Sentinel Safety

Sentinel Safety brings the latest AI technology to protect pedestrians in the rail industry

The risk of injury and near misses for workers around mobile plant and machinery is always a concern during the construction and maintenance of railways where separation of people and plant is not possible. To date, most hazard prevention technologies involve a tag-based system or administrative controls which are not always the most comprehensive solution to the dynamic rail maintenance and construction environment.

Developed over three years by an in-house team of Brisbane-based engineers, PRM Engineering Services have produced the Sentinel Vision A.I. pedestrian detection system, for the specific safety requirements of the rail industry. The system uses the latest in artificial intelligence and pedestrian recognition to detect when a pedestrian is in a machine’s blind spots or enters hazardous zones near mobile machinery, warning both the operator and the pedestrian.

Working in real-time, the Sentinel Vision A.I. system incorporates multiple cameras, that are mounted to mobile machines such as wheel loaders, excavators, and on-track vehicles with as many or as few cameras installed as needed to cover blind spots and assist the operator in identifying people around the machine. Sentinel Vision A.I. is the first pedestrian detection system to alert both the operator and pedestrian. This innovation has been found to cause long-term behavioural change in pedestrians working around mobile plant and machinery. Sentinel Vision A.I. uses a unique voice alarm system to ‘talk’ to the pedestrian, cutting through the beeps and buzzes that that workers hear on sites every day.

The system has been trialled by several top tier rail authorities with positive results and many have reported that when people are alerted by the voice alarm, behavioural change and greater awareness of risky behaviour around mobile plant is achieved. Ideally, over time, Sentinel Vision A.I. will be activated less as people have learnt not to walk in front or behind active vehicles, reducing the risk of accidents and injuries.

The system takes images from the detection cameras and then processes the information through an A.I. neural network to determine if there is anything that looks like a person, or part of a person, and if there is, it triggers internal and external alarms. Detection zones are customisable and determined with an easy to use drag and drop interface, and an additional option of pre-warning zones.  The system has been trialled and used in a range of different operating environments and environmental conditions with positive feedback.

Sentinel Vision A.I. is one of a number of innovative products developed by PRM Engineering Services. Part of the PRM Group of companies, which has been providing safety systems and equipment to the rail industry for over 20 years, PRM Engineering Services designs bespoke safety and control systems that meet the unique needs of operators. Through our partnerships and experience gained in the rail and heavy machinery industries, PRM Engineering Services’ range of Sentinel Safety systems were developed to meet the changing safety and risk management requirements of rail authorities.

The Sentinel Safety range also includes several Height and Slew limiters used throughout the rail and construction industries to allow safe operation around powerlines and within confined spaces. The Sentinel Height and Slew limiters have been used by rail authorities Australia wide for a several years and can be retrofitted to any machine with articulated booms.

Based on this experience PRM has also recently released additional optional features including HV detection and RFID for attachment recognition. By combining the functionality of our widely used Sentinel Height and Slew limiters with a patented Sentinel HV Aerial Module, the system can ensure safe operation around powerlines from the moment the machine is turned on. The system prevents the machine moving within the exclusion zone around powerlines and motion-cut valving prevents the machine from moving closer while allowing the operator to direct the machine away from the electricity source.

The Sentinel Height and Slew limiters are perfect for the safe operation of excavators, loaders, skid steers and backhoes when working under overhead powerlines, in and around bridges and inside tunnels and can be installed on new and old machines alike. The Sentinel Height and Slew limiters also have are range of rail specific systems to the meet the machine safety requirements of multiple rail authorities and councils.

PRM Engineering Services are passionate about safety and have a long-standing heritage of safety system design and installation since 2002. With experience in the rail and earthmoving industries, PRM Engineering Services have become integrators and developers of a number of unique safety and control systems that meet customer requirements. These projects have ranged in scope from customisations of height or slew systems through to full redesign of control systems for on-track rail vehicles. Along with our team of talented engineers, the PRM Group of companies can also assist with the installation or modification of electrical, hydraulic, and control systems for heavy machinery, enabling PRM Engineering Services to offer end-to end innovative and customised solutions to our wide range of customers.

Find out more at: https://www.sentinelsafety.com.au/.

Innovation in the world’s largest tram network

Melbourne’s iconic tram network operates across 250km of double track. Xavier Leal from Keolis Downer shares Yarra Trams’ latest innovation strategy that is digitising the network’s 5,000 daily services.

The world’s largest operational tram network has been transporting passengers in Melbourne for over one hundred years. Xavier Leal, manager of innovation and knowledge at Keolis Downer, acknowledges that operations throughout the urban tram network have considerably advanced since the first tram line was pulled by horses in 1884. As the operator of Yarra Trams, Keolis Downer has been investing in its digital strategy to prioritise data collection and improve passenger experience.

Leal has almost fifteen years of experience in strategy and innovation management. Since he joined Yarra Trams in two years ago, he has been driving forward innovations in the business that support enhanced passenger experience, operational effectiveness, and safety in the network.

Before his current role at Keolis Downer, Leal worked in the mobility and transport sectors in Europe. He has led a wide range of international projects that explored digital innovations and defining technology diffusion processes. His previous projects include developing innovative information and technology services, including T-TRANS and Collective Intelligence for Public Transport in European Cities (CIPTEC). Leal said Keolis Downer leverages its worldwide operational experience to explore innovations in smart cities through a digital mobility observatory.

Leal highlighted that it is important to note the difference between tram networks in Europe and Melbourne to understand how investment in processes will allow Melbourne to set an international benchmark for light rail infrastructure.

“Melbourne has a unique tram network. Trams elsewhere don’t have the same challenges that we have here. Not only is it the world’s largest operational tram network with over 250km of track and more than 1,700 stops across the city, but 75 per cent of the network is shared with road vehicles,” Leal said.

This means trams do not have separated corridors on Melbourne roads and operate amid buses, cars, cyclists, and pedestrians. This brings particular challenges with safety and operational performance, particularly travel times. Melbourne’s tram network could run more efficiently. To enhance network capability, Yarra Trams have used technology to enable faster services.

However, due to the nature of having assets distributed widely across the network, including the vehicles themselves, stations, and other monitoring points, there is the potential for the accumulation of digital data to support the more efficient operation of the network. Yarra Trams has recognised this, and is looking to digital innovation, with a number of projects deployed to target priorities including faster travel times, reduced disruptions, and customer safety. These initiatives include digitising asset management through real time-based platforms, to exploring crowdsourcing of data for safety and unplanned disruption management.

One project that Yarra Trams has trialled is the on-board collection of image-based data on traffic. In developing the technology, Yarra Trams took a consultative and collaborative approach by incorporating feedback from multiple stakeholders which come into contact with the relatively open network.

The development team looked to how they could incorporate real time data on traffic volumes to maximise operational efficiency and passenger experience. However, solutions were not always going to come from within the organisation, and Yarra Trams looked for partners who could enable this digital data project.

“Effectively engaging with the innovation ecosystem is another critical success factor to maximise digital technologies,” Leal said.

Keolis Downer collaborated with the Australian Integrated Multimodal Ecosystem (AIMES) to procure Toshiba’s traffic sensing technology. Leal said the data collection and analysis system was based on image processing and deep learning technology in a smart transport cloud system. A trial of traffic sensing by on-board unit (OBU) based image processing technology took place in March 2019 with two C2 trams travelling on route 96 from Brunswick East to St Kilda Beach.

Leal said the trial tested the capability of the technology to detect various states of traffic by deploying image processing techniques and transmitting the results to a cloud system. The OBU could detect traffic in terms of volume, vehicle queues, vulnerable road users, pedestrians and obstacles.

HD cameras captured real time traffic and processed and measured the information as it happened. The information collected from vehicle queue lengths waiting at red signal and pedestrian flow assessed traffic conditions to
a degree, while also detecting obstacles and service adjustment.

The OBU system consists of three units, a stereo camera, image processing hardware, and a signal divider. The OBU system sends detection results back to a central server. These results include images that have been tagged with GPS data. The trail enabled Yarra Trams to obtain geographically precise data to illustrate issues in the network in real time, enabling faster responses and comparisons with historical data.

The digital data collected throughout this trial may allow traffic management and operation control staff to instantly evaluate risks as well as predict needed safety measures.

Images taken by trams are used to map pedestrians and crowds.

“It was a successful project,” said Leal. “We assessed the system capabilities
to detect traffic volumes, vehicle queue lengths at intersections, pedestrian crowd volume detection and estimation around tram infrastructure. Now we are discussing with Toshiba, government stakeholders, and Melbourne University researchers the next steps to further evolve the system,” Leal said. Leal is proud to pioneer the use of digital data to evaluate complex transport networks. He said it’s not uncommon for large networks such as the Melbourne tram network to experience unplanned disruptions, so managing data from Yarra Tram allows a clearer understanding of behaviour of motorists, pedestrians, and other vehicles which the network comes into contact with.

Leal said trams and light rail services are the lifeblood of Melbourne, as they are the primary mode of public transport for inner suburban residents. Globally, more than 200 cities are now recreating, building, or planning tram networks. If the Melbourne network were to be rebuilt today, it would cost more than $20 billion and take several decades to complete.

“It’s important to us to have a holistic approach to our digital strategy, that leverages Keolis’s expertise in mobility and digital technology with a robust data management platform that aligns with the Department of Transport’s systems and tools,” Leal said.

“We are increasingly gaining more data flowing from digital channels. From a passenger experience perspective, it is important for us to integrate reporting capabilities with analysis of inputs coming from diverse channels,” Lead said. He said the company expects these channels to grow and further diversify as new streams of data and incorporated into the network.

“We are committed to keep pushing for further integration of information and data to ensure the right actions are taken to enhance Melbourne’s dynamic network,” he said.