Performance Characterization of 3D MEMs Lidar with Nonlinear Scanning Pattern

The performance characterization of a 3D sensor from a relatively under-reported category of depth measurement sensors - a MEMs, laser light-based 3D sensor - is presented. A laser-light 3D MEMs sensor is distinct and different from canonical depth, ToF sensors or LIDARs but in terms of scan characteristics represents a confluence of all these types. Notably the intriguing feature of this sensor is the non-linear, non-monotonically varying scan pattern composed of overlapping curvilinear lines in the form of Lissajous trajectories such that the scan surface appears tesselated with variable-sized curved polygons, using a single laser diode. The sensing performance and merits of this type of scanning are analyzed and discussed. A noise model for the sensor jitter using simplified 3D Perlin noise is presented with justifications. Real data from several use cases and applications (obstacle detection, object segmentation, navigation and mapping) of the sensor are also provided. Considering criteria such as low noise levels, evenness of point distribution as well as higher scan density per unit area we find that this sensor provides superior performance in certain areas compared to other 3D sensors.