Building Safer, Smarter Cities With Artificial Neural Networks

Data is not only collected through the city’s IoT network of sensors (monitoring temperature, air pressure, humidity, pollution levels and more), but also from video (intelligent video surveillance, traffic cameras, etc) and - as eGovernment services gain in popularity - existing systems of record (billings, customer relationship management data and more).

Building Safer, Smarter Cities With Artificial Neural Networks

by Kameshwar Rao Sorda, solutions director at Huawei Enterprise Southern Africa

While a number of cities globally are turning to the Internet of Things (IoT), wireless connectivity and cloud computing to improve internal efficiency and service delivery to communities, the future lies in harnessing Artificial Neural Networks to automate and continually improve many of these operations.

While machines have long helped us speed up our daily tasks, the requirement for them to solve more complex mathematical and logical challenges means that we need a different kind of computer. Based on the structure of our brain, artificial neural networks include adaptive learning, self-organisation and fault tolerance features. By using new data and identifying patterns, it can learn or improve upon its performance without being programmed by humans.

This technology in itself is not new, the first artificial neuron was produced in 1943 by the neurophysiologist Warren McCulloch and the logician Walter Pits. But the technology available at that time did not allow them to do too much, but has grown in popularity due to our ability to collect and transmit information from a multitude of sources. It has already been applied in many industries, including risk management and industrial process control.

Where AI can help cities improve

Kameshwar Rao Sorda, Solutions Director, Huawei Enterprise Southern Africa

Kameshwar Rao Sorda, Solutions Director, Huawei Enterprise Southern Africa

The growth of smart city project implementations has already provided authorities with a huge volume of information from various systems and sensors in place. According to Gartner, cities are the fastest growing area of IoT, with 3.3 billion devices predicted to be connected by the end of 2018.

By sifting through this big data, we can identify trends and patterns to enhance city operations and service delivery to residents. Data is not only collected through the city’s IoT network of sensors (monitoring temperature, air pressure, humidity, pollution levels and more), but also from video (intelligent video surveillance, traffic cameras, etc) and – as eGovernment services gain in popularity – existing systems of record (billings, customer relationship management data and more).

Artificial neural networks work by analysing the millions of data points collected, for ‘training purposes’. The process gradually provides computers with the capability of calculating with a certain degree of probability the identification of basic patterns from raw data such as shapes, sequences, frequencies, colours, order, and more. By stacking layers upon layers of patterns in a neural network model, they are able to accurately identify patterns of greater complexity.

This provides cities with the opportunity to improve traffic management, reduce water and electricity wastage, and even improve public safety and security in the urban environment by identifying and predicting crime trends, and empowering authorities allocate resources to the right areas.

Smart connected street lights – which could include CCTV cameras and WiFi hotspots – not only provides residents with a safer environment, but also dims the lights during periods of inactivity or low traffic, helping municipalities cut down on lighting energy costs.

The city of Santander in Spain, often seen as a role model for smart city development, has a project in place that uses sensors to cut down on time wasted by motorists looking for available parking spots, as well as the resultant air pollution. This concept can be taken further, with artificial neural networks being used to predict – with very high accuracy – the availability of a free parking spot at a future time.

Infrastructure, skills investment for the future

As with any Safe City or Smart City project, cities will need to invest in the required infrastructure, including reliable connectivity – whether broadband or narrowband – as well as in the cloud data centres that provide the higher CPU (serial processing) and GPU (parallel processing) computation power needed to collate and analyse all the data collected, and drive the automation and continual improvement of many functions.

It is not just about hardware and software either; human resources will be just as important, with cities and corporations needing to invest to ensure that they have the basic skills required to sustain an AI-led future. This includes data preparation capabilities, basic and advanced algorithms, modelling, automation, iterative processes, cloud management, infrastructure and facility management, and more.

Artificial neural networks are also being used in areas such as texture analysis, recognition of speakers in video-based communications, recovery of telecommunications during faults, interpretation of multi-meaning words in languages that use logograms (such as Chinese and Japanese characters), three dimensional object recognition, handwriting recognition and facial recognition.

Artificial neural networks might one day be able to perform tasks that a human brain cannot. From across industry verticals, to specific functions such as automation, forecasting, cybersecurity and more, there are endless opportunities in future for how this technology will be used to improve the way in which we live and work.

COMMENTS

WORDPRESS: 0
DISQUS: 0
%d bloggers like this: