Clever use of connected car data isn’t just a matter of driver convenience, comfort and efficiency, but also of safety, writes Wim Stoop, senior product marketing manager for EMEA at big data software company Cloudera.
As vehicles increasingly become laden with sensors and tracking tools, a wide range of businesses are facing a dilemma over how to manage and understand the barrage of data they amass. Many are under-prepared for this barrage.
Connected cars are swiftly becoming one of the leading and largest areas of the IoT. Market analyst company Gartner has projected that there will be 250 million connected vehicles on the roads worldwide by 2020, with those vehicles representing a combined market value of $150 billion.
Three quarters of new vehicles (75%) will be connected by 2020, according to another research company, TrendForce, and the number of sensors deployed within them is set to increase more than three-fold. It’s therefore clear that we are on the verge of every action that drivers perform being connected, tracked and instrumented.
Not only are manufacturers and the wider automotive industry seeing the possibilities of connected cars, consumers are increasingly keen, too. McKinsey Research finds that over a third of car drivers (37%) would switch to a different manufacturer if that was the only way they could guarantee having full access to all their apps, data and media. They’re willing to pay for connected services on a subscription-based model – a potential goldmine of new and so far untapped revenue for the industry.
However, connected cars are not just the latest cool fad enabled by the IoT revolution, they’re also making our lives much easier, more convenient and, most importantly, vastly safer.
More often than not, the benefits we associate with connected cars include cost reduction through better vehicle management or improved comfort through assisted or even self-driving capabilities.
But connected cars are also crucial in helping drivers reach their destination safely or in providing traffic assistance. This is now possible through sensors that warn drivers of external hazards, trigger hazard-warning signals and operate emergency call functions, in addition to more advanced capabilities such as self-parking, motorway assistance and even self-driving. The safety options also extend to predictive maintenance and providing diagnostics which will alert motorists when a fault may be about to occur on their vehicle – again, reducing the chance of a potential traffic accident.
In the US, the Department of Transportation and the National Highway Transportation Safety Agency estimate that sensors in connected cars could reduce non-alcohol-related traffic accidents by as much as 80 percent, preventing an estimated 5 million accidents and saving 18,000 lives each year.
In other words, this is not just a crucial technological opportunity for the automotive industry, but for society in general.
Another interesting element to this issue is that all this data being collected by these cars can be used by insurance companies, to advise them on motorists’ individual driving habits and their potential risk. This then further enhances public safety by advising drivers of traffic patterns, hazards and accidents in real-time.
A world of analytical opportunities
Google’s batch of connected cars has been estimated to generate a massive level of data collection that equates to 2 petabytes of data per car annually. This is because cars are now able to track all manner of indicators, from speed, RPM and fuel efficiency to temperature, pressure braking, location and safety data. However, Google’s level of data pales into insignificance compared to that available to Tesla, which is in a strong position to provide a much closer vision of fully autonomous vehicles.
There are two types of analytics required to work on the data being generated. It needs to be analyzed not only to make sense of it, but also to provide a spur to immediate action when it comes to ensuring hazard avoidance and enabling collision response. A lot of this data needs to be brought into the cloud or company data centers for trend analysis, to assess how vehicles perform over time and understand why collisions occur.
For this reason, the application of machine learning is vital. It is especially important when it comes to adding context; for example, combining traffic data with weather data to understand a situation and provide specific, correct recommendations to drivers – such as the need to replace brake discs.
The real value here lies in combining sensor data with understanding. Adding contextual data into anomalies helps organizations evaluate data and predict future events. Combining customer information such as demographics, location and behavior with social data allows manufacturers to make recommendations when a car requires an oil change or needs an update, as well as improving providers’ understanding of their customers to offer the services they desire.
This is where the real cleverness of customer insight comes in. By analyzing car and driver behavior, providers can now be certain which product the customer needs before they figure it out themselves. For example, if they drive on a lot of back roads, then the provider can recommend a set of tyres and tweak the suspension system to their individual requirements, to ensure they achieve optimal fuel efficiency.
Another great example of this working in practice is when a customer with brake problems is identified. It’s now possible to advise that the brake needs replacing, but then also take a 360-degree view, offering a coupon for the work, perhaps, but also helping the customer understand the performance issues they’re experiencing and offering a full solution to get the vehicle serviced.
Future vehicles in practice
As the number of signals and data points rises, it becomes increasingly vital for businesses to minimize vehicle downtime and ensure constant performance monitoring.
Manufacturer Navistar runs real-time visibility on its fleet of over 250,000 trucks, to increase fleet uptime and reduce overall maintenance cost per mile. By gathering telematics and geolocation data and then analyzing data in ways that weren’t possible before, Navistar has reduced unplanned downtime and maintenance costs by 40 percent, delivering tremendous return on investment.
Similarly, an insurance telematics provider has collected 130 billion miles’ worth of driving data from connected cars, then used it to analyze behavior, risks, speeding and braking patterns, crash information alongside contextual data like weather and traffic information. That provider might then build a risk profile for individual drivers based on driving habits, shared with insurance companies to help them provide individual offers to customers, reducing claims by 30 percent and saving millions of dollars annually.
It’s not 2020 yet and connected cars still have a long way to go, but there are some truly exciting innovations just around the corner. We’re already seeing how vital this innovation and advances in technology might be in helping to save lives and reduce accidents. As IoT and the sensors connected to devices and vehicles mature, data will be increasingly important to helping businesses understand and process this information in a manner that adds business value and is cost-effective.
By Wim Stoop, Senior Product Marketing Manager EMEA at Cloudera