Improving automated collection of Social Distancing data from camera feeds

The Covid-19 Pandemic has brought about an unprecedented lockdown to global society. Data on the impact and effectiveness of the lockdown is critically important for the following reasons:

* Epidemiology modelling needs to know the number of interactions that there are in society, and the change in these interactions following each intervention.
* The lockdown rules are new, unprecedented and hard to enforce. As a result, there is a lot of uncertainty around compliance, and the effectiveness of government messaging and enforcement strategies. Data is critical in determining what the next move should be.
* The police are under significant pressure, with many of their ranks in isolation. Being able to target lockdown enforcement to areas where there is a higher number of social interactions (<2m) is important.

Vivacity currently provides anonymous road usage data from our smart sensors. This is the only available multi-modal data source which provides accurate data on the volume of pedestrians, cyclists, cars and commercial vehicles across the UK. Our data is already forming a key part of the Governments analysis on the effectiveness of the lockdown.

In the first few weeks following the lockdown, we developed a new data output from the sensors, which measured the distance between pedestrians, and provided statistics on the number of <2m interactions. This current method does not give sufficient information to enable more detailed decisions to be made on the exact rules and enforcement of the lockdown.

In this project we will:

* Measurement of the duration of interactions - the longer people interact, the increased risk of infection
* Provide social distancing measurement for cyclists, and between cyclists and pedestrians, so that the government can understand which outdoor exercise modes are more effective for social distancing
* Look to classify the difference between "household groups" and interactions between "strangers", so that the data gathered is not used to discourage "household groups" who are currently allowed to go outside together from spending time together

We will apply these new analysis methods to our existing network of sensors to provide over 450 data feeds across 16 cities in the UK. This data will help the Government plan the removal of the lockdown measures, and monitor that social interactions do not grow too quickly as restrictions are lifted.