Get real-time distance between running cars on Azure Stream Analytics - azure

Question

We will get streaming data of many cars on a particular Stream Analytics. Each row will have vehicleId, latitude and longitude of vehicle. I need to raise an alarm whenever distance between ANY two cars is less than suppose x meters.
Right now we can consider radial distance to keep it simple. Hence, we need to calculate distance of cars NOT from a fix point but from other cars ( near by cars can keep changing with time ). Hence, we cannot hard-code vehicle id in the query for sure.
We do have Geo-spatial functions support https://learn.microsoft.com/en-us/stream-analytics-query/geospatial-functions ..
I am not sure if this can even be done by Stream Analytics query directly.

Answer 1

I created a small example of the potential solution, not perfect one perhaps, but it resolves the problem in the ASA job.
Essentially I have re-used javascript function that expects simple latitude and longitude and gives the distance in meters. You can use potentially the geospatial embedded function - I haven't tried to play with that.
So, idea is to cross join input, for all input messages(unfortunately yes, you get duplicated result but it works), and then you apply distance function and filter to the output only those that have a distance less than a threshold value. The following example propagates to the output only if the distance is not zero(it means it compared with itself) and if it is less than 5 meters:
with inputData as (select * from input i1 inner join input i2 ON DATEDIFF(second,i1,i2) BETWEEN 0 AND 5),
distances as (select *, udf.getGeoDistance(i1.lat,i1.long,i2.lat,i2.long) as distance from inputData)
select *
into output
from distances
where distance <> 0 and distance < 5
UDF function:
// Sample UDF which returns sum of two values.
function getDistanceFromLatLonInKm(lat1, lon1, lat2, lon2) {
'use strict';
var R = 6371; // Radius of the earth in km
var dLat = deg2rad(lat2 - lat1); // deg2rad below
var dLon = deg2rad(lon2 - lon1);
var a =
Math.sin(dLat / 2) * Math.sin(dLat / 2) +
Math.cos(deg2rad(lat1)) * Math.cos(deg2rad(lat2)) *
Math.sin(dLon / 2) * Math.sin(dLon / 2)
;
var c = 2 * Math.atan2(Math.sqrt(a), Math.sqrt(1 - a));
var d = R * c * 1000; // Distance in m
return d;
}
function deg2rad(deg) {
return deg * (Math.PI / 180);
}
Input:
[
{
"name" : "car1",
"lat" : 59.3293371,
"long" : 13.4877472
},
{
"name" : "car2",
"lat" : 59.293371,
"long" : 13.2619422
},
{
"name" : "car3",
"lat" : 59.3293371,
"long" : 13.4877040
}
]
And result(car1 and car3 are close to each other):

Answer 2

Follow your latest comment,you use tumbling-window and set 5 seconds timeunit to get slice of data.Per my knowledge, you still could not calculate distance of cars each other by sql and Geo-spatial functions directly.Not to mention a warning.
I came up with an idea that you may could use Azure Function as the output of ASA job.Collect the slice of data and send them into Azure Function as json parameter.Inside function,you could write code to calculate the distances between cars each other,even alert warnings to other destinations.