RI: Medium: Collaborative Research: Novel microLIDAR Design and Sensing Algorithms for Flapping-Wing Micro-Aerial Vehicles

这个项目使一个微小的机器蜜蜂能够感知它与附近任何物体的距离。这种小型机器昆虫的深度传感技术在尺寸、重量、计算和功率方面突破了传感器和算法设计的极限。关键思想是关节设计；从机翼设计和光学到智能算法和高效计算，机器昆虫的每个部分都进行了优化。这可以通过来自不同背景的科学家和工程师的跨学科工作来实现。通过该项目吸取的经验教训可以应用于其他涉及小型设备的应用，包括医疗传感器和内窥镜成像、智能家居和物联网、农业和工业监控系统，以及用于搜索和救援的移动视觉。激光雷达传感使大型机器人汽车能够在复杂的环境中导航。该提案介绍了可用于昆虫级空中机器人的“微型激光雷达”的设计。让微型激光雷达在小型平台上工作涉及四个相互交织的研究重点。第一次推力使用MEMS反射镜和广角光学来感知和调制激光脉冲。第二个重点是适应信号处理算法来估计这种规模的距离数据。第三个重点是开发新的感知和导航算法，利用微型飞行器绘制室内环境地图。第四个重点是改进机器人昆虫飞行，使其能够进行新的操作，这些操作需要了解周围的范围图。这些传感器的效用将在机器人上进行演示，用于新颖的机动和构建室内环境的地形特征地图。

This project makes it possible for a tiny robotic bee to sense its distance to any nearby object. Such depth sensing for a small robot insect pushes the limits of sensor and algorithm design in terms size, weight, computing, and power. The key idea is joint design; every part of the robotic insect is optimized together, from wing design and optics to intelligent algorithms and efficient computation. This is possible by inter-disciplinary work across scientists and engineers from diverse backgrounds. The lessons learned through this project can be applied to transform other applications that involve small devices including medical sensors and endoscope imaging, smart homes and the internet of things, agricultural and industrial monitoring systems, and mobile vision for search and rescue.Lidar sensing has enabled large robotic cars to navigate complex environments. This proposal introduces designs for "micro-lidar" that can be used on insect-scale aerial robots. Making micro-lidar work on small platforms involves four intertwined research thrusts. The first thrust uses MEMS mirrors and wide-angle optics to sense and modulate the laser pulses. The second thrust is adapting signal processing algorithms to estimate range data at this scale. The third thrust is developing novel perception and navigation algorithms to map the indoor environments using a micro-aerial vehicle. The fourth thrust is to improve robotic insect flight to allow novel manipulations that require knowledge of the surrounding range map. The utility of these sensors will be demonstrated on the robobee for novel maneuvers and building topo-feature maps of indoor environments.