Teach your Go2 to autonomously inspect a factory!

Join DroneBlocks’ Quadruped Expert and MIT Researcher Gabriel Margolis as he takes us on a deeper dive into robot simulation with the Unitree Go2 in our latest course! Gabe will also break down and explain the amazing technology behind the Go2’s onboard sensors – and explore some of the real-world applications of the Go2. This combined knowledge will bring you on a mission in the simulator to run an autonomous robot factory inspection! This course can be explored on its own – or as a follow-up to Gabe’s previous course on Go1 Low-Level Simulation – where students could gain familiarity with configuring a simulator for quadruped robots.

Although training the Go2 to shake its paw or rollover is exciting, some students may wonder: what is the practical use of quadruped robots such as the Unitree Go2? Gabe explains, “The Unitree Go2 was mainly designed for research and education purposes, but if you think of the scaled-up versions of the Go2 that are a bit more robust, such as Boston Dynamics’ Spot, these robots are serving the inspection industry, or serving in emergency response situations.” Tasks like these are where the advantage of using a robot becomes clearer because these situations are where a robot can excel. Using robotics can streamline the inspection process, covering large surface areas or environments, and is ideal for automation. We are also aware of the advantage of using a highly mobile robot in a dangerous emergency situation rather than putting a human in harm’s way. Other future applications are beginning to become unlocked as the technology proceeds, such as delivery or transportation of small payloads. It’s DroneBlocks’ hope that you or one of your students will be the one to unlock the next clever application of these robots after taking this course!

All of these robotic applications are made possible by the advanced hardware aboard these mobile-legged robots. In this course, we focus on the Go2’s upgraded sensor speeds and answer the question: What is LIDAR? Gabe breaks down how sensors work – and can be leveraged in useful ways in the real world – similar to how self-driving cars operate or how a camera captures images using sensors. The Go2’s multiple sensors, such as its RGB Camera and its spherical LIDAR sensor, are covered in depth. Through demonstrations in the simulator, students can view these sensors in action by simulating them and interpreting the data they absorb.

Finally, once students have built an understanding of the Go2’s sensors and configured the simulator – they will be able to test their new skills in a simulated factory environment provided by DroneBlocks. Students can send their virtual Go2 to explore this simulated factory inspection environment and create autonomous algorithms to solve the problem of inspecting the “Factory” without manually controlling the robot’s every move. Gabe explains the final task in short, “Students will simulate an inspection task – and consider how to make it an understandable problem for the robots to solve.”

As in all DroneBlocks courses, an expertly crafted step-by-step video guide will lead you on your journey, and downloadable exercise files are provided. It is also worth noting that a physical Unitree Go2 is not needed to benefit from this course!

In this course, students will:

• Setup the simulation environment using the provided exercise files
• Learn about LIDAR and Camera Sensors on Go2 and how they can be simulated
• Consider the realistic application of factory inspection
• Apply the application in the simulator 
• Learn to automate the inspection using an algorithm
• Establish a framework for a factory simulation with legged robots

In a natural progression, Gabe suggests that “in the next course, we will apply the same to the real Go2 in a physical environment.”

Visit the course HERE to get started! Interested in learning more about bringing the Unitree Go2 to your school or organization, learn more HERE