Laser Power Efficiency of Partial Histogram Direct Time-of-Flight LiDAR Sensors

On-chip capacity for storing temporal photon data in direct time-of-flight (dToF) lidar sensors is limited. This has prompted the development of various partial histogram approaches to reduce the amount of data stored on-chip. The aim of this paper is to inform sensor design by providing a taxonomy of these approaches, models for evaluating their impact on system laser power and identification of additional trade-offs which must be considered. All published on-chip partial histogram lidar approaches to-date are reviewed and two main categories are established: zooming and sliding. A means of evaluating any specific configuration based on its histogram reduction ratio (HRR) is also established. To quantitatively evaluate partial histogram approaches, a model to determine the minimum number of required laser cycles is developed. Both zooming and sliding are compared to an ideal baseline using this model, in order to establish a laser power penalty benchmark for each approach. These are evaluated over a range of real-world design conditions for two contrasting designs: short-range indoor and long-range outdoor. In general, a sliding approach is found to be the most laser power-efficient for long-range outdoor applications, while a zooming approach becomes increasingly more effective under low ambient conditions. Power efficient cycle-scaled variations on the conventional zooming and sliding approaches are introduced. These are shown to consistently reduce the laser power penalty across all tested design conditions. It is also shown that a cycle-scaled sliding histogram approach can be adopted to reduce the required on-chip histogram storage capacity by half, with almost no additional laser power penalty. Finally, a qualitative discussion of zooming and sliding compares additional key design considerations such as sensitivity to motion artefacts.