**1. Introduction**

Despite the fact that computational technology continues to rapidly develop, edge devices and embedded systems are still limited in terms of their computation resources due to such factors as power consumption, physical size constraints, and manufacturing cost. This poses a challenge for critical applications such as mobile robots, cell phones, and AR and VR devices, which require efficient sensing with sensors and on-board computational resources. To effectively process the abundance of sensor measurements using resourceconstrained computation platforms, there is a need to limit the computation complexity of the methods deployed. This is true whether the method is data-driven or principle-driven, and high efficiency is typically a critical requirement.

This Special Issue is focused on both practical and theoretical technologies in the field of efficient intelligence and how they can be applied to diverse embedded devices such as industrial robots, unmanned vehicles, and fuel cells. The ten research papers published in this Special Issue cover a wide range of topics, including collaborative autonomous navigation with unmanned surface and aerial vehicles, multi-modal simultaneous localization and mapping (SLAM), target object tracking, LiDAR point cloud loop closure detection, motion distortion compensation for LiDAR point cloud, hybrid prognostic methods for proton-exchange-membrane fuel cells (PEMFC), detection of fabric defects during factory manufacturing, state recognition of elevator traction machines, efficient object detection neural networks, accurate pantograph detection for high-speed railways, and vision-based autonomous forklifts. It is our hope that these published papers will be beneficial for both academic researchers and relevant industrial practitioners alike.
