Code Monkey home page Code Monkey logo

pren2's Introduction

PREN 2019 - Panzertraktor 37

banner-smoke

Build Status

Documents:

Project overview

The goal of the project was to develop a prototype of a train like vehicle that can drive on rails. Hardware and software was custom made for this project.

The prototype should be able to...

  • pick up a wooden cube from the ground next to the rails
  • drive 2 rounds on the track as fast as possible
  • recognize signals with differnet digits in different colors on them, that are located next to the rails
    • signals on a pole provide info about where to stop (only 1 exists)
    • signals close to the ground mark possible stopping locations (multiple signals with different digits)
  • stop at a location determined by the recognized digit on the signal

Facts & Numbers

At the final Presentation, our prototype achieved the following results:

Speed Info signal correctly detected Stop at the correct position Pick up Cube Ranking
1.9m/s Successful Failed Successful Rank 6 out of 27 teams

Our team consisted of 2 mechanical engineers, 2 electircal engineers and 2 software engineers.

Signal detection

Done by the software engineers: boardcomputer source code

The numbers on the signals and the special "start-signal" are detected with a custom trained neural network based on the YOLOv3 Tiny architecture.

An Intel Neural Compute Stick 2 was used to accelerate inference on the Raspberry Pi. With that we were able to process ~10 frames per second.

We detect the numbers 1 to 9 and the "start-signal". The position of the objects (upper or lower half of the frame) determines if the signal is an info- or a stop-signal.

predictions

Because the camera is able to take more than 10 images / second and we need to analyze as many images as possible due to the high speed, the images are spilt in half and 2 consecutive frames are merged into one image. Like this the amount of images processed is almost doubled!

A complete guide on how we did the signal detection can be found in here: https://github.com/eddex/tiny-yolov3-on-intel-neural-compute-stick-2

Picking up the cube

Done by the electrical and mechanical engineers: microcontroller source code

The cube is detected with a Time of Flight (ToF) sensor and then picked up by a magnet. The magnet is moved outwards on top of the wooden cube. It then attracts the iron hook on top of the wooden cube and is pulled back in again by the servo motor.

cube-pickup

pren2's People

Contributors

andilanz avatar eddex avatar

Stargazers

 avatar

Watchers

 avatar  avatar  avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.