Amid a spate of recent announcements relating to real time mobile services it would be easy to miss the greater significance of a joint development between Nokia and geographic information system (GIS) mapping software vendor Esri, that takes account of live traffic updates for route planning. The advance lies in the deeper level of data crunching and analysis required to combine GIS mapping with live information about traffic from multiple sources, compared say with a service delivering progress updates of individual trains, or planes during a flight.
The progress made could also be significant for delivery of real time video updates, including targeted adverts, in response to live information such as the mobile handset user’s location.
The ability to combine real time traffic information effectively with live GIS based route planning has been a goal ever since the first effective and generally available satellite navigation systems were introduced two decades ago. The stumbling block at first was over data delivery to the car and integration with the mapping software, but it turned out the greater challenge was in processing and assimilating the information from the various relevant sources. We might now call it a Big Data problem, though mercifully Nokia and Esri prefer the term ‘smart data’.
The main immediate target is traffic fleets, given the potential for big savings by enabling vehicles to avoid bottlenecks or points of congestion through the ability to continually optimize routes as journeys unfold. Nokia’s contribution comes from its subsidiary called Here, which specializes in real time traffic information as a potential source of revenue and business for its mobile division, and has now combined this with Esri’s existing web based location software packages called ArcGIS Online and ArcGIS for Transportation Analytics. Esri admitted that it has taken the two companies 10 years from concept to final testing, highlighting the extent of the challenge involved in combining traffic and map data.
Nokia introduced its new Here system on Nokia phones only last year, and put it into the all-important US market last November. This week it has said it is now offering Here Drive+ and Here Transit for all Windows Phone 8 smartphones. Until now, the apps were only available on select Nokia Lumia devices.
Here Drive+ offers voice-guided turn-by-turn navigation with offline maps and the new app has something called My Commute, a feature designed to improve the user’s everyday commute by learning your daily driving habits and predicting changes to your route based on current traffic conditions. Customers can upgrade navigation from local to global coverage from within the app.
But how can Here work? It was faced with getting to grips with event based data from a variety of government and commercial sources, along with billions of historical traffic records.
The key to keeping the data fresh lies in real time sources of data. Part of this is provided by Here’s own sensor network but this comes nowhere near to covering a whole road network, so it is augmented by probe data from consumers’ smartphones, exploiting the GPS capability they all have to pinpoint their location. All this combined contains a lot of information about traffic, but extracting it and assigning the relevant weight to each item of data has taken time and involved a lot of perfecting through testing. Essentially the system has to be compared with plain vanilla satellite navigation on its own as a base point and then with different levels of real time traffic processing to determine the best combination. There is little doubt that the system will be continually improved after its deployment in the field but Esri and Here have decided it was now sufficiently robust and adds enough value, to be launched.
The most important source of data is also the greatest in volume, coming from the consumer probes. This amounts to 10 billion data points that are all less than one minute old and therefore of greatest use for real time traffic analysis, plus a further 5 billion points between one and five minutes old that are still quite applicable. The problem is that this data does not immediately equate to traffic levels since it only detects the speed of each particular smartphone. The handset in question might be in a vehicle but it might also be in the pocket of somebody walking. If that handset had recently been moving at a higher speed it would be likely to be in say a car but could also be in a train, or the owner may have just completed the journey. Distinguishing between these outcomes is a classic application of smart data analysis that has taken Here a long time to get to grips with. The reward is that having done so, the same tools and techniques are applicable to other mobile data applications.
Real time traffic analysis itself is probably the telemetry application with greatest immediate commercial potential. We only have to look at surveys of the cost of congestion to different country economies to see that. One of the most recent surveys came from the UK Centre for Economics and Business Research and traffic information provider Inrix, which estimated that traffic congestion is costing the UK economy nearly £4.3 billion ($6.6 billion) a year. That sounds conservative alongside studies elsewhere, such as the Texas Transport Institute’s report that using standard measures traffic congestion caused $101 billion of delay and fuel cost for the US in 2010, the latest year for which data was then available.
This equates more closely with a study by an initiative designed to reduce greenhouse gas emissions called Trasnport2020.org, suggesting that congestion in Europe costs €200 billion a year or 2% of the continent’s GDP. That may be over egging it given that this group has an agenda to reduce fossil fuel consumption, but looks closer to the mark than the UK study, given that we have seen past estimates from that country’s CBI (Confederation of British Industry) that traffic congestion costs £15 billion a year there.
While the economic burden of congestion is felt particularly by commercial fleet traffic it also falls on individuals, with an impact on overall quality of life as well. Nokia’s HERE and Esri recognize that for individuals, whether travelling to work or making recreational journeys, route optimization is not just about minimizing time, so the new system supports personalization. Then people can set their favorite routes and be advised to deviate only when anticipated delays exceed a certain threshold.
Until now route planning systems incorporating real time information have largely been confined to rail networks, which is easier to accomplish because it is effectively a walled garden with often a single source of information about positions pertaining to a relative small number of trains. There are Windows Phone 7 Apps for this for use by passengers to plan trips, check live station arrivals and departures, and retrieve real-time notifications of service disruptions or delays to their journey.
A similar real-time GPS tracking system has been deployed for the larger Indian rail network, locating trains with an accuracy of 10 meters and maximum latency of two minutes. This is being incorporated effectively into the country’s signaling system as an extra layer of protection against collisions by providing drivers with information about where other trains are. However it has to improve on the two minute latency for this to be effective.
The key thing about Nokia’s new approach is that this system can be used anywhere in the world, and Here contains data from 196 countries. When you compared these features with the fiasco that Apple had around creating its own mapping system, you can see that the Navteq owned Nokia mapping system, already being used to great advantage to offer free navigation on Nokia phones, if it has the Here features as well, this would make Nokia phones a MUST in places like New York, cursed by commuter delays.
Meanwhile there is growing interest in application of real time processing to video and ad targeting. The same information sources and analysis methods deployed by Nokia and Esri would be applicable here, but there is another dimension, the delivery infrastructure. This is partly about capacity, but addressing that is not specific to real time interactive applications and involves developments such as Wi-Fi offload and build out of fiber. The real time communications challenge is more about ensuring QoS, traffic prioritization and supporting full duplex two way communications over the web. This is not just a mobile problem, since the web evolved initially to deliver best effort traffic on a one way journey, with a request and response approach. This made it hard to achieve real time capabilities even over the fixed Internet, since communications were half duplex involving short lived connections, with requests for web pages for example being distinct from the subsequent delivery of the data.
This has been fixed now by the network protocol WebSocket, which facilitates full-duplex communications channels over a single TCP connection. It was standardized by the IETF in 2011, and as the WebSocket API in Web IDL (interface definition language) by the W3C (World Wide Web Consortium). It has also been incorporated in HTML5, which is likely to take over from traditional native apps for browser based applications across the mobile web.
There will be some resistance to the advance of HTML5 on quality grounds, since native apps currently perform better and can offer a richer experience having been better optimized ‘ important considerations for games particularly. But momentum behind HTML5 is growing for several reasons, notably its potential for cost savings as it works across multiple platforms including the desktop, so that development costs are lower. There is a bigger and growing base of developers, noting that more than half of mobile developers have already built in HTML and nearly 90% are now planning to do so in HTML5.
Above all though HTML5 armed with WebSockets has in effect become an overlay across the web able to transmit real time updates and control distribution. More recently an important development that will stimulate development of mobile advertising around HTML5 was the launch of HTML5 for Digital Advertising 1.0 by the Interactive Advertising Bureau (IAB). This provides guidance on optimizing HTML5 ad code for better performance across multiple platforms. It establishes common ground for optimized HTML5 ads, aiming to reduce operational costs resulting as companies make the shift from graphical development tools like Flash to HTML5. This is a recognition that the scarcity of design tools is a handicap because it requires additional expertise, even though there are plenty of bread and butter HTML5 developers.
So with the platform and the analytical methods coming together, the stage is set for real time ad and video delivery to mobiles on the basis of relevant and accurate information.