Depth cameras and computer vision have this week been rolled out at Wynyard Station in Sydney, in an effort to reduce platform crowding and cut the time it takes for trains to load and unload passengers.
The aim is to minimise what is known as the train’s ‘dwell time’ – the time it needs to stops at each station to allow passengers to board and alight, and the biggest variable affecting whether a commuter train is able to stay on schedule.
Sixteen Ethernet connected depth cameras have been installed along the length of Wynyard’s platform three, which fire an irregular pattern of dots from an infrared projector at waiting passengers.
An infrared camera then captures the pattern and sends it to a processor to work out depth from the displacement of the dots – which are more spread out close to the camera, and denser on objects further away.
The system then applies computer vision techniques in order to identify people in the footage, by constructing a ‘Head-to-Shoulder Signature’ nicknamed by its developers as ‘the two watermelons’.
As people’s movements are tracked along the platform, the system detects what Sydney Trains called the “hot spots and friction points” of overcrowding.
This data is combined with API feeds of carriage weight – indicative of how full a carriage is – and accessed by platform marshals via a tablet so they can distribute people more evenly along the platform and move them away from crowded carriage doors.
“This could be a technological solution to a very human problem. Customers often become creatures of habit and wait at the same spot at train platforms, without realising that by moving down the platform they are more likely to board faster and find a less populated carriage,” said New South Wales transport minister Andrew Constance.
“The amount of time a train is stationary at a platform while people alight and board is a major factor in whether customers reach their destination on-time. Precise mapping of crowd behaviour and what we call ‘train dwell times’ will help us improve systems to manage customers and make sure they get where they need to go,” he added.
Sydney Trains said if the trial was successful, Dwell Track could be made available to staff via a downloadable mobile app.
The Dwell Track system has been developed by researchers at UTS working with the Rail Manufacturing Cooperative Research Centre and railway engineering group Downer Rail. It was demonstrated at the CeBIT technology exhibition in Sydney last year and has been tested previously at Brisbane Central, Town Hall and Redfern stations.
Potentially the system could be hooked up to red, amber and green ‘traffic lights’ along the platform so people can see for themselves the best spot to wait.
Always looking at new ways to improve
The trial continues Sydney Trains’ effort to reduce delays and improve customer service with technology. In May last year, the network introduced real-time train occupancy data – based on carriage weight readings – on Waratah trains via an app. Since earlier this year carriage occupancy indicators have also been displayed on screens at certain Sydney stations.
Downer Group has also rolled out a data analytics and machine learning platform to Sydney’s fleet of Waratah trains which captures data from trains and predicts the remaining life of components to better schedule maintenance.
“With more than 1.4 million customer journeys a day, we are always looking at new ways to improve our customers’ experience. We have already had great success with our Fast Track Teams and are also exploring a range of cutting edge technology initiatives, including CCTV data modelling,” said Sydney Trains chief executive Howard Collins.
“We look forward to seeing the outcomes of these trials in the coming months,” he added.
In 2017/18 Sydney Trains provided 359.2 million customer journeys, up from 81.2 million customer journeys five years prior.