Tiny-PULP-Dronets：压缩神经网络以在多任务自主纳米无人机上实现更快、更轻量的推理
Tiny-PULP-Dronets: Squeezing Neural Networks for Faster and Lighter Inference on Multi-Tasking Autonomous Nano-Drones
AI总结：
本文提出 Tiny-PULP-Dronet 压缩方法，在保持 PULP-Dronet 飞行性能的同时将参数减少50倍、乘加运算减少27倍，为资源受限纳米无人机实现多任务智能提供轻量推理方案。
摘要：
口袋大小的自主纳米无人机凭借其极小的外形尺寸和几十克的重量，能够革新许多机器人应用场景，例如在狭窄、受限空间中的视觉检测，并确保更安全的人机交互。这一引人注目的愿景受到机载所需高智能水平的挑战，因为这种智能需求与可搭载于机上的 PULP（并行超低功耗）MCU 级导航和任务控制器所具备的有限计算和存储资源相冲突。本工作从 PULP-Dronet 出发，后者是一个用于纳米无人机自主导航的最先进卷积神经网络。我们提出 Tiny-PULP-Dronet：这是一种新颖的方法，可将运行推理所需的模型大小压缩超过一个数量级（参数减少 50 倍），并将运算数量减少 27 倍（乘加运算减少 27 倍），同时保持与 PULP-Dronet 相似的飞行性能。这种大幅缩减为纳米无人机上可负担的多任务处理铺平了道路，而多任务处理是实现高级智能的基本要求。
英文摘要：
Pocket-sized autonomous nano-drones can revolutionize many robotic use cases, such as visual inspection in narrow, constrained spaces, and ensure safer human-robot interaction due to their tiny form factor and weight -- i.e., tens of grams. This compelling vision is challenged by the high level of intelligence needed aboard, which clashes against the limited computational and storage resources available on PULP (parallel-ultra-low-power) MCU class navigation and mission controllers that can be hosted aboard. This work moves from PULP-Dronet, a State-of-the-Art convolutional neural network for autonomous navigation on nano-drones. We introduce Tiny-PULP-Dronet: a novel methodology to squeeze by more than one order of magnitude model size (50x fewer parameters), and number of operations (27x less multiply-and-accumulate) required to run inference with similar flight performance as PULP-Dronet. This massive reduction paves the way towards affordable multi-tasking on nano-drones, a fundamental requirement for achieving high-level intelligence.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
