Autonomous Car - Neural Network / PID Controller

Description

1) MACHINE LEARNING: In this project we will create a NEURAL NETWORK with "Python" and copy its weights to a network with forward propagation on the "Artemis RedBoard ATP" board programmed with "Arduino IDE", and that will allow the robot car to drive alone and without hitting the walls. We will have four inputs, three correspond to 3 analog infrarred sensors and the fourth is for the BIAS. We will make the neural network have 4 outputs: two for each motor.

2) PID CONTROLLER: In this phase, also we will use the same 3 analog infrared sensors, which detect the distances at which the walls are, one in front and two on the left and right sides. The robot uses PID controller to maintain a central distance between the left and right walls. If the robot is near the left wall, then it can decrease the speed of the right motor and increase the speed of the left motor, to make the robot move to the right, and moving away from the left wall, and vice versa.

Details

The goals of this project are the following:

Apply the machine learning to a robot car.
Add the PID controller to this robot.

To facilitate the understanding of this project, we have divided it into the following sections:

HARDWARE,
MACHINE PEARNING,
PID CONTROLLER,
PRINTING AND ASSEMBLING THE CHASSIS,
TEST, and
CONCLUSION.
REPOSITORIES ON GITHUB

You can see my complete project at: https://www.hackster.io/guillengap/self-driving-car-using-redboard-artemis-atp-521c63

Files

chassis_up.stl

Secondary STL file of the chassis

Standard Tesselated Geometry - 11.61 kB - 05/19/2020 at 20:24

Download

chassis_1.stl

Main STL file of the chassis

Standard Tesselated Geometry - 32.41 kB - 05/19/2020 at 20:22

Download

microrobot_artemis.png

Schematic Diagram

Portable Network Graphics (PNG) - 1.08 MB - 05/19/2020 at 20:03

Preview

Components

1 × SparkFunn RedBoard Artimis ATP

1 × L298N Driver

1 × Rechargeable battery 7.2 V

3 × IR distance sensor

2 × Gearmotor

Project Logs

Collapse

HARDWARE

Guillermo Perez Guillen • 08/23/2020 at 01:24 • 0 comments

In the figure below I show you the schematic diagram.

SparkFun RedBoard Artemis

Features: Arduino Uno R3 Footprint, 1M Flash / 384k RAM, 48MHz / 96MHz turbo available, 24 GPIO - all interrupt capable, 21 PWM channels, Built in BLE radio, 10 ADC channels with 14-bit precision, 2 UARTs, 6 I2C buses, 4 SPI buses, PDM Interface, I2S Interface and Qwiic Connector.

Build Instructions

Collapse

Software apps and online services:

Arduino IDE
Python 3.7.3

Hand tools and fabrication machines:

3D Printer (generic)

Discussions

Jon wrote 09/03/2020 at 14:19

I was interested to read about your experience with ML on the Artemis platform. Although not in the scope for our current project ( https://hackaday.io/project/174098-lighting-color-control-with-commodity-lamps ), one of our team members purchased the Artemis board. There are definitely some significant implications for ML as it relates to our own proposed design that could take the project in some future iteration from "this thing we could program with some rules" to "this amazing device that could solve for scenarios we've not even imagined." Good luck!

Are you sure? yes | no

Guillermo Perez Guillen wrote 09/03/2020 at 16:57

Hi Jon, good luck to your project ... next year I will post some ML with Artemis and using voice command ... the issue is that I've to spend weeks to get only two words voice commands with a prediction of 70-80% for example, it's not so easy, you should have a lot of patience ... Here is something that can help you so much: https://www.hackster.io/javagoza/voice-controlled-faucet-09e066

Are you sure? yes | no

Guillermo Perez Guillen wrote 08/25/2020 at 02:52

This project addresses the challenge of "Wildcard", and complies with what is indicated on the official website: https://hackaday.io/contest/171491-supplyframe-designlab-2020-hackaday-prize/log/177551-challenges

Are you sure? yes | no

Guillermo Perez Guillen wrote 08/25/2020 at 02:51

This project complies with "The 2020 Hackaday Prize Official Contest Rules": https://hackaday.io/contest/171491-supplyframe-designlab-2020-hackaday-prize/log/177665-the-2020-hackaday-prize-official-contest-rules

Are you sure? yes | no

Autonomous Car - Neural Network / PID Controller

Description

Details