Duckietown

🚀 The new Duckiedrone (DD24) is now shipping, and is available worldwide! 🛸 Autonomous Drones are going to be a significant part of our future. But how do they work? Get your own Duckiedrone now and start learning autonomy at: https://hubs.la/Q02Gl9Sw0 🤖#robotics #AI #drones #autonomousvehicles #learningautonomy #engineeringeducation 📚

The 6th Annual MassRobotics #Duckiedrone Academy is in full swing! Autonomous Drones are finding increasing applications in the real world and have the potential to become game changers in industries such as agriculture, maintenance, safety and security, logistics, etc. At Duckietown, we believe it is key to provide accessible, state-of-the-art learning experiences in robot autonomy to everyone, as today’s youth will be shaping the next generations of robot autonomy. Big shout out to MassRobotics (Kevin Smith, Joyce Sidopoulos for organizing, Brown University (Stefanie Tellex) and Amazon Fulfillment Technologies & Robotics for supporting this Duckietown #Duckiedrone enabled summer camp! Learn more about Duckietown at: duckietown.com and about the Duckiedrones: docs: https://lnkd.in/dV62UGzk #stem #robotics #engineeringeducation #vocationaltraining #drones #autonomousvehicles #AVs #learningautonomy #quadcopters #AI #coding

Ever been in a rush, but that (self-driving) car in front of you is taking it particularly easy? Knowing when and how to safely pass a moving obstacle can be quite the challenge for autonomous vehicles. Former Duckietown students Nikolaj Witting, Fidel Esquivel Estay, Johannes Lienhart, and Paula Wulkop, developed and implemented a solution that enables Duckiebots to detect and maneuver around dynamic as well as static obstacles, while maintaining awareness of oncoming traffic. (Yes, those are two way streets!) Dynamic obstacle avoidance was achieved by enhancing the Duckietown visual lane-following pipeline to detect and navigate around static and moving obstacles. The approach followed was to estimate the pose relative to other Duckiebots based on LED detections, and integrating a dynamic controller node to manage overtaking maneuvers. Want to try and reproduce this at home or for your class? You can start here: https://hubs.la/Q02FPjqN0 #engineeringeducation #robotics #engineering #highereducation #self-drivingcars #math #computervision #robotautonomy #diy #autonomousvehicles #agv #AVs

We started our journey at the Massachusetts Institute of Technology in 2016 to make the science and technology of robot autonomy accessible and fun by open-sourcing the first Duckiebot and Duckietown class. We are glad to see that this effort contributed to a shift in the trend of published papers in the field of vehicle autonomy! “Towards Autonomous Driving with Small-Scale Cars: A Survey of Recent Development” by Dianzhao L., Paul Auerbach, and Ostap Okhrin is a thorough review that highlights the rapid development of the industry and the important contributions of small-scale car platforms to robot autonomy research. This survey is a valuable resource for anyone looking to get their bearings in the landscape of autonomous driving research. Congratulations to all involved in the continuing dissemination of robot autonomy literacy, we are excited to see what the future of self-driving cars and embodied AI holds! https://hubs.la/Q02FrlMk0 #S-Bot, #Kobot, #e-puck (EPFL) #Turtlebot (Open Robotics, Clearpath Robotics by Rockwell Automation), Thymio ll, #RACECAR, #GRITSBot, #BARC (University of California, Berkeley), #Autorally (Georgia Institute of Technology), #Deepracer (Amazon Web Services (AWS)), #DonkeyCar, #MuSHR, F1Tenth Foundation, #JetBot (NVIDIA Robotics), #JetRacer, #XTENTH-Car, #ART/ATK

🤖How can we extract maximum information from heterogeneous data streams, e.g., vision and encoders in an autonomous car?🚗 The job of the estimation block in an agent is to provide the robot with an educated guess of its relation to the environment (position and orientation), transforming the raw data from sensors into actionable information for the robot to accomplish its task. Like a detective piecing together the versions of different eyewitnesses to reconstruct the underlying truth, sensor fusion comes to the rescue! 🕵️♂️🔍 Learn more about sensor fusion and robot autonomy with Duckietown at: https://hubs.la/Q02F0yNF0 #SensorFusion #RobotAutonomy #AutonomousVehicles #AI #Robotics #EngineeringEducation

Prof. Marek Długosz shares his experience teaching robot autonomy at the AGH University of Krakow. Read the full interview here: https://hubs.la/Q02DyttR0

From 3 to 4 wheels, check out an Ackermann Steering Duckiebot! ● Realistic car dynamics ● One DC motor, one servo ● Increase in mathematical modeling complexity Do you think it is worth increasing the vehicle complexity in learning robot autonomy? Merlin Hosner and Rafael Fröhlich, former Duckietown students thought it was and not only built this amazing prototype, but produced one of our favorite videos ever! Learn more about their project: https://hubs.la/Q02CdGxz0 Follow Duckietown for more ideas on reproducible robotics projects. #DIY #engineeringeducation #robotics #engineering #highereducation #selfdrivingcars #math #computervision #python #robotautonomy

This paper describes the application of vision-based reinforcement learning for lane-tracking control, using a #Duckiebot. Robert Moni and team developed a controller policy using convolutional networks trained with proximal policy optimization and domain randomization. Sim2Real transfer learning was validated by metrics measured on a real track and compared with simulation outcomes. Read more about their research - https://hubs.ly/Q02BW_yJ0 and check out the comments for additional resources to reproduce these results.

Researchers from the Department of Computer Science and Engineering of the Qatar University in Doha, Qatar, investigate the potential security risks associated with end-to-end deep reinforcement learning (DRL) systems in autonomous driving environments that rely on visual input for vehicle control. Authors: Abdullah Hosseini, Saeid Al-Houti and Junaid Qadir. Read more about their study and conclusions: https://hubs.ly/Q02zXynC0

BCI Initiative thank you for giving us the opportunity to support the dissemination of your amazing work! We look forward to continue making #Duckietown available to support your efforts, and providing curious people worldwide with an accessible platform for learning the science and technology of robot autonomy. Congratulations to the winning team!