EMLL On-device​ Edge AI has the following advantages: Edge AI Challenges: ARM processors dominate smart devices and are the mainstream platform for edge AI. NPUs, DSPs, and GPUs offer higher computing power and have specific application scenarios for edge AI, but the ecosystem is limited and maturity is still a long way off. The most […]

Arm Performance Libraries When replacing Eigen’s backend with the Arm Performance Libraries (APL), which includes optimized linear algebra implementations such as Arm BLAS and LAPACK, the core idea is to have Eigen call the underlying linear algebra functions (such as matrix multiplication and decomposition) provided by APL, which are optimized for the Arm architecture, rather

NavRL Source: https://github.com/Zhefan-Xu/NavRL Welcome to the NavRL repository! This repository provides an implementation of the NavRL framework, which aims to enable robots to safely navigate dynamic environments using reinforcement learning. While the original paper focuses on drone navigation, NavRL can be extended to any robot with a velocity-based control system. For more details, see the related paper here:

DiffPhysDrone Source: https://github.com/HenryHuYu/DiffPhysDrone A swarm of drones can fly as fast as birds through uncharted forests, urban ruins, and even obstacle-filled indoor spaces, without relying on maps, communications, or expensive equipment. This vision has now become a reality! A research team from Shanghai Jiao Tong University proposed an end-to-end method that integrates drone physical modeling and

Swift Source: https://www.nature.com/articles/s41586-023-06419-4 Abstract​ First-person perspective (FPV) drone racing is a televised sport in which professional competitors pilot high-speed aircraft across a 3D track. Each pilot observes the surrounding environment from the drone’s perspective using video feeds from an onboard camera. Achieving the level of autonomous drone control achieved by professional pilots is challenging because the

PX4 HITL PX4+Gazebo Hardware-in-the-Loop Simulation HITL​ 1. Connect the flight controller to the computer via USB, open QGC, and select Vehicle Setup -> Security -> Hardware-in-the-Loop Simulation -> HILT enable. 2. Select Vehicle Setup -> Airframe -> HILT Quadcopter X (you will need to restart the flight controller for this to take effect). 3. Calibrate

HITL-PX4 Tutorial Hardware-in-the-loop simulation​ Hardware-in-the-loop (HITL or HIL) is a simulation mode where regular PX4 firmware runs on real flight controller hardware. The benefit of this approach is that most of the actual flight code can be tested on real hardware. PX4 supports both HITL (using jMAVSim or Gazebo Classic ) and VTOL (using Gazebo Classic) for multirotors.

Mujoco Bitcraze Crazyflie 2 Description​ [!IMPORTANT] Requires MuJoCo 2.2.2 or later. Changelog​ See CHANGELOG.md for a full history of changes. Overview​ This package contains a simplified robot description (MJCF) of the Crazyflie 2 model from Bitcraze. It is derived from the publicly available ROS description. URDF → MJCF Conversion​ License​ This model is released under an MIT License.

Pegasus Pegasus​ Overview​ Pegasus Simulator is a framework built on NVIDIA Omniverse and Isaac Sim. It aims to provide a simple yet powerful way to simulate the dynamics of multirotor aircraft. It provides a simulation interface for PX4 integration as well as a custom Python control interface. Currently, it only supports multirotor aircraft, but support for other aircraft topologies is planned

PX4 SITL Using the Gazebo​ First, we will build a simulated target using the console environment. This will allow us to verify the system setup before committing to the real hardware and IDE. Navigate to the PX4-Autopilot directory and start Gazebo SITL using the following command: If you have Gazebo Classic installed, start Gazebo Classic SITL using the following