Power of Prediction: Developing Context-Aware Cars

The connected car revolution is gearing up, transforming a conveyance that gets you from point A to point B into a powerful personalized device that keeps you seamlessly connected from any location and at any time and predicts your wants and needs along the way.

An altogether new type of car will soon hit our streets: Cars that connect with consumers’ digital lives. Cars that get better and better the more they are driven. Cars that become increasingly intelligent as they learn and adapt to the users’ personality over time. These cars will recommend likely destinations, maneuvers or routes based on the user’s previous actions and current situation. Like a good butler, they will anticipate the driver’s needs, including making recommendations for places to visit along the route.

Car manufacturers have become good at gathering and storing data. Now it’s all about leveraging this wealth of data and making use of it. To develop personalized, contextually relevant applications and meet next generation customer expectations, OEMs have to put data into the overall context of driving and making it intelligent. Clearly, data and intelligence are not the same thing. Automakers need to work out algorithms that reduce noise as they define and extract truly valuable information.

[click to continue]

Data gathered by the Connected Car has limited value if not combined with context – Tweet That!

Raw data is not actionable or useful, but it does become useful by adding context. In order to accurately summarize and represent the data that has been collected, layers of driver-dependent context, or personalization, should be applied. These in turn feeds the user experiences that drivers see and feel when driving the vehicle. Contextualization and personalization is the key to transforming data into real information and delivering personalized context-aware experiences and superior user experiences. With a contextual understanding of the driver, the car can offer a huge choice of applications that anticipate the driver’s wants and needs and thus make the overall experience of driving the vehicle more interesting, safer and more compelling. To master contextualization, demographic, historical and contextual data has to be analyzed:

  • Contextual data refers not only to geo-location but also to road conditions caused by traffic, weather, time of day/week/year or other factors
  • Historical data will contain a person’s visited places, prior routes and actions.

CloudMade is devoted to bringing contextual relevance to drivers. Our products are designed to give both drivers and OEMs actionable insights, thus creating a better experience for drivers and generating significant business insights for OEMs. CloudMade understands the context of the driver, using machine learning against a multitude of content sources like traffic, current and predicted weather, news, events, POIs and more. Cars with CloudMade’s technology are smarter and feel particularly personal to their driver.

Learn more about CloudMade’s context awareness engine: Predictive Learning System

November 12th, 2014 - Posted by in for OEMs

Tech Talk Video with CloudMade’s CTO

In this short video, CloudMade’s CTO, Jim Brown, talks about CloudMade’s technologies and how they are used to create better user experiences in automotive, outdoors, fitness and wearable products.

The technologies Jim talked about can be found here:

  • Hybrid – aggregation, sync and management of location content
  • Mapsafe  – cloud-based personal data and preference storage, sync, access and backup
  • Micromap – a compressed, routable, searchable geodata format
  • UMDb – a tool chain that imports data from numerous formats into Micromap

To learn more about CloudMade’s technologies and for a demonstration of their capabilities, get in touch.

[Show previous posts]

October 7th, 2013 - Posted by Nick Black in for OEMs, geodata, products, talks, technologies

Single Line Search in Navigation

CloudMade’s On-the-Go Search

We’re re-shaping search-on-the-go at CloudMade, making search faster, easier and more relevant while in the car read on to find out more

In this post we look at how CloudMade’s On-Dash and In-Dash platforms use our Hybrid Data and Mapsafe technologies to let users search through point of interest data from many different providers and see the results represented as a single item.  For some real-world examples of this in action, check out Magellan’s SmartGPS product that lets users search through Yelp, Foursquare and TomTom data from a single interface.

Global Search

Users want access to a rich world of location information from a single search interface.  After all, Google lets us search the entire content of the web from a single interface, so why should we need to go into separate “apps” when searching for a place to stop for lunch?

When the user searches for “Pizza”, search results are shown, sorted by distance, from different providers on a single screen.  In this example you can see results that appear both in Yelp and Foursquare, results that only appear in Yelp and results that only appear in TomTom data – indicated by the icons at the far right of each returned result.

Search

CloudMade’s Mapsafe technology provides social context to search results, in this case associating each place with check-ins, reviews and posts made by the user’s friends on social networks like Facebook, Twitter and Foursquare.

Disambiguated Results

The next screens show how the user is able to view the details associated with the results.  Different data providers have different attributes and strengths.  Yelp for example includes a 1-5 star rating, Foursquare data is typically updated the most frequently, Tom-Tom often has the most accurate positional data.

CloudMade’s Hybrid Data technology handles a process of disambiguation which takes unique results from three (or more) datasets, for example a pizza place in Yelp, Foursquare and TeleAtlas and understands through a process of fuzzy matching that the three separate entities in-fact refer to the same real-world location and presents a single, disambiguated result to the user.  This gives you the best of all worlds: the user gets the latest, richest information from sources like Yelp and Foursquare whilst you can rely on the positional accuracy of a provider like TeleAtlas.

Search-Disambiguation

A note on merging and disambiguation – at no point does CloudMade’s platform merge results together.  The different databases (Yelp, Foursquare and TomTom in this case) are kept as separate databases with the user being able to search from a single interface, see all results on a single page and quickly filter through different providers.

Live Results

Where devices have an internet connection, through a Bluetooth tether for example, live results from an online search provider like Google Places can be fed into the search results.

Search-Live
Discussion

The capabilities of CloudMade’s On-Dash and In-Dash platforms to handle large volumes of rapidly changing geo-data from a vast number of different sources and make the geo-data available to OEMs through a set of cross platform APIs opens up a new range of use cases and possibilities for device OEMs to create vibrant, exciting search experiences.  Some of the new possibilities include:

– Global search across a huge number of different datasets

– Including photos of locations, products, menus in search results

– Including user generated reviews and ratings in search results

– Integrating sources from small, niche providers that may cover one vertical in particular metros but not offer full regional coverage

– Integrating sources that may lack accurate positional data

Contact CloudMade to learn more about the capabilities of our On-Dash and In-Dash platforms.

May 7th, 2013 - Posted by Nick Black in for OEMs, geodata, products, technologies

« Previous Page