DriveAwAI - Final Year Project - Nathan Lloyd

Welcome to my Final Year Project as an undergraduate, the aim of this project is to have a semi autonomous vehicle that naturally has the ability to toggle between assisted. This will be done utilising a Pi 4, Coral TPU and explorerHAT Pro.

Objectives

• Create a Raspberry Pi RC Car
• Provide a suitable GUI control
• Include autonomous driving elements, utilising computer vision
• Provide suitable security

Shopping List

Item	COST	Link
Raspberry Pi 4	£30 - 50	https://thepihut.com/products/raspberry-pi-4-model-b?variant=20064052740158
SD Card (Recommend 32GG+)	??	UP2U
ExplorerHat Pro	£20	https://thepihut.com/products/explorer-hat
Coral TPU	£75	https://www.mouser.co.uk/ProductDetail/Coral/G950-01456-01?qs=u16ybLDytRbcxxqFKdbhgQ%3D%3D&vip=1&gclid=Cj0KCQiAyp7yBRCwARIsABfQsnTXXBOMc39uSGN2DqE6U5CDuxEE6uXcuEYrENukWt5UyDaJZwTOfbQaAoTaEALw_wcB
PiCam V2	£25	https://thepihut.com/products/raspberry-pi-camera-module
PiCam Mount	£3	https://shop.pimoroni.com/products/raspberry-pi-camera-mount
Extra Webcam	£?	https://www.amazon.co.uk/TeckNet%C2%AE-1080P-Webcam-Built-Microphone/dp/B003VOEENG
Motors & Wheels	£8	2 PCS DC Gear Motor + Tire Wheel for Arduino DC 3V-6V
Fan Shim	£10	https://thepihut.com/products/fan-shim-for-raspberry-pi
Ninja Case	£8.50	https://thepihut.com/products/pibow-4-coupe-case-for-raspberry-pi-4b?variant=20452597661758
Dragging Wheel (Shooping Cart)	£3	https://www.amazon.co.uk/gp/product/B007OXBMQM/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1
Jumper Wires M-M	£5	https://www.amazon.co.uk/gp/product/B0144HG2RE/ref=ppx_yo_dt_b_asin_title_o01_s00?ie=UTF8&psc=1
LED's x2	£2.50	https://thepihut.com/products/adafruit-diffused-5mm-led-pack-5-leds-each-in-5-colors-25-pack-ada4203
Resistor x2	£6	https://thepihut.com/products/ultimate-resistor-kit
Blue Tape	£1	https://www.amazon.co.uk/gp/product/B0001IXA76/ref=ppx_yo_dt_b_asin_title_o02_s00?ie=UTF8&psc=1
Pi Cam Ribbon Ext	£4	https://www.amazon.co.uk/gp/product/B01MU2HAOM/ref=ppx_yo_dt_b_asin_title_o03_s00?ie=UTF8&psc=1
Toy Traffic Signs	£12	https://www.amazon.co.uk/gp/product/B07PD2M9C4/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1
Battery Pack	TBD	TBD
Webcam up2u	up2u
Total	£200 -£250	I've not gone out to look for the best deal, so please do that yourself.

Some of the above aren't exactly what I bought and in some cases I had parts laying around from previous projects. This is just a guide to what you may need if you fully recreate.

Car Body

I have had a lot of help from a friend designing a chassis to fit the items I've accumilated, you can use whatever you want for your body... but the CAD models and designs will be displayed within a folder shortly.

Extra Camera

As noted in my 2nd report, an Issue was running the Object detection model as well as onboard camera display using the same input. A simple workaround was using two cameras and mounting the second on top.

Installing Coral TPU

Step 1: Install Coral TPU - https://coral.ai/docs/accelerator/get-started/

echo "deb https://packages.cloud.google.com/apt coral-edgetpu-stable main" | sudo tee /etc/apt/sources.list.d/coral-edgetpu.list curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key add - sudo apt-get update

sudo apt-get install libedgetpu1-std

Step 2: Install TF for Python - https://www.tensorflow.org/lite/guide/python

pip3 install https://dl.google.com/coral/python/tflite_runtime-2.1.0.post1-cp37-cp37m-linux_armv7l.whl

Step 3: Check Installs by running a model

mkdir coral && cd coral git clone https://github.com/google-coral/tflite.git

cd tflite/python/examples/classification bash install_requirements.sh

python3 classify_image.py
--model models/mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite
--labels models/inat_bird_labels.txt
--input images/parrot.jpg

Installing Requirements

Step 1:

Enable i2c, SSH, Camera and VNC in Raspberry Pi config.

Step 2:

Download the project if you haven't already.

Step 3:

Unzip

Step 4:

Navigate to the directory (/DriveAwAI-Nathanlloyd7) through terminal, here you can find the requirements doc. Then Run
pip install --user --requirement Requirements.txt

Step 5:

Run main: default password on SenseHat PIN password = 4321. Have fun with your new car