Piezoelectric MEMS Mirror Optimized by Particle Swarm Optimization Algorithm

Micro-mirror which is capable of steering light at a reasonably high speed, is an important component in MEMS solid-state LiDAR systems. Longer detection range and larger field of view (FOV) are often ideal in many applications, such as autonomous driving, and those aspects can be achieved by increasing the mechanical angle of the micro-mirror as well as the size of the aperture. However, as the aperture and rotational angle (theta(opt)center dot D) get bigger, the dynamic deformation inevitably becomes larger, thus affecting the collimation performance. One potential solution is to add a backside rib support to the mirror which can reduce the dynamic deformation while keeping its moment of inertia low. Conventional backside rib designs are primarily based on intuitive structural patterns, and the design process is time-consuming. Also, the performance improvement is based on trial and error which does not guarantee success in the end. To shed light on an optimized pattern with the focus of large theta(opt)center dot D and low dynamic deformation, in this paper, we propose a piezoelectrically-driven micro-mirror with an optimized backside rib enabled by a particle swarm optimization (PSO) algorithm and iterative FEA modeling. Experimental results show that compared with an intuitive pattern, the automatically-generated pattern can reduce the beam divergence by 30% while keeping the same moment of inertia.