Description: Quadruped technologies are booming worldwide with expected global terrain robot market to rise to 1.43 billion in 2030. Recently, Boston Dynamics, Xiaomi, Unitree, and others have introduced their product to the market. It should be emphasized that the available technologies have an extremely limited autonomy and AI algorithms that considerably shortens their application in industry. Despite a huge request from industry in Russia there is no ready-developed product on the market. We propose to develop an AI-driven quadruped robot platform, which is able to recognize the environment and terrain conditions based on deep learning and computer vision algorithms in different lighting environments. With this ability the robot will easily navigate to hard-to-reach places avoiding obstacles and memorizing the environment, and, eventually, investigate mining and construction areas and check measurement device data and thereby securing operation safety in such industries as oil and gas, mining, nuclear stations, construction, and warehouses. The project will focus on the quadruped robot design, BLDC motor control, MPC control of robot, simulation of locomotion in NVIDIA Isaac Sim, and DL-based navigation. 

Project status: under development.