From air
Jump to navigation Jump to search


This project idea is to achieve autonomous scale model cars capable of traveling several laps of a set on a race. The route is signposted with panels between which the cars must pass avoiding obstacles and other competitors. The signs shall be identified by unique markers (QRCode type which are pinpointed). Obstacles can also be equipped with markers indicating a slowdown, circumvention by the right or the left. Cars already "know" the circuit (ie panels embodying the circuit). The track can be fitted with "radars" identifying the vehicles and computing their velocity.

More information on Project CannonBall de voitures autonomes



  • Tutors : Vivien Quema, Didier Donsez
  • Members : Jules LEGROS, Benoit PERRUCHE

Project objectives

Implement an autonomous car driving itself by identifying markers.

Progress of the project

The project started January 14th, 2013.

Week 1 (January 13th - Janurary 19th)

  • Project discovery
  • Discovery of OpenCV

Week 2 (January 20th - Janurary 27th)

Week 3 (January 28th - February 02nd)

  • We tried to run ArUco (markers recognition) with eclipse but we had to face a lot of problems on windows (libraries, path...)

Thus, we downloaded Visual Studio and we will work with this IDE.

  • Thought about how we will arrange the tablet on the car
  • We change the lenovo tablet for an icona 7 because of some issue with the power

Week 4 (February 03rd - February 09th)

  • Still some problems to run ArUco on Windows
  • Obtaining an USB hub
  • An other exchange of tablet, back to the thinkpad2

Week 5(February 10th - February 16th)

  • Installation of a rudimental workshop
  • Measuring different elements in order to build a box for the tablet
  • Drawing and thinking of plan for the tablet box

Week 6 (February 17th - February 23rd)

  • Finalisation of the plans, and correction of them
  • Implementation and test of basic movement of the motors

Week 7 (February 24th - march 2nd)

To Do

  • Get the movement sequences to apply in different cases for the remote controlled car.
  • Calibrate the camera

Last weeks

  • Abandon of trying to run ArUco on windows
  • Real time colored object tracking to replace the tracking of markers
  • Discovery of Arduino
  • Algorithms of movements for the car via Arduino.IDE (used for control of car motors)
  • Problems to catch the datas from the cam (position of the tracked object and its size) to the arduino (and the movement algorithms)

---> Solution : Serial communication

  • Adjustments (car acceleration / direction) with tests