A 64-channel SPAD-based LiDAR sensor with Compact Coincidence Detection Circuit and Inter-frame Correlation-based Nonlinear Partial Histogram Builder

The SPAD-based Light Detection and Ranging (LiDAR) sensor [1–3] has found widespread applications across various fields, including consumer electronics, surveillance, and autonomous driving [4]. Leveraging collimated laser beams, line sensors are particularly suitable for outdoor applications due to its inherently higher optical signal-to-background ratio (SBR). However, detection range can still be limited in the presence of high background noise and constrained laser power budget. Macro pixels that incorporate coincidence detection circuits (CDC) aid in filtering events generated by background light through the utilization of temporal and spatial correlations among the detected signal photons [4]. Nonetheless, concurrently achieving compact coincidence detection, adjustable coincidence threshold settings, and PVT-insensitive correlation window generation remains a challenge using current methods [2, 4, 5]. To achieve high spatial resolution, scanning technique can be employed, albeit at the cost of potentially reducing the frame rate. In order to maintain an acceptable frame rate, the number of accumulated laser shots within each frame should be limited. In this scenario, the use of the traditional linear histogram builder method can decrease the SBR and consequently impact the detection range. To address these challenges, a 64-channel SPAD-based scanning LiDAR sensor is proposed. The main contributions of this work include: (1) Proposal of a compact and precise CDC for filtering out uncorrelated events. (2) Introduction of an inter-frame correlation-based nonlinear histogram builder and frame adjustment strategy intended to resolve the tradeoff between SBR and frame rate.