We met with one of the CEO of a mining company, who over couple of drinks, shared with us that there is sudden rise in attrition rate of his workers employed in factories in 10+ countries. He hired one of the management consulting firms to find out the root cause of this sudden attrition. Consulting firm did one month study, interviews with HR and workers and submitted the report that, Workers are taking early retirement/resignation due to serious health problems. Study suggested that, these workers were healthy when employed in health hazardous job roles like mining, drilling, etc., but over the period of time operating conditions started taking toll on workers’ health. Further study revealed that, this problem could have been avoided if management had taken timely intervention, like – Giving workers rest/leave as soon as signs of deteriorating health are hinted. It was not like that company management was ignorant towards workers health, but the actual issue was the payout! Most of the time, workers hide their falling health because they don’t want to be sent on forced leave or rest and loose on the payout (which is typically per hour). And, management had no mechanism to check if a worker who is doing his job is health or not? After listening to him, we told the CEO what if he has a way or solution which helps him to capture real time health vitals of workers so that he can have visibility on the healthy state of workers and can send workers on rest or leave so that they are not exposed to health hazardous operating conditions.
After listening to us, he finally said, Yes! This is amazing. Can you do it? And from there, we started developing a POC - “Connected Factory Workers” and we are submitting the same as part of AWS IOT Challenge.
What it does
Out IOT solution, collects real time data from the sensors embedded into the special suits and wearables which workers wears before they enter health hazardous zone. These sensors capture body temperature, heart rate, breathing rate, Cadence and Activity of each workers in real time. All these readings are then analysed to monitor health or workers and to find out who need to be send on rest/leave
How we built it
Since this is POC, we are not integrating any real sensors with AWS IOT. Instead, we are using on CSV file with sample health vitals which are uploaded to AWS IOT Things. From there, data is passed onwards to AWS Kinesis Analytics, AWS Kinesis firehose, AWS Analytics for processing and event generation in form of Alerts/Poly. AWS Machine learning is also leveraged for training the solution about optimal health vitals so that supervisor can know which worker is healthy and who is not. Real time dashboard is visible in our portal and as well as on QuickSigh dashboard.
Challenges we ran into
Even during the POC, we encountered many challenges like seamless flow of data between AWS components and configuring rules and actions. Although, our pitch looked easy, but managing the multiple stream of data and processing them real time was not easy.
Accomplishments that we're proud of
We are proud of our solution because with the help of this solution we would be able to generate human values out of our codes!
What we learned
During this POC we learned a lot about AWS IOT and Kinesis and how to manage multiple stream sof data.
What's next for Connected Factory Workers
We will take our POC ahead to the Pilot stage in which we will integrate real sensors to capture health vitals