[[projects:projects]]:[[robots]] ====== Skateboard ====== ==== todo ==== Battery specs for nano Boot nano in embedded mode Run python program on boot Python webserver https://docs.python.org/3/library/http.server.html https://github.com/howCodeORG/Simple-Python-Web-Server/blob/master/serv.py Python Ajax server https://stackoverflow.com/questions/336866/how-to-implement-a-minimal-server-for-ajax-in-python How to get IP? Polling? hostname -i start webserver.py and ajaxserver.py simultaneously ==== State of the Art ==== As of June 2020, a youtube search reveals only one project involving a robotic or autonomous skateboard. * https://www.youtube.com/watch?v=LEmKBrcBGcM This manipulates weight distribution to "pump" a skateboard in a small half-pipe. No steering. ==== Physics ==== * steer - tilt the board right or left to twist the trucks * lean - the faster the speed, the more the rider must lean into the turn ==== Intelligence ==== * run a course - steer around traffic cones, accelerate to maximum possible speed ==== Vehicle design ==== * Jetson Nano mounted on skateboard * one electric hub motor on the skateboard for forward and backward acceleration * two linear servos distributing a weight side-to-side for steering * one downward-facing camera drone connected to the Jetson via wifi * batteries for the drone, the hub motor, the linear servos, and the Jetson * the Jetson plots the course and steers the skateboard and drone around the course * a web interface to the Jetson allows humans to stop and start the vehicle * a smart-phone wifi hotspot, the Jetson and drone connect to the hotspot Step 1. RC, forward, stop, right, left Drone * Step 1. Aerial photo. Identify cones, obstacles. * Step 2. Plot course. Options * Drone can fly over the course. * Drone only sends video and follows skateboard. * System can work with camera mounted on the skateboard. Works better airborne. electric skateboard that drives itself with weight-shifter for steering * and a drone with downward camera for vision * drone launches off the skateboard and follows the board * video from the drone transmits to computer and steering instructions then transmit to skateboard * software on the computer draws the scene, identifies cones and obstacles, draws lines, calculates angles step 1, use an off the rack jetson nano vehicle * but instead of a camera on the vehicle, use one from a drone * establish comm between vehicle and drone * build indoor courses with cones Buy * USB web cam * Battery for Jetson * Drone * Skateboard Photos of Cones and vehicle Detect with camera and Jetson Use drone camera handheld Camera feed from drone to Jetson Identify the course Reference each frame Draw angles among cones Plot course Jetson control drone Fly the course Keep overhead Jetson control skateboard Run the course Web interface to Jetson What is WiFi range? (Phone controlled drones use wifi) Cheap drones send camera feed to phone How can we get that feed to the Jetson? https://youtu.be/7ukDKVHnac4 Transistor