [1][2][3]. The respiratory motion of the chest provides a dynamic feature used to separate the target's radar response from the strong clutter present in this use case [4]. We derive a Kronecker-factorized signal model for vital sign estimation in (strong) clutter using ultra-wideband (UWB) radar and present a variational message passing (VMP) algorithm [5,6] to detect the presence of people and estimate the respiratory chest motion.…”

“…However, in a tightly enclosed space, such as the interior of a car, the MPCs that interact with the target carry additional information and can be used to increase the signal-to-noise ratio (SNR), which is of critical importance in the given use case. In this work, we apply VMP to improve upon the results of [4], which already incorporates MPCs into the detection.…”