Smart Cities and Artificial Intelligence

Subscribe To Download This Insight

By Dominique Bonte | 3Q 2017 | IN-4706

While Artificial Intelligence (AI) is making inroads into all technologies and verticals, it will prove to be particularly relevant in smart city contexts. Legacy smart cities approaches have mainly focused on connectivity, sensors, and monitoring to offer better information and achieve modest, incremental cost savings and improved public services to citizens. Typical examples include smart waste management, smart street lights, and parking sensors. To take smart cities to the next level, both in terms of the quality of services and the cost at which they can be provided, AI and Deep Learning (DL) in particular, will be required.

Registered users can unlock up to five pieces of premium content each month.

Log in or register to unlock this Insight.

 

Can AI take Smart Cities to the next Level?

NEWS


While Artificial Intelligence (AI) is making inroads into all technologies and verticals, it will prove to be particularly relevant in smart city contexts. Legacy smart cities approaches have mainly focused on connectivity, sensors, and monitoring to offer better information and achieve modest, incremental cost savings and improved public services to citizens. Typical examples include smart waste management, smart street lights, and parking sensors. To take smart cities to the next level, both in terms of the quality of services and the cost at which they can be provided, AI and Deep Learning (DL) in particular, will be required. 

AI, and the automation it enables, is going to be instrumental to lower costs, by reducing the number of employees performing manual monitoring and other tasks. Example use cases of AI in smart city contexts include:

  • Public Safety and Traffic Management – Automated safety and traffic camera systems; with the deployment of surveillance cameras expected to continue aggressively, the downstream employee cost of monitoring and analyzing video footage in real-time risks becoming prohibitive, prompting the deployment of AI-based automation
  • Chatbots and Digital Assistants– Providing citizens seamless access to government services and information; replacing human operators with machine equivalents is already being trialed by the city of Los Angeles
  • Intelligent traffic lights – Optimizing signal phases to maximize traffic flow based on collective learning
  • Automation – Driverless cars and shuttles for Mobility as a Service (MaaS), drone and robot-based freight delivery, autonomous utility vehicles for street cleaning and garbage collection, drones for maintenance inspection
  • Security – Detection of terrorist threats and planned cyber-attacks; crowdsourced intelligence from social sites will be analyzed and used to prevent both physical and cyber-crime through advanced AI technology.   

However, the biggest challenge for smart cities will be to tie all technologies, platforms, and solutions together into one holistic, automated, ultra-complex closed-loop system which will work autonomously; automatically altering parameters based on data analytics to regulate supply to match demand.  Examples include imposing maximum speeds to driverless vehicles to optimize traffic flow, dynamic prices for toll, transit, and car sharing to limit demand during peak times, postponing recharging of electric vehicles to grid off-peak periods, and ensuring that affordable resources are available when and where they are needed. This represents the end game for smart cities. It will require the deployment of advanced AI tech to build demand-response systems which are able to automatically adapt and reconfigure themselves to match fluctuating demand levels for mobility, freight transportation, energy, communication, housing and office space.

Local governments adopting AI will need to make sure solutions are tested exhaustively and validated formally, especially when it relates to allocating resources or setting safety relevant parameters, such as speed restrictions. The problem isn’t unique to smart cities, for example NVIDIA spends a lot of effort on testing and validation, but accountability is a particularly important element of local government.

A Closer look at NVIDIA’s Metropolis Solution

IMPACT


NVIDIA has been one of the vendors leading the way in making available GPU-based hardware platforms that support deep learning across a wide range of applications with driverless cars being the most spectacular one. At the GTC event, held in California earlier this year, the Metropolis intelligent video analytics edge-to-cloud platform was announced, squarely targeted at smart cities. Metropolis is about applying deep learning to video streams and deriving insights for smart city applications such as public safety, law enforcement, traffic management, and resource optimization with cameras deployed at government properties, in public transit, commercial buildings, and along roadways. The solution leverages the NVIDIA Jetson TX2 at the edge and NVIDIA Tesla in the data center. NVIDIA also announced 50 AI city partner companies, among which Avigilon, Dahua, Hanwha Techwin, Hikvision, and Milestone. This could represent the start of a fledgling smart city AI ecosystem. 

Metropolis offers the promise of exhaustively analyzing all available video footage - NVIDIA expects around 1 billion cameras in operation by 2020 - with a ultra-high accuracy and reliability of which human operators could never achieve.  

It’s all about doing more and doing it better with less money

COMMENTARY


While the full power of AI and DL is still being discovered, concerns about the effect on employment are starting to be raised, especially within government circles.  The deployment of AI-powered, job-killing automation is far from obvious from a political perspective. On the other hand, in the face of having to address a growing number of urbanization issues within limited city budgets, governments will be forced to embrace AI, rather sooner than later. At the same time, across the globe, governments are exposed to increasing levels of pressure to spend tax payers’ money more efficiently. AI is the obvious and the only answer.  

Services

Companies Mentioned