Comparative evaluation of background-rejection techniques for SPAD-based LiDAR systems

Abstract

This paper presents the analysis and comparative evaluation of three background-rejection techniques implemented in CMOS processes for Single Photon Avalanche Diode (SPAD)-based Light Detection and Ranging (LiDAR) systems, using both Monte Carlo simulations and laboratory measurements. The first technique, known as photon coincidence technique, uses the temporal proximity of multiple detections to reject background light and maximize the detection of photons belonging to the reflected laser pulse. With the second technique, named Auto-Sensitivity (AS) technique, background light is rejected by automatically reducing the SPAD photon-detection probability (PDP), in order to avoid sensor saturation due to the intense environment illumination. The third technique we consider is the last-hit detection, which is able to detect and timestamp the last event impinging on the sensor over the acquisition window rather than the first, maximizing the system performance for long distance targets. The photon coincidence technique and AS technique are first compared to each other, showing the photon coincidence technique to outperform the AS technique. Then, the two techniques are applied together, resulting in an increase of the measurement range. Furthermore, a detailed analysis considering three different implementations of the photon coincidence technique is presented, showing pros and cons of each implementation and how the performance is affected. The last set of results focuses on the last-hit detection, which is compared against the standard detection paradigm (first-hit), showing not only an improvement for long distance targets (as expected), but an overall increase in system performance in terms of both success rate and SNR.