An autonomous AI city infrastructure management system will be launched in Tyumen. The project is part of a general agreement between the regional government and Cognitive Technologies, a Russian IT company, on innovations development. The agreement was signed on June 4 on the sidelines of InnoWeek-2018 forum in Tyumen. The company which developed Skolkovo smart city concept, has promised to launch the Cognitive City pilot project in Siberia in the fall of 2019. Andrei Chernogorov, Cognitive Technologies’ CEO, explained to Invest Foresight what AI is capable of and what kind of intelligence cities need.
– Mr. Chernogorov, what is the difference between a neural Cognitive City and a smart city?
– The smart cities concept is gradually fading out. It intended to equip residential areas with lots of sensors which in fact ensured a one-way relay-like information flows. We offer a much more sensitive use of AI and neural networks technologies when a city not just delivers data to a certain management center but collects all of respective information on itself and makes adequate decisions.
– What sort of decisions?
– We see an automatic city environment diagnostication and a subsequent issue of instructions and their execution monitoring as a key element of the concept. We suggested to the municipal authorities that we install data collection devices at the city vehicles. In Tyumen, local buses will be used for that, but generally those can be any terrestrial means of public transportation employed in a city, i.e. public transport and municipally owned vehicles. Such devices combine cameras and computers and are capable of recognizing traffic signs, road marks, traffic lights, etc. The data collected is used to produce a 3D city environment model which is continuously monitored thereafter. A deep learning neural network makes it easy to monitor streets cleanness, garbage removal timeliness, road pavement and marks state, lighting systems operation, parking rules obedience, etc.
Deviations from standards and any breaches are immediately recorded and then instructions are forwarded to contractors responsible for a particular item. The AI’s task will therefore be not just detecting a problem, but monitoring it until it is rectified.
– One of the most pressing issues is the potholes that are where they are forever.
– The execution of instructions is nowadays monitored occasionally. A thousand irregularities and violations may be registered daily. Yet just a tiny portion thereof, the most blatant ones as a rule, are occasionally monitored by a municipal commission. All others remain intact causing people’s discontent. The problems keep growing and over time their solution becomes more cost and effort intensive.
– When may the first results be expected?
– We hope Tyumen will become a pilot city. The system prototype will be installed in the city streets in 2019 fall, its full-scale operation will start in 2020. The Tyumen project has far-reaching goals. Once we have it action tested, we will offer it to all cities with a million-plus population worldwide. In their case municipal infrastructure does not require any preparations for its implementation. According to our assessments, the cost of the system operation (including equipment leasing) will be about $200K per annum. For the budgets of large cities, that is negligible money.
– How do you see the demand?
– Our assessment of the annual market of such solutions is $0.5 bln. Our advantages will include the urbanization trend. Urbanization is progressing exponentially and, according to Ernst & Young, 80% of the global population will live in cities by 2020. Over the past ten years, the number of the cities with a million or more residents, has doubled. Given the said growth of urban agglomerations and sophistication of city management systems, it becomes impossible to handle cities manually. Procedures which can be computerized through the AI use, must be computer managed. At the moment, there are about 500 cities worldwide which generate a stable demand for such solutions, and further 1,500 are sources of some partial demand. Among those, Russian cities are not at the bottom. Neural networks based solutions for municipal economies make up a promising market.
– Are there any hardships with the project’s implementation?
– As far as technologies are concerned, there are no problems whatsoever. An operational prototype is ready, since we use Cognitive Pilot which has been successfully employed for solutions in UGVs. Our present-day task is imposing a neural network onto a geoinformation subbase and preparing it for integration into municipal systems. We therefore are rather tasked with adjustment and integration.
Yet, we require a year to fully prepare the project. We need to fully teach a neural network at a year-round data set, to test the system in all sorts of weather conditions.
– Why Tyumen?
– The climate is perfect here. Investment climate, I mean. For over a year, the group of our companies has been present in Tyumen. We have one of our development centers here. The regions needs innovations, and it has a dynamic team which is truly willing to fully implement projects, not just list the problems which make such implementation impossible. Institutional environment often traps out lots of brilliant ideas without producing any visible results. Having operated in Tyumen for a year, we realized they have good organizational and financial climate for breakthrough projects. That is one of Russia’s most innovative regions which also has resources to bring large scale solutions to full implementation.
By Ann Oreshkina