Driving road safety

Share this on social media:

Topic tags: 

Tom Eddershaw looks at the latest vision technology for monitoring traffic, including systems designed to make cyclists safer on the roads

Intelligent Transport Systems (ITS) are being used in more imaginative ways on the road and are starting to make their way into vehicles. From 25 to 28 March, Intertraffic 2014 in Amsterdam will be showing how machine vision is making roads safer and traffic run smoother by selectively monitoring images of traffic flow, potential hazards, and recording traffic violations.

Volvo provides an example of this trend. The company has designed an automatic braking system to assist drivers if a pedestrian steps out in front of the car. Identifying dangers on the road is a logical step for vehicle manufacturers, and more complex, intelligent vision systems are being integrated into cars and buses with the safety of other road users in mind.

At Intertraffic, the SmartCycle system from Iteris has been shortlisted for an Innovation Award for intelligent traffic systems. The software identifies cyclists at intersections, so that the time the traffic lights stay green or red can be changed accordingly. This is an improvement to the old way of managing the increasing number of cyclists on Europe’s roads which simply added time whether or not a cyclist was present. Now, if a bike is visible to the SmartCycle system, a reasonable amount of time is allowed for the cyclist to exit the intersection safely.

The safety of cyclists is not only a concern at traffic lights. Bicycle collisions with motorised vehicles are becoming more frequent as the number of cyclists in city areas increase. During November in London last year, six cyclists were killed on the road in two weeks, highlighting the need for protection for cyclists in the capital. All bar one of the incidents involved some kind of large vehicle. With ever busier roads, bus drivers in cities are often at the point of sensory overload. Intelligent systems could offer a helping hand to the drivers and improve the safety of road users.

One such system, from Fusion Processing named Cycle Eye, uses a combination of intelligent camera systems and radar to detect cyclists that are in a dangerous position relative to a bus. By alerting the driver only to cyclists and ignoring cars and other vehicles, this new system hopes to keep the effectiveness of each alert at a high level. Until now, systems have alerted the driver to anything on the road that got too close to the vehicle or used cameras that gave a live feed to the driver. Both of these types of system never fully succeeded as the driver became annoyed by the system and ignored the frequent alerts or was distracted by the screen.

‘We thought we could provide something that tells them when a cyclist is in a blind spot instead of using a screen,’ explained Jim Hutchinson, CEO of Fusion Processing. ‘[The screen] could be installed on request but we believe it is something that makes this worse instead of better.’

The company saw a niche in the market. He said: ‘We set out to find a way to create a product that had a very high detection rate and a very low false detection rate. The prototype is very close to 100 per cent detection with very low false detection.’

After completing research and holding discussions with the bus manufacturers the company decided to use a camera and radar system, because, as Hutchinson explained: ‘The system needs to work in all weather so a camera-radar system is used.’

The main two cities that have been involved so far are London and Bristol, and the company is interested in approaching other cyclist-heavy cities such as Cambridge, Oxford and York.

It is key to the system’s success that drivers know that technology like this is an aid, explained Hutchinson, not a substitute for good driving, and drivers still need to be trained to use their own eyes and not rely wholly on the system. ‘They still need to do everything that they did; it’s effectively given them an extra set of eyes to help when they already have so much going on.’

In the future, the system could be used on smaller vehicles to alert a driver to cyclists or to warn HGV drivers of cars sitting in the lorry’s blind spot. ‘It would be very easy to do, possibly just with a software upgrade,’ Hutchinson said.

Hutchinson also said that the company would be interested in providing for cars but they are yet to have a manufacturer approach them. He explained that there is, however, an issue with the system being installed on smaller vehicles: ‘With cars, these are typically added as optional extra and I am unsure how many people will be willing to pay for the increased safety.’

The HGV market also has an issue. The problem is, on a busy motorway, the system could then become a distraction for the driver with many cars on the road. ‘It’s key that the product is something useful to the driver instead of something that begins to hinder them.’

Speed kills

The government sets rules for road users to help keep drivers safe and large advertising campaigns have attempted to raise the public’s awareness to the dangers of speeding – ‘speed kills’ is a common slogan used in adverts in the UK. Breaking these restrictions often leads to fixed penalty fines, but how can you monitor the 245 thousand miles of road that are used daily in the UK?

Nations are turning to ITS which uses Automatic Number Plate Recognition (ANPR) to identify those who do break the speed limits. By monitoring the average speed over a certain distance, as opposed to spot speed checks, means that vehicles abide by the limit for a longer period of time.

Europe’s move to intelligent transport systems is being led by the UK and the Netherlands with the installation of average speed cameras and other enforcement systems that use ANPR.

These systems have been adapted from cameras for industrial processes such as line scanning and the logistics industry, both of which use Optical Character Recognition (OCR) technology. It is, by and large, a similar process; both tasks involve a fast moving object passing through the field of view and have to recognise and capture an image of a small identifying code within the frame.

However, road-side systems also have to comply with government requirements to preserve personal privacy, and each image has to be reliable enough to be used as evidence in court.

Most of the regulations relate to the software as opposed to the hardware, as the restrictions focus on the protection of data. For example in the UK, where the requirements are decided by the Home Office, a separation between an enforcement camera system and any other system is required.

Once an image has been taken and the system identifies an issue, the data is burnt to a DVD and passed to the police. A trained reviewer looks at the images and the information to decide whether an offence has occurred.

By using a DVD transferring method, the evidence file becomes ‘write once, read many’; the possibility that it might be tampered with is reduced, and the file is admissible in court.

However, this ‘air gap’ means that there may be two very similar products installed for different purposes next to each other at the roadside.

An example of this is a Siemens Mobility product called Sicore. Sicore is an ANPR camera used in the Siemens system, SafeZone, to enforce average speed limits over long or short distances down to 75m. But it can also measure journey times and help predict possible traffic congestion. This information is processed at the UK’s National Traffic Control Centre (NTCC), which is just outside Birmingham, then sent out by broadcasters or displayed by on-road notification boards.

Mark Bodger, research analyst for Siemens Mobility UK, explains how this helps his travelling from the company’s Poole-based headquarters. ‘We often drive up the A34 from Poole to the Midlands. Around Southampton you get signs saying how long it is going to take to get to the M3 or the M4 and that information typically comes from the NTCC.’

However, because of the legal requirements attached to law enforcement cameras, the traffic monitoring and speed enforcement images cannot be captured by the same system. If the objective is traffic monitoring, then the system will use the number plate as a marker but discard the images.

On the other hand, for average speed enforcement an image needs to be kept in order to identify the offender, as Bodger described: ‘If the speed [of the vehicle] is less than that of the speed limit then the images are discarded. If it is above the threshold, the in-station system requests the images and data that have been recorded, in the form of a fully encrypted evidential record. The system combines the data from both cameras with the speed calculation to create a violation record. This record is then burnt to a DVD and transferred to a machine that is not connected to the initial system. The police can then decrypt the data and identify the offender.’ This degree of separation between the enforcement system and the police is the air gap.

The image of the plate is normally infrared, illuminated by a built-in infrared flash. Currently in the UK this second image is usually in colour; however, in similar systems in the Netherlands the context image is also infrared. This makes the image more reliable in adverse conditions and removes the need for visible, supplementary lighting to create a colour image at night.

Because of the extensive testing process which takes around two years for enforcement solutions, a product is already two years old when it ‘hits the road’. There is often new technology that could improve the installed system and the infrared image is an example. ‘This is a possible enhancement in the UK,’ states Bodger. ‘We are in continual conversations about improvements with our clients and the Home Office. However, each element requires more testing.’

If a new element is to be introduced as an upgrade it has to undergo some or all of the testing process again. As it is often one part of the whole system, it only needs to satisfy the relevant criterion which reduces the testing time. What is more, ‘incremental improvements are allowed as the pre-upgraded systems and parts are still approved’, says Bodger.

The clients still have some say over the pace at which the improvements happen. In the UK the Home Office reserves the right to make any improvement in all of the operational systems. Bodger explains: ‘That’s normally if a fault is found or there are questions on the reliability of the operation, i.e. if an operational error meant that it did not satisfy the efficiency requirements of the end user.’

These intelligent transport systems were born out of technology that was first applied in the warehouse, where the line scan cameras operated in an optimal lighting environment. Traffic systems in contrast, must be able to work during the day, the night, in snowy conditions, in foggy conditions and still recognise a number plate moving at speeds of 70mph or more. Bodger states: ‘Normal recognition cameras have lighting and a controlled environment. These [ITS cameras] have to deal with pretty much everything the environment throws at it.’

Enzio Schneider is the product manager for the traffic market at Basler, which is involved in global projects supplying cameras to integrators of ITS and which will also be present at this year’s Intertraffic event. Schneider explained that one of the problems facing traffic cameras is that, at night, they can be dazzled by extraneous sources of light. He described how the Basler cameras cope with this by using infrared light pass filters that block out other ambient light during the night. ‘Especially during the night, you will get a very stable performance as the cameras have a defined supportive light and you have a short exposure time. During the day you block off most of the sunlight which allows for stable performance.

‘If you need colour images, it is different and more difficult. You need the ambient light for colour which means the sunlight will change a lot and you will need to adapt. Some cameras either use auto iris technology, which is similar to a consumer camera where a small part of the optics is closed in order to protect it from over exposure. For high speed applications the cameras use constant monitoring of the brightness of the image and the on board computer decides if the exposure requires alteration to get a better image,’ Schneider added.

In the EU there is normally internet access nearby which allows the systems to communicate with the traffic controller. This does depend on the infrastructure, but more often than not the installation process is quite simple. This means that drivers can expect to see more and more of these cameras throughout Europe.

Schneider believes ITS is currently the best way to ensure driver and pedestrian safety and that such systems will be used for years to come.

Other tags: 

Related features and analysis & opinion

19 February 2019

Matthew Dale discovers how emerging imaging technologies could make our roads safer and help manage congestion

18 July 2018

Dr Constantinos Bouroussis, a tunnel lighting expert and a research associate at the National Technical University of Athens’ Lighting Laboratory, describes a mobile imaging system developed to inspect tunnel and road lighting