City Brain-type transport co-ordination coming

This story first appeared in the November issue of EVtalk – CLICK HERE to download the magazine FREE

We may all soon be part of a connected transport artificial intelligence (AI) network.

Alibaba has developed an AI system called City Brain, designed to co-ordinate the transportation system in Hangzhou, China.

The system resulted in a 15% decrease in congestion in the first year.

In addition to the time saved because of congestion (estimated to be an average of three minutes per trip), crashes are tracked in real time, allowing for faster response.
Traffic lights are also co-ordinated for emergency vehicle access.
In fact, it is estimated that emergency vehicle response time has been reduced by 50%, allowing emergency vehicles to arrive an average of seven minutes faster.

The system has also made illegal parking a thing of the past due to the tracking of vehicles and the use of image recognition technology to analyse over 3000 cameras in real time.

City Brain is built upon a powerful AI-machine algorithm.
Connected technologies combined with the ubiquity of cameras creates a transport network where essentially every vehicle is connected.
This data is collected and fed into the AI.

In addition to monitoring traffic and parking, cameras are used to collect information on queues for public transport.

The City Brain project began by giving Alibaba access to traffic data from the transportation bureau, public transportation systems, mapping apps and video feeds from thousands of cameras.
Then the system was given control of 104 traffic lights.

The AI has grown and learned as more sensors are added to the network and it was given control over more systems.
Today, the system includes 1300 traffic lights and is connected to 4500 traffic monitoring video cameras; covering an area of 420 square kilometres.

The system is supposed to be set up so the city owns the data and Alibaba owns the software.
It is interesting to consider how this system could evolve.
For instance, the AutoNavi mapping system – which is also owned by Alibaba and integrated into City Brain – has been used to power a new ride-hailing service.

What are the potential impacts of that integration? It is easy to envision the system giving preferential treatment to its own services.

This might sound unfair, but a truly smart city will require connectivity.
Piecemeal solutions will not be able to deliver the same level of integrated data and services, the AI will not be as intelligent.

What would be the consequences for competition? If this is the smart city of the future, how can we ensure neutrality for competition and innovation?
What about privacy? How can we guarantee oversight of current, or more importantly, future use of data?
Alibaba has recently agreed to develop a City Brain for Kuala Lumpur, Malaysia.
One of the focuses of this second initiative is to open the data up to innovators for use by enterprises, start-ups, and research institutions.

With potential benefits like this, we might ask ourselves why this has not already been done in cities in New Zealand.
It is inevitable that this will be part of our future, starting in cities, then connecting between cities, creating a networking country.

Of course, in New Zealand we must worry about things like privacy.
There are other risks associated with highly centralised systems such as hacking and even simply failure.

So, what would this look like in New Zealand? Obviously it will start in Auckland.
Would we be willing to give up so much of our data, how would we provide oversight? Would it be private or government run? What benefits would justify the cost?
There are options other than large centralised systems which might mitigate some of the risks and costs.

Imagine, for instance, Kiwis volunteer their personal and commercial computer resources.
These computers link to a distributed network when idle and that network then provides the analytics necessary for the system.

The best part of this idea is it provides an opportunity to utilise resources already available.
Systems like this (such as the BOINC software platform) have been used for analysing big data sets in astronomy and other scientific fields for over a decade.

Most importantly, however, is that we start thinking about every solution as part of an evolving system.
When planning and building transport corridors, we should be thinking about what solutions will replace it in the future.