by Raman Mehta

How data science can turn the vision of connected vehicles into reality

Aug 10, 2015
AnalyticsBig DataData and Information Security

Data-science-driven predictive and prescriptive analytics can transform connected vehicles to improve safety and mobility, provide new 'infotainment' options, and result in a cleaner environment.

As per NHTSA statistics, more than 32,000 people lost their lives in the United States in 2013 in road accidents. There is no better use for technology than saving lives. Connected vehicles represent a seismic movement that is ready for prime time. It is at an inflection point where automobiles, telemetry, infrastructure, technology and most importantly the mind set change are converging to make connected vehicles a reality.

The term “connected vehicles” holistically refers to vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) connectivity. Safety benefits of connected vehicles include blind spot detection, pedestrian warning, and collision avoidance. Imagine when the infrastructure interacts with the vehicle about upcoming school zones, construction areas, dangerous curves, black ice conditions and immediately alerts the driver. In one of the NHTSA studies, up to 80 percent of accidents involving non-impaired drivers can be avoided using connected vehicle technologies.

The goal of connected vehicles is to improve safety, mobility, and environmental impact; provide new ways for “infotainment;” and a slew of other business benefits. Making drivers safer, greener and smarter, and providing a way of naturally extending our connected lives through vehicles is one of the biggest application of Internet of things (IoT). With every model year change over, more and more vehicles are being fitted with sensors, telematics and connectivity solutions.

This is all great but the infrastructure, regulations and above all massive data required to make this a reality are still evolving. The automotive companies will struggle from the variety and unstructured nature of the data. The variety, velocity and volume of the data would make it unsuitable for traditional database processing. We need more sophisticated big data platforms, such as Hadoop, to process this massive data.

Auto companies have been capturing telematics and sensor data for a number of years. They just did not know how to fully store, leverage and monetize it. With the emerging big data and analytics, it is now possible to crunch this massive amount of data in real time and make it available to provide an actionable alert through vehicle infotainment systems. The Automotive IoT driven (pun intended) advances in sensor data collection and analytics based on sound data science principles is the game changer.

Bits to Information

The true value of the big data is in creating unique customers insights. It is about consuming inbound feeds from telematics, sensors, infrastructure and the environment, and then creating outbound information, alerts and marketing feeds using the right channels appropriate times. The insurance companies can change the consumer mindset from “big brother” is watching to how they can educate drivers with big data insights on their driving habits and reduce premiums for safe drivers.

The advantage of connected vehicles powered by analytics is that now we can extend connections to traffic lights, highway sensors, tunnels and bridges. This tremendous amount of data can be coupled with information gathered from weather services and other sources to provide a true picture of the road conditions, safety hazards and traffic congestion at an individual vehicle level.

The context and reliability of alerts is very important. There is a real danger that due to lack of rigor in applying data analytics consumer may get flooded with multitude of meaningless alerts. Think of transmitting an alert when there is a sudden increase in hard braking events from vehicles in close proximity to each other. It is useful if it happens on a small stretch of a highway during an unexpected whiteout condition. This type of alerts could avoid chain accidents. Whereas, the same braking event happening when a large numbers of automobiles exit a sporting venue in a crowded downtown after a football game may not be that alarming. We need sophisticated data-science-driven predictive and prescriptive models that can separate the noise from real signals.

Another big challenge is information security. The recent Chrysler Jeep hack that went viral is a good example of the uncharted territory we are entering. In this case hackers took control of an on-road Jeep and manipulated its air-conditioning and speed among other things. Think of this scenario happening on a massive scale where traffic signals are hacked into, or tunnels and bridges are maliciously jammed by sending false signals to vehicles. It could easily turn into a homeland security issue. Automakers who do not pay strategic attention to cyber security are risking brand dilution due to direct impact on reliability, safety and potential compromise with the personal information of their customers.

Other industries such as publishing and retail have undergone complete transformation due to the Internet. The automotive industry has taken advantage of the efficiencies due to the Internet, such as global product design and supply chain improvements, but the connected vehicle is the one thing that will complete the transformation journey.

Innovative manufacturers are moving away from just “selling cars.” They are moving towards providing mobility solutions. A solution that involves an ecosystem of partners in which infrastructure providers, mobile data companies, repair shops, insurance providers and a host of other partners maximize the value to the customer.