August 1, 2007

What is the Probe Person Technology? /   Dr Eiji Hato, the University of Tokyo

The Probe Research Group, which I serve as representative, commenced its activities in 2003. The origin of this research group, or rather the root of the Probe Person Technology lies in a long-term monitoring project of individual travel behaviors, using attenuation property of the electric field intensity of PHS (Personal Handyphone System), which was conducted in 1998 by a research group created centered around Ehime University.

As the remote booting program for the positioning system did not perform well, we monitored the system for two weeks in front of the computer displays. I can clearly recall how excited we were for being able to immediately follow the daily travel behaviors of some 20 remote respondents.

This project, conducted with support from a highway company in Kansai Region, aimed to replace the existing traffic control system based on the Euler method utilizing data obtained from roadside detectors, with a new portable and more accurate traffic planning and control system based on the Lagrange method.

The project is now known to be the first trial of the Probe Person Monitoring in the world, and initiated the development of algorithms for route matching and auto travel detecting that are used in most of the probe systems today.

The collection of detailed data on trips and activities for traffic planning, and the implementation of a traffic information providing system were technologically possible; however, we could not establish it as a commercial service. Reasons were high communication cost of PHS; and the sufficient accuracy of the traffic information obtainable from existing roadside detectors. I admit with some sense of self-reproach that we could not present any positive reasons except "fun to do" when proposing our new system to replace the existing advanced system.

Following this our very first experiment, numerous projects by JSK (Association of Electronic Technology for Automobile Traffic & Driving) and Internet ITS Consortium have been conducted. And then widely distributed probe vehicle system utilizing DoPa network (packet communication service network provided by DoCoMo) has been implemented. The service, to which a considerable amount of budget was allocated, initially received criticism from taxi companies and others arguing that cell-phone service providers were the only beneficiary of the service due to its communication charges, has now developed to have commercial probe information services operated by automobile and electric appliances companies.

Encouraged by the huge budget allocated to the probe vehicle projects mentioned above, we started seeking other possible applications that are inexpensive and stimulating. This directly led to the development of the Probe Person Technologies. One of the technologies is the data cloning approach. In 2001, an accident at a firework festival occurred on an overcrowded overpass in Akashi City, Japan and claimed some spectators' lives. To prevent a recurrence of such accidents, we developed the cloning algorithm, which allows us to infer the whole situation from a small number of samples. In 2002, we successfully implemented this cloning technology for delivering the spectator congestion forecasts for FIFA Korea/ Japan World Cup games. I myself was almost dragged to a police station for inquiry when we were installing temporal base stations for PHS positioning at the stadium of Sapporo Dome where security was quite tight against expected hooligans.

Back then, the potential market size for ITS was said to be trillions or even tens of trillions yen, suggesting explosive dissemination of the road traffic information services using probe vehicles. However, in reality, they are not as widely spread as expected. Although the ETC (Electronic Toll Collection) system is fully installed on expressways and widely used by drivers in Japan, the application of automated driving technology has faced with a difficult challenge. Consequently, some say "ITS's Rosy Future Theory was wrong," and others set up a weblog with a joking title of "I doubt ITS." The necessity for the services beyond what currently performed with conventional VICS (Vehicle Information and Communication System) is being questioned.

This question is best answered by simply reminding people that "ITS stands for Intelligent Transport Systems." Automobiles are only a part of the intelligent transport systems, and the information people really need is not only "highly accurate information on expected travel time for automobiles," which VICS now provides. I believe that the real objective of ITS is to evolve mobility in the cities to maximize people's fundamental needs for traveling. To regard the ITS as solely a problem-solving technology for vehicle traffic, and Probe Technology as its key technology would make them difficult to be commercially viable, and, above all, that would confine the intrinsic value of ITS to a narrow range.

In the field of the Internet, technologies called "Web 2.0" - such an equivocal term seems to be trumpeted by novices who have never written a single program - appear to have attracted much attention. Based on the information on Web 2.0 gathered from the Internet, the boundary between Web 1.0 and Web 2.0 seems to be whether or not it provides customized service for each individual using huge databases, and has adopted the concepts of MMO (Massively Multiplayer Online) or CGM (Consumer Generated Media).

The essential part of the Probe Person Service is Web Diary, which is a combination of a service similar to weblog that allows participants to submit comments on their lives, and a technology that automatically infers their behavioral contexts from the data obtained from multiple sensors. The service can also deliver traffic forecasts customized for each user based on the information stored in its databases, and it also contains a search engine, temporal /spatial reminders that also to be optimized for thin client systems. In this sense, the Probe Person Service shall be the Web 2.0.

However, such service items are regarded as a-matter-of-course user services provided over the Internet, nothing special today. But providing these services seems to be difficult in the arena of probe vehicles. The service that only provides "the information on expected travel time for automobiles," with sufficient attention paid to privacy, and that charges expensive fees to users, is somewhat rigid compared with the Web 2.0-type service. Keeping in mind the situation where sales of automobiles face downward trend, I cannot imagine that the users without sufficient money to spare due to bills for mobile phones and occasional drinking parties are willing to pay for such a service.

Also the standardization of APIs that is mentioned in the realms of probe vehicles / person and GIS seems to be difficult. In the case of Google, its API is actually standardized but only a limited part is disclosed. Should the use of the API be prohibited after a full-pledged service has commenced, it would ruin the whole thing. It may be no acceptable if the trial of standardization is demonstrated in a research paper by a graduating student, but standardization in the business arena where data is vital seems to be unrealistic.

The above discussions lead us to suspect that the only areas where probe information can be commercially utilized, if any, are Business-to-Government and Business-to-Business or B2G and B2B. Among them, safety and security, and distribution are sectors where information value is claimed to be tremendously high. In these sectors where quality customer service is emphasized, the needs for highly accurate traffic information or consultancy services are high.

It is needless to say that the areas with business potentials will be developed regardless of users' expectation. Conversely, from the users' perspective, they simply seek for highly accurate probe information services regardless of its business potentials. However, vast majority of the users who are accustomed to the Internet world must want to have the service "for free of charge". It seems rational to secure data infrastructure "in some way" for more accurate probe information services since traffic information has a highly public nature. The current situation, however, is almost unmistakably that of Prisoners' Dilemma, where multiple players are suspicious one another. Consequently, I believe an independent third party is necessary to coordinate the effort of releasing them from the trap of the false belief.

As for the business model that Google, a company proclaims itself as a "Web 2.0 company," has established, I do not see any innovative technologies in it. The algorithm of the PageRank that measures the importance of web pages has now become a commonly used technology. Today Google is widely known rather as a company that manages a tremendous number of databases; it is well-known that their core technology is increasingly shifting to securing power supply for a vast numbers of database servers.

Regarding the Probe Person Technology, the technologies that this research group develops do not cover the whole scope; various other technologies will certainly be developed. In the past, the development of algorithms that accelerate map matching and other applications was vital; now, the development of new technologies such as navigation with panoramic-view, librarian for data search, and fabrication of indexes and canonicalization of data for data mining has become vital in the arena. This is because the applications of the Probe Person are now considered to be smart mobility network such as transportation-related eco-money, as well as marketing and advertising, health management, and mobility design.

The Probe Research Group started its activities at Ehime University in 2003 with 15 researchers of Japan under the research title of "A Long-term Monitoring of Trips and Activities in Urban Space." Recalling those days today I am grateful for as many as 15 researchers gathered at Ehime, some 400 miles from Tokyo. Honestly speaking, to which direction a technology proceeds in the next is only known to the people who implement the technology.

There goes a slogan-like saying "Input everything in the world into Google's DB." This means Google is extraordinary because, for example, as you use the Internet only via Google, it can tell you what you like, the time and date you were on the Internet, and even it will help you manage your healthcare. However, I am suspicious about the idea of Google having such technologies because real-time behavioral data is never obtainable through keyboards.

If asked "What is the Probe Person Technology?," I would answer with "Integrated technologies designed for recording people's awareness and behaviors in real urban space." It is nothing more or nothing less. And that is the nature of the Probe Person Technology. I have a feeling that it will grow to be a vital social technology that plays a fundamental role both in the Internet world and the real world.