During the one year study in Finland, all our teammates face the problem of falling down on the ice. So we decided to develop a solution by using Nokia 5G light pole to help reducing the number of falls in the city and save money for government in medical treatment area.

What it does

The Rescuer helps city managers to detect the icy area and save citizens from falling. Basically when passengers falling down, the 5g light pole records the accident and sends the data to Rescuer. On the user interface of light pole, city manager can easily see the amount of falls near the dedicated light pole, and also know whether there is a emergency by checking the status of light pole - indicated by color

How we built it

The project is setup with uMEC, LuxTurrim5G and Kubernetes. Each light pole has a uMEC and a raspberry Pi. Each Kubernetes Pod watches one light pole and handles traffic from the pole. If someone falls down near a pole, signals will be transmitted to the corresponding Pod before being sent to database. Signals are sent in UDP protocol. A website is available to read the data from database and visualise the data.

Challenges we ran into

Debugging the applications running in Kubernetes cluster was quite challenging. Pods would crash and this will drop all the logs after the Pod is deleted. Configuring the Kubernetes network to run on UDP and route traffic to a correct network is also challenging. IoT usually concerns huge number of components and cooperating those components not that easy even with the help of orchestration tools.

Accomplishments that we're proud of

We finished that and it works!!! We succeed in managing many components and we are proud of it. In the era of IoT, the number of devices will grow rapidly and traditional ways to manage those devices would become impossible. We are also proud that we developed an app that could save people's lives and reduce healthcare costs.

What we learned

All of our team members learned different techniques and enhanced their abilities. We learned how to use OpenCV, Socket, css, html, jquery, Ruuvitag and Kubernetes.

What's next for Rescuer

Using machine learning to detect whether the person is really falling down in the terms of accuracy.

Built With

Share this project: