Robots that fly ... and cooperate | Vijay Kumar

Good morning. Introduction to Autonomous Flying Robots

In this section, the speaker introduces the topic of autonomous flying robots and highlights the challenges and opportunities associated with building and applying this technology.

Challenges in Building Autonomous Flying Robots

• Autonomous flying robots are agile aerial robots that present unique challenges in their development.

• These robots are related to unmanned aerial vehicles (UAVs), but unlike large UAVs, they are small, lightweight, and capable of autonomous flight.

• The speaker showcases a small robot created by two students that weighs a little more than a tenth of a pound and is about eight inches in diameter.

How Autonomous Flying Robots Work

• The robot operates using four rotors that can be controlled independently to achieve different motions.

• By adjusting the speed of each rotor, the robot can hover, fly up or down, tilt, pitch forward or backward, and yaw about the vertical axis.

• An on-board processor analyzes sensor data from accelerometers and gyros to stabilize the robot's motion by sending commands to the motors 600 times per second.

Advantages of Small Scale Robotics

• When scaling down robotic systems like these flying robots, they naturally become more agile due to reduced inertia.

• As the characteristic length (R) decreases, inertia reduces dramatically, allowing for quicker turns and maneuvers.

• Videos demonstrate how these small robots can perform flips and recover quickly due to their agile nature.

Applications of Autonomous Flying Robots

This section explores various applications where autonomous flying robots can be utilized effectively.

First Responders and Safety Applications

• Autonomous flying robots can be deployed as first responders inside buildings for tasks such as intruder detection or detecting biochemical or gaseous leaks.

Construction Applications

• These robots have potential applications in construction where they can carry beams, columns, and assemble structures.

Cargo Transport

• Although small robots have limited payload capacity, multiple robots can work together to transport cargo efficiently.

• An experiment in Sendai after an earthquake demonstrated the use of these robots for assessing damage in collapsed buildings or mapping radiation levels in reactor buildings.

Navigation Challenges

• Autonomous flying robots face the challenge of planning trajectories from one point to another in a complex 12-dimensional space.

• To simplify navigation, a four-dimensional space is used, consisting of X, Y, Z coordinates, and yaw angle.

• The robot plans minimum-snap trajectories to optimize its path.

Conclusion

Autonomous flying robots offer exciting possibilities across various fields. Their agility and compact size make them suitable for tasks that were previously challenging or impossible for larger UAVs. By overcoming navigation challenges and leveraging their capabilities, these robots can contribute significantly to safety, construction, and transportation applications.

Smooth and Graceful Motion with Avoidance of Obstacles

In this section, the speaker discusses how minimum-snap trajectories are used to produce smooth and graceful motion while avoiding obstacles.

Minimum-Snap Trajectories in a 12-Dimensional Space

• Minimum-snap trajectories are transformed from a flat space into a complicated 12-dimensional space for robot control and execution.

Examples of Minimum-Snap Trajectories

• The first video demonstrates the robot moving from point A to point B through an intermediate point.

• The robot is capable of executing circular trajectories, experiencing up to two G's of force.

Obstacle Avoidance and Hoop Navigation

• Overhead motion capture cameras provide the robot with information about its position and the location of obstacles.

• The robot can navigate through moving obstacles, such as when Daniel throws a hoop into the air.

• The robot combines various trajectory segments to accomplish complex tasks, similar to a diver performing intricate movements.

Coordination of Multiple Robots Inspired by Nature

This section focuses on coordinating multiple robots using inspiration from nature, specifically studying coordination in Aphaenogaster desert ants.

Coordination Inspired by Ant Behavior

• Aphaenogaster desert ants exhibit implicit coordination without explicit communication or central coordination. They sense their neighbors and objects to achieve group coordination.

Decentralized Coordination in Robot Swarms

• Robots monitor separation between each other during formation flight using decentralized control commands calculated 100 times per second. Motor commands are executed 600 times per second based on local information from neighboring robots.

• Coordinated actions are achieved without centralized coordination, allowing for scalability and agnosticism towards neighbors.

Cooperative Flight and Construction with Robot Swarms

This section explores the capabilities of robot swarms in cooperative flight and construction tasks.

Formation Flight and Adaptive Formations

• Robot swarms can fly in formation, maintaining positions relative to their neighbors. Formations can be planar or three-dimensional, adapting to obstacles on the fly.

• Despite close proximity and aerodynamic interactions, stable flight is maintained during complex maneuvers.

Cooperative Object Manipulation

• Robot swarms can cooperatively pick up objects, increasing payload-carrying capacity by teaming up with neighboring robots. However, agility may be compromised as inertia increases with more robots carrying the same object.

Autonomous Construction

• Robots autonomously build cubic structures using truss-like elements based on an algorithm that guides their actions. The design blueprint is provided to the robots for autonomous construction.

Conclusion

The speaker discusses the use of minimum-snap trajectories for smooth motion and obstacle avoidance in robotics. They also explore coordination inspired by ant behavior for multi-robot systems and showcase cooperative flight and construction capabilities of robot swarms.

New Section

In this section, the speaker discusses how the robot navigates without GPS and instead uses sensors to build a map of its environment. The coordinate system is defined based on the robot's position and what it observes.

Robot Navigation Without GPS

• The robot is equipped with a camera and a laser rangefinder, which it uses to build a map of the environment.

• The map consists of features such as doorways, windows, people, and furniture.

• The robot determines its position with respect to these features, rather than using a global coordinate system.

• It navigates based on its understanding of these features.

New Section

In this section, the speaker shows a clip demonstrating how the robot enters a building for the first time and creates a map in real-time.

Real-Time Mapping

• Algorithms developed by Frank Shen and Professor Nathan Michael are used to create a map while the robot explores an unknown building.

• The robot identifies features, builds the map, and estimates its position 100 times per second.

• This allows for precise control algorithms to be applied in navigating the environment.

New Section

In this section, the speaker explains how the robot can autonomously explore an unknown building and provide information about its layout.

Autonomous Exploration

• The robot can be sent into an unknown building to create a map and gather information about its layout.

• It not only solves the problem of navigating from point A to point B but also determines where to go next based on areas with limited information.

• This autonomous exploration capability enables efficient mapping of unfamiliar environments.

New Section

In this section, the speaker mentions the potential applications of this technology, particularly in education and entertainment.

Applications in Education and Entertainment

• Robots like this have the potential to revolutionize K-12 education by changing the way we teach.

• The speaker emphasizes their location in Southern California, close to Los Angeles, and concludes with a focus on entertainment.

• A music video is introduced featuring nine autonomous robots playing six different instruments, created exclusively for TED 2012.

New Section

In this section, a music video featuring autonomous robots playing instruments is shown.

Music Video with Autonomous Robots

• The music video showcases nine robots playing six different instruments.

• The robots are completely autonomous and were created specifically for TED 2012.