Location Intelligence, Spatial Data Analysis | Pitney Bowes

Location Intelligence: Using APIs to Avoid Accidents in Driverless Cars

driverless car is capable of sensing its environment and navigating without human input by using radar, GPS and computer vision. Advanced control systems interpret sensory information to identify appropriate navigation paths.

 

But even considering all the technological acumen that driverless cars possess, accidents can still happen. A report on accident rate of self-driving cars gives surprising numbers[1].

 

A recent accident involving a Tesla model S[2] has grabbed a lot of media attention.

Although these self-driving cars have sensors and detect obstacles ranging from pedestrians to other cars, the same cannot be expected of traditional cars without such sensors. This makes it necessary to create better and safer driving zones for these vehicles.

How can we create a safer environment for self-driving cars? Data! Actionable insights can be made using accurate data derived from both locations and sensors. Data from traffic light sensors can be used alongside location intelligence to determine the population density in an area. While real-time demographic data is almost impossible to obtain, sensor data for a given timestamp tied with demographic data obtained from Location Intelligence APIs can be used to determine the feasibility of a driverless car in a given zone. Data from the ongoing traffic of driverless cars can be used to further develop a better and more efficient system.

 

A Use Case in this aspect might be: When a road is identified, the data supplied through the Pitney Bowes Geo Life API would give the driverless car a context of that particular throughway. The following image shows a demonstration.

This insight, coupled with the data obtained from street light sensors, will offer enough information to provide a recommendation – whether or not the car should be put into autopilot. An approach like the one described can give better context to the driver for making a conscious decision on autopilot mode given the road demographics.

Further, the Geo Enhance API can give a context of the location-based places of interest. So, auto pilot mode can be used with caution when driving near a school or a hospital. An example of this is given below.

It is predicted that the number of self-driving cars on the road will rise significantly in the years to come[3].

In a perfect world, all the cars on any given road, at any given time, would communicate with each other through sensors, forming a fully operational Internet-of-Things (cars), which would remove the need for any use of traffic information. But until then, we need better policies and safer zones, which can be achieved through intelligence supplied by crucial traffic data concerning both pedestrians and vehicles.

References

[1]https://www.theguardian.com/technology/2016/jan/12/google-self-driving-cars-mistakes-data- reports

[2]http://www.huffingtonpost.com/entry/tesla-crash-autopilot-mode_us_577c2b55e4b0416464110cbc

[3]  http://www.businessinsider.com/report-10-million-self-driving-cars-will-be-on-the-road-by-2020-2015-5-6