Published May 27, 2021 © Apache-2.0

Climate - control - AI - remote - vision- telehealth

Indoor/Outdoor climate control concepts using AI /ML. Eliminate travel requirements using tele - commute and telehealth. Funny App !!

IntermediateProtipOver 1 day573

Climate - control - AI - remote - vision- telehealth

Things used in this project

Hardware components

QuickLogic Corp. QuickFeather Dev Kit with UART Cable + SensiML

In addition to board -- connector , debugger ..etc..

SmartThings Shield for Arduino

Smart items --- added to make it IOT compatible for remote access...

Software apps and online services

Microsoft windows as starting point

Public domain ubantu linux download. Partially use google cloud with linux features in developing and testing perl/python code. Some Visual studio debug/run python.

Hand tools and fabrication machines

Digilent Mastech MS8217 Autorange Digital Multimeter

Multiple tools are needed -- from digikey and microchip, quick logic .. attached later.. Started to include linux as a second OS to Windows. Hard to swich and not sure of comming mack to full functional Windows. Bought few webcams . May need dedicated laptop attached to hardware.

Story

tele health smart AI project

-------------------------------------------

Principle : Reduce travel, emissions.

The main objective of the project is to have a

tele health system. The whole system is too big to

handle by one system. Need to have echo system of

products.

The initial focus is on using core AI models from

model library (zoo) and test it on simple images. The

open system is supported by openvino on windows systems.

This require many inputs and outputs analyzed and tested.

Windows system require to pre install many tools. These

are described in code/schematic section. Simple models used

are face recognition, vehicle detection, vehicle parts, type

and number plate read tests.

Typical system involves webcam as input device collecting

video frames at remote location connected to internet. The

image is processed/inferred and results are provided to host

locations.

Use the smart windows technology. The temperature inside

building and outside are different at different times of the day.

Smart way of saving energy is by managing heat transfer inside

and out through smart fans. The smart fan can be prgrammed by

remote application based on time of the day, temperature difference.

In the simplest form smart programmable plugs are used.

Diagrams will be added later... at different sections...

Focussed project objective to programmable features in tele-health.

The body sensor patch is hard to re-design and use it for different purpose of bio sensing. Wired sensors are not convenient to keep it attached to body for hours. Unhealthy air due to fire and smoke is dangerous to live and breath.

Programmable sensors are the solutions. Program the IOs to the required applications. FPGA. QuickFeather QORC type eFPGA development board.

..........

Schematics

Code

References and backgrounds:
  
https://github.com/adventuresinML/adventures-in-ml-code

https://github.com/krishnajy/colab/commits?author=krishnajy


at Cloud:

https://console.cloud.google.com/home/dashboard?walkthrough_id=go_gae_quickstart&project=nth-glider-290301&cloudshell=true


...

Welcome to Cloud Shell! Type "help" to get started.
Your Cloud Platform project in this session is set to nth-glider-290301.
Use “gcloud config set project [PROJECT_ID]” to change to a different project.
krishnajy@cloudshell:~ (nth-glider-290301)$
krishnajy@cloudshell:~ (nth-glider-290301)$ pwd
/home/krishnajy
krishnajy@cloudshell:~ (nth-glider-290301)$
krishnajy@cloudshell:~ (nth-glider-290301)$
...

https://github.com/Xilinx/Vitis-AI

....

The laptop used to test code crashed away..
Used another laptop and github/cloud to work
out the details. 

...

from tensorflow.keras.applications.inception_v3 import decode_predictions
# Predicted class index.
idx = np.argsort(res[0])[-1]
top3_pred = decode_predictions(res, top=3)[0]

print('Top 3 Predictions:')
for pred in top3_pred:
    print(f"{pred[1]}: {pred[2]*100 :.2f} %")
    
#Display the result on screen
from IPython.display import display 
pred_class = top3_pred[0][1]
pred_prob = top3_pred[0][2]
print(f"Prediction: {pred_class}: {pred_prob*100:.2f} %")
display(Image.open(inp_file_name))
-------
Top 3 Predictions:
golden_retriever: 95.70 %
Labrador_retriever: 1.40 %
kuvasz: 0.58 %
Prediction: golden_retriever: 95.70 %

------
; tele - AI vision 
:
similar to flower  detection

sample Python code : detection

==========

from tensorflow.keras.applications.inception_v3 import decode_predictions
# Predicted class index.
idx = np.argsort(res[0])[-1]
top3_pred = decode_predictions(res, top=3)[0]

print('Top 3 Predictions:')
for pred in top3_pred:
    print(f"{pred[1]}: {pred[2]*100 :.2f} %")
    
#Display the result on screen
from IPython.display import display 
pred_class = top3_pred[0][1]
pred_prob = top3_pred[0][2]
print(f"Prediction: {pred_class}: {pred_prob*100:.2f} %")
display(Image.open(inp_file_name))

==========
Top 3 Predictions:
golden_retriever: 95.70 %
Labrador_retriever: 1.40 %
kuvasz: 0.58 %
Prediction: golden_retriever: 95.70 %

Credits

Krishna P Rao

3 projects • 2 followers

Embedded system design, 5G , wireless, chip design Masters degree

Climate - control - AI - remote - vision- telehealth

Things used in this project

Hardware components

Software apps and online services

Hand tools and fabrication machines

Story

Custom parts and enclosures

Smart HOME

Use windows terminal using mobx kind

fpga use to programmable IOs --etc..

Schematics

Real front page

Day/night timer set for auto on/off

Linux platform source ..

some steps convenient/minor details..

AI- Tensorflow

AI-remote-vision-detection

smart window weather

Code

The code will be at GITHUB.

Weather smart control , tele-health , remote AI applications.

AI -- Tensorflow

AI code in python tensorflow

AI/ML/Python -- cloud applications

Credits

Krishna P Rao

Comments

Embed the widget on your own site

Climate - control - AI - remote - vision- telehealth

Climate - control - AI - remote - vision- telehealth

Things used in this project

Hardware components

Software apps and online services

Hand tools and fabrication machines

Story

Custom parts and enclosures

Smart HOME

Use windows terminal using mobx kind

fpga use to programmable IOs --etc..

Schematics

Real front page

Day/night timer set for auto on/off

Linux platform source ..

some steps convenient/minor details..

AI- Tensorflow

AI-remote-vision-detection

smart window weather

Code

The code will be at GITHUB.

Weather smart control , tele-health , remote AI applications.

AI -- Tensorflow

AI code in python tensorflow

AI/ML/Python -- cloud applications

Credits

Krishna P Rao

Comments

Related channels and tags