The frame work adopted for the safety of the road transport in Australia notifies the contribution of the road user, vehicle and road environment and infrastructure, and the relation between these components at the time of crashes. Level crossings can also be considered within the same framework. Human Factors is the scientific discipline taking into consideration the needs, abilities and limitations of humans concerned with understanding of the interactions among them and other elements of a system, and is focussed on the design and evaluations of systems.
The safety system approach forms the backbone for human factors. In contrast to the traditional person-based approach which focuses only on the errors made by individual operators with in a system, the safety system approach acknowledges that accidents can be attributed to a combination of active operator-level errors and inadequate, or latent, conditions that reside throughout the system. For systems that involve a human component, consideration must be must be given to the information the human needs to enable them to operate safely and efficiently within that system. The level crossing system must provide enough information to support the human operator to make decisions in order to safely negotiate the crossing (Edquist, J., Stephan, K., Wigglesworth, E., Lenne, M et al., 2009)
Industries such as aviation have been adopting this model for nearly two decades to the extent that it is now entrenched in everyday practice. For instance, the methods used to understand safety failures and to investigate accidents consider the interaction between errors made by individual workers and the contribution of other system, elements such as the task design and work environment, supervision and training and policies under which they operate. In this sense it is recognised that the system must be designed to support human performance, rather than human performance needing to fit within a system.
For systems, that involve a human component, consideration must be given to the information of the human needs to enable them to operate safely and efficiently within that system. The level crossing system must provide enough information to support the human operator to make decisions in order to safely negotiate the crossing. As Pfeiffer (1990) notes, there is less emphasis on providing information to the train driver, as the train has little capacity to stop or slow down to avoid a collision. Thus the emphasis is on providing information to the road users, who have more capacity than the trains to avoid a collision.
The human acting within a system must engage in four main tasks;
- receiving the information via the senses,
- process that information,
- make a decision based the situation, and
- carry out that course of action.
- A failure in any of these tasks leads to errors.
The critical issue is therefore to provide the road-user with adequate information with sufficient time for them to take a decision and act on that. This is where the design of the level crossing environment and infrastructure is important. In particular, the information available in the level crossing environment can support the human to perform first two tasks, by providing them with the necessary information in a way in which they can readily perceive and process that information.
The first step in this process is ensuring the operator is 'fit' to operate within that system. They need the knowledge and skills to be able to use the information provided to them to make decisions and execute an appropriate response. This is the role of driver training (including knowledge of relevant laws) and education and awareness campaigns. A driver who is aware of required actions and has the skills to apply them is thus in a position to use the information provided to safely negotiate the crossing environment.
With regard to the information that the driver needs on approach to level crossings, first, they need to know they are approaching a level crossing and, in particular, the type of level crossing they are approaching. This is because the actions required of the driver differ according to whether the crossing has active protection (active crossing) or no protection (passive crossing). At a passive crossing, there are no specific warnings that notify the driver that a train is approaching, apart from the presence of the train itself and the sounding of the whistle when the driver passes the whistle board. Thus, the road user needs to be aware that they are approaching a passive crossing, so that they can slow down on approach to the crossing and search for approaching trains. This requirement in itself puts the driver at risk of crashing on approach to the crossing as it requires them to take the eyes from the road, in order to execute a thorough search for trains (Edquist, J., Stephan, K., Wigglesworth, E., Lenne, M et al., 2009).
If the driver observes that a train is approaching a passive crossing, they must judge whether or not it is safe to cross. The Road Rules state that a driver must not enter a level crossing if a train approaching the crossing can be seen, or is sounding a warning, and there would be a danger of a collision with the train if the driver entered crossing. Thus, it is up to the driver to judge whether or not they have enough time to cross prior to the arrival of the train, or conversely, whether or not they have enough time to stop before the crossing at the speed which they are travelling. To perform this difficult task, drivers need a sufficient sight distance, down the track, both on approach to, and at the crossing.
In addition, the train itself needs to be conspicuous to be detection. Several authors have noted that a train seen from the front at a distance is a small and not very conspicuous object (Caird, J.K., Creaser, J.I., Edwards, C.J., Dewar, R.E et al., 2002; Cairney, P et al., 2003; Green, M et al., 2002; Mortimer, R.G et al., 1993). It is difficult to see such an object, particularly when it is not expected; and it is more difficult to judge how far away it is and how fast it is approaching (Leibowitz, H.W et al., 1985). Mortimer (1993) notes that road users differ in their ability to judge speed and distance for an approaching train. Many drivers (particularly from those urban areas where passive crossing are rare) may have little experience in this task, and may be uncertain as to the course of action in such an unfamiliar situation. Although this is a similar perceptual task to determining whether to give way to other traffic at a sign posted cross-road, the consequences of an incorrect decision, to proceed are potentially worse. Unlike other traffic, trains have little capacity to stop or avoid a crash, as a train cannot change its direction of travel nor can it slow appreciably with little notice.
At active crossings, the driver is provided with information about the presence of a train via warnings that activate when a train is approaching the crossing. Once the warnings (either flashing lights alone or in combination with boom gates) are activated, drivers are required by law to stop at the crossing. This removes the uncertainty and the onus on the driver to make a decision as to safety of the crossing. Instead, the driver needs only to perceive that the warning is active, and to act according to the road rules and stop at the crossing.
It is worth noting that some drivers choose to intentionally violate the road rules, and to cross if they feel they have time, even if signals are activated. This is one reason why violations are less common at level crossings with flashing lights and boom gates than at crossings with flashing lights alone, as the boom gates provide a physical barrier that make it more difficult to cross.
Although at first glance it appears that the demands on the driver are lower at active crossings than passive crossings, it must be kept in mind that active crossings are generally at locations with much more traffic (rail and road), and therefore the driving environment is more demanding. The driver will be under more pressure from competing task demands, and will have fewer resources to deal with. Thus it is possible that they might fail to perceive that the warnings are active and that a train is approaching the crossing (Wigglesworth, E et al., 1978). It is therefore important that the warnings are salient and designed taking into account the abilities and limitations of the people who will be using them.
In addition to the human factors issues that are specific to level crossings, safety will also be affected by some common issues to road transport in general, such as driver experience and training, current fitness for duty(e.g., fatigue, stress), situation awareness, distractions, and mental models/expectations about the environment and the other objects in it.
As a road user approaches a rail level crossing, there are potentially multiple points of intervention to prevent possible train-vehicle collisions, most of which involve presenting the road user with the appropriate information.
The first is in the knowledge, skills, attitudes, motivations and experiences of road users, which together shape road user's behaviour in the immediate vicinity of the crossing. These can be affected by education campaigns to provide information about what is required at various crossing intersections, and enforcement to bolster intentions to perform these behaviours.
A second intervention opportunity is the speed at which road users (and trains) approach the crossing. Usually must have the time to respond to threat of collisions as trains find little bit difficult to stop or slow down at crossings rather than the road vehicles. In the event of a collision, the severity of injury depends on the relative speeds, of the colliding objects and also on their masses. For both of these reasons, a reduction in speed on approach to the crossing is both a primary countermeasure in itself, and a desired outcome for other countermeasures such as system to alert the road user of crossing/train. To respond appropriately at a crossing, road users must be aware of the existence of the crossing within sufficient time to respond. Road users also need to know exactly where the crossing is.
Once aware of the crossing, the road user needs to know if there is a train or approaching the crossing. It is thus of critical importance that road users are given the opportunity to detect approaching trains. This requires adequate sight distance. Information about the trains can also be provided by the train itself, an in-vehicle system, or infrastructure at the crossing. Where warnings of train presence are provided, these warnings must be configured so that the road user both detects the warning and is able to extract the salient information from it, and sufficiently credible for the road user to believe and obey them. Due to high cost the current warning systems are not practicable for all crossings; however they are low cost alternatives. Where circumstances make it possible that road user who proceed onto the crossing in the absence of a train may be trapped there (e.g. by traffic congestion or a nearby intersection), information can be provided to discourage queuing over the crossing or signals can be integrated.
Some road users may choose to ignore warnings, in which case there are interventions that aim to make it more difficult for the road user to intrude on the crossing while the train is approaching. Finally there are countermeasures which aim to inform the train driver if the crossing is not clear so that the train can slow or stop.
Countermeasures targeting road users
The effect of educations campaigns vary widely depending how they are structured targeting the road users and whether it is supported by other strategies if necessary.
The campaigns which rise fear levels without providing any instructions how to behave at the crossings are not effective. In the US, President Eisenhower's Safe Driving Campaign ran over ten days and promoted the idea of avoiding accidents on a specific day. After the first campaign (for Wednesday December 15, 1954) there were nine fewer deaths on the specified day compared to the previous year, however after the second campaign (for December 1, 1955) there were eight more deaths (Weingroff, R.F et al., 2003). The idea of Safe Driving Day was subsequently abandoned. A similar campaign for Victorian drivers to avoid accidents on a particular weekend in 1978 also failed.
Perhaps the most well publicised campaign for level crossing safety is Operational Lifesaver. This is an educational program that began in Idaho in 1965, and has spread throughout the United States and to other countries. It claims a reduction in crossing accidents of 75% in Canada (Di Tota, D et al., 2008) and the USA (Sramek, H et al., 2008) since its inception. A study of crash rates across 49 states of the US found that the presence of Operation Lifesaver programs in a state correlated with a 15% drop in crashes, and a 19% drop in fatalities (Mok, S., Savage, I et al., 2005). However, the authors note the expansion of Operational Lifesaver coincided with an economic downturn and the deregulation of railroads and trucking industry, so as a reduction in traffic and the number of trains may account for this effect. A later study controlled for the number of trains and the amount of traffic, and studied level of activity for Operational Lifesaver programs rather than just whether Operational Lifesaver programs were present or not. This study found that increasing educational activity reduced collisions, but the effect on fatalities was uncertain (Savage, I et al., 2006). As both of these studies observed historical correlations it is possible that other factors not included in the statistical model could account for this effect.
Investigating the feasibility of Operational Lifesaver for Australia was recommended by a Commonwealth inquiry following a major train smash (Neville, P et al., 2004); the Australian Government gave in-principle support to the recommendation (Truss, W et al., 2005). A report on reducing crashes at passive crossings noted that Operation Lifesaver could be successful in Australia, but materials would need to be adapted (e.g., translating materials into different languages for rural migrant communities (Cairney, P., Gunatillake, T., Wigglesworth, E et al., 2002).
These enforcement cameras identify the drivers who violate the rules at the crossings and thus giving a warnings or fines to them respectively. These types of cameras were already in use at traffic signals and recently they are also used at rail level crossings. A small study in the UK found an unexpected increase in incidents after the installation of cameras, however including both violations and collisions as 'incidents' and the normalising procedure used may have affected these results (Atkins Rail, et al., 2007). A larger study from the US found that cameras produced decreases in crossing violations of 34% to 92% (Carroll, A.A., Warren, J.D et al., 2002). However the details of evaluations at each site reported were not sufficient to assess the quality of the evaluations. A more recent meta-analysis found that photo/video enforcement resulted in a 75% drop in collisions, however again insufficient detail was reported to assess study quality (Park, P.Y.J et al., 2007). A 96-hour demonstration project found no effect of automatic enforcement, but studies need to include sufficient time to show an effect (Fitzpatrick, K., Carlson, P.J., Bartoskewitz, R.T., Bean, J.A et al., 1998). It is not only important to gather the required data to perform a statistical analysis, but time must be allowed for road users to understand that crossing laws are being enforced and to change their behaviour accordingly.
While more data is still necessary, the overall evidence shows that these enforcement cameras have a positive effect at rail road crossings.
Factors to be considered at the time of crossing
Speed is the primary reason of both crash and injury risk. Therefore measures designed to reduce travel and impact speeds are considered separately. The measure described here are those where the primary objective is to reduce vehicle speeds.
The perception, information processing, decision, action loop is a dynamic process; the speed at which a road user is travelling makes an important difference to their ability to drive safely. The faster road user's travel, the less time they have to gather information, process it, make a decision, and execute the required action. Thus reducing speeds can enhance the road user's capability to process information from the environment and respond appropriately. Speed is an important factor for optimising information processing at both active and passive crossings.
Reducing Speed Limits
One method is to reduce the speed limit on the road. Peltola and Pajunen (2006) found that lowering the speed limit from 80 to 60km/hr had worked successfully.
This is an alternative method to reduce the speed by installing traffic calm measures such as chicanes and speed humps to slow down for road users. A Korean study of crash rates (Oh, J., Washington, S.P., Nam, D et al., 2006) noted that crossings with speed humps had fewer crashes. However as this was a co relational rather than experimental study, it is not possible to determine whether the lower crash rate is attributable to the speed humps themselves or to some other factor that these crossings had in common. As the study was performed in Korea it may not transfer to Australian conditions (e.g., high-speed road environments).
A further concern is that these traffic calming measures may just 'move the problem elsewhere' (Nelson, R.C., Clark, M., Lamont, I., Cooksey, A et al., 1996), or result in unintended consequences. Davis (2005) observed vehicles approaching a closed level crossing positioning them around traffic calming islands so that they were stopped on the wrong side of the road, preventing oncoming vehicles from moving through and creating a queue on the crossing.
Sight distance and signs
This is also another factor to be considered at the time of crossing whether visibility is sufficient for the road user to perform the necessary action. There are three types of sight distance at level crossings (Hensley, M.M., Heathington, K.W et al., 1998). The first is the distance to the crossing - visibility needs to be sufficient to detect a train on the crossing, or to detect active traffic control devices. For active crossings this is the only relevant sight distance, assuming that the crossing signals always work perfectly.
For passive crossings there are two further sight distances. The second is the distance to and along the track from where a train might be approaching the crossing in either direction. Once a driver knows there is a crossing ahead, they must be able to see along the track to detect the presence of a train. The third sight distance is that for a driver stopped at the crossing: they need to be able to see whether the train is coming and determine approximate speed. There must be adequate sight distance to allow time to accelerate across before the train reaches the crossing.
The current Australian Standard for railway crossings (AS1742-7, 2007) states that the minimum sight distance 'shall be sufficient for the road vehicle driver stopped at the railway crossing stop line to be able to start off and clear the crossing before the arrival of a previously unseen train'. For crossings to be equipped with 'GIVEWAY' signs, in addition to this first criteria, 'the sight distance shall be sufficient for the road vehicle driver approaching the crossing at the 85th percentile speed to see an approaching train in time and stop at the crossing if necessary. If the sight distance is in adequate, and the crossing is to remain open, active control must be used.
The use of STOP signs at crossings is controversial (Park, P.Y.J., et al., 2007; Yeh, M., Multer, J et al., 2007). Providing instructions as to whether it is safe to continue while looking for trains, or whether it is necessary to stop and look, reduces the burden of decision making on the road user. However, there are concerns that the use of STOP signs at level crossings with very low train frequencies could lead to road users disregarding these signs at level crossings and potentially the entire road network (Russell, E.R., Burnham, A et al., 1999).
Information about the presence of a crossing
At passive crossings, which comprise 72% of Australian crossings (Ford, G., Matthews, A et al., 2002) the signals that indicate a train is approaching (train horn, light or the train outline) are all peripheral and distant. There is also no information provided in the immediate roadside environment, particularly the central ellipse of vision, within which the driver is accustomed to scanning for hazards. The onus is entirely on the road user to perform the appropriate action of searching for an approaching train, and yielding when required. Where the frequency of trains at the crossings is low, road users may expect that a train will not be coming and when they use it, it leads to failure. When the expectancy is violated and train appears, road users may take longer to respond, and even fail to respond.
Alerting the road user about the crossing doesn't totally solve the problem. It is even a possibility that novel treatments to alert road users to crossing presence (for example rumble strips) may distract road users from noticing an approaching train (Raslear, T.G et al., 1996).
Raslear (1996) applied signal detection theory to the level crossing scenario. The problem can be seen as one in which the road user must detect the 'signal' of an approaching train from the noise of other stimuli in the environment, and then make a decision in the face of uncertainty. Countermeasures which increase the signal, such as conspicuous warnings that are only active when a train is approaching, will make this task easier, while countermeasures which increase the noise such as rumble strips which provide auditory/tactile stimulation whether a train is approaching or not, will make this task more difficult. Whether a road user responds to a signal is also dependent on their 'bias to stop', that is, the likelihood that an individual will make the decision to stop, which depends on expectancy and motivations. Increasing the bias to stop will also reduce crashes.
Therefore alerting the road user about the presence of a crossing is unlikely to be sufficient to avoid all collisions. The optical approach is to provide some form of active warning of train approach for road users at all crossings.
Alerting the road user
Providing advance warning of the existence of a crossing can assist road users approaching both active and passive crossings, as they will be primed to search for warnings or trains. One desired outcome from alerting a driver to the presence of a crossing is a reduction in speed; another is increased looking behaviour, although this is more difficult to observe.
Lerner, Llaneras, McGee and Stephens (2002) summarised the information required for road users to safely interact with passive crossings:
- That there is a rail-road crossing ahead;
- That the crossing is not protected by bells, lights, or gates (i.e., is a passive crossing) and, therefore, it will be up to the road user to determine whether the train is at or in proximity of the crossing;
- What actions are required of the road user in approaching and traversing the crossing (i.e., maintain speed or slow down, look for trains, and possibly stop); and
- If appropriate, that there is some special condition or situation at the crossing (e.g., a skewed crossing, limited sight distance, a humped crossing, or high-speed trains) that requires more road user attention and will influence the action described in third item above.
The Australian Standard specifies that 'RAIL X' should be marked on the road in advance of the crossing on all high-speed approaches of adequate seal width (AS1742-7, et al., 2007). Richards and Heathington (1988) found that 72% of drivers could correctly identify the 'RAIL X' pavement markings used in advance of crossings, which is somewhat better than the 64% who could identify railroad advance warning signs. Tierney (1991) cites an Australian study which found that all of the unspecified number of interviewees could indentify 'RAIL X' pavement marking. A review of seven American studies found that pavement markings reduced crash rates by 21%, however no details were given on the type of markings included in the review (Park, P.Y.J et al., 2007), and the studies included road intersections as well as rail-road crossings.
As the drivers have to slow down at crossing and look whether a train is coming or not, they must have a proper vision of the crossing and decide to take a necessary action. This is will be significant particularly at nights. A national survey of 4400 Australian drivers found that approximately 20% had travelled over a level crossing without noticing it until they had crossed (Sochon, P et al., 2008). Passive crossings in Australia are marked by sign assembly RX-1 or RX-2, which include a cross buck, the number of tracks, and GIVE WAY (RX-1) or STOP (RX-2) sign depending on the available sight distance at the crossing, and with the appropriate line marked on the pavement (AS1742-7, et al., 2007).
The Australian Standard (AS1742-7, et al., 2007) specifies that all signs used in the immediate vicinity of crossings shall be reflectorised.
An unpublished study cited in Russell (1992) found that reflectorised posts were effective in increasing the conspicuity of the crossing. Later field tests of reflectorised signs and posts together with roadside delineators showed that these devices resulted in drivers slowing down and looking for trains (Russell, E.R., Cathcart, A et al., 1998).
In vehicle warnings
These warnings are the new ones which indicate the presence of a train that does not rely on potentially expensive infrastructure changes. There have been a number of small-scale trials of in-vehicle systems that provide warnings of an approaching train (Carroll, A.A., Passera, A., Tingos, I et al., 2001; Osemenam, B., Bahlen, S et al., 1998; Smailes, J.A., Carroll, A.A., Anderson, J.F et al., 2002; U.S. Department of Transportation, et al., 2001, 2007). As these systems are achievable, there is little evidence regarding their contribution at level crossing crashes. The evidence is that these systems can warn or threat the driver 'in advance' potentially thus reducing the crashes (Lenne, M.G., Mulvihill, C., Triggs, T., Regan, M., Corben, B et al., 2008). In the Lenne et al., (2008) study, the "threat" was an emergency vehicle, but the results could just as easily be transferable to trains. These warning systems are generally useful for heavy vehicles which take time for decision and also vehicles which carry hazardous materials or substances.
Alternative warning systems
As the mentioned above warning systems appears to be costly is it is not possible to establish them at all crossings. There is also another way to detect the trains like using of low-cost warning systems, audible or manual warning.
Low-cost train detector system
These systems are generally used to detect a train where there is no association of human factors issues in itself. These systems are generally used where the standard system is not practical, thus improving the level of information to the road user.
A variety of low-cost detection systems exist, including
- inductive wheel sensors,
- Doppler radar detector;
- geophone (picks up acoustic wave from train propagated through earth);
- fibre-optic cable bonded to rail;
- buried fibre-optic cable with high-power laser;
- video cameras;
- speed radar;
- speed and distance radar;
- acoustic detection;
- pressure sensor on rail;
- magnetic sensor;
- active infrared (emits electromagnetic radiation and detects reflections);
- passive infrared ( detects infrared energy from train);
- LIDAR (light radar);
- Wheel counting.
Carroll, Gordon, Reiff and Gage (2002) conducted tests of three train detection systems that used magnetic anomaly and vibration, or wheel sensors. Out of all wheel counting performed best.
Nookala and Melby (2005) conducted a test on GPS based detection system by applying them at the existing signals. Their results showed that there were no activation failures and they are successful for heavy vehicles by providing adequate information about the presence of train.
Roop, Roco, Olson and Zimmer (2005) ranked all of the above systems, except for wheel counting, by taking in to consideration of various objectives such as safe, cost, reliability, installation, maintenance and how it enriches the railroad property. They found that probably it's not possible to install these systems instead of current warning systems for lower cost, but at the same time it is possible to get cheaper systems that are safer than passive crossings.
Induction loop train detection systems have been tried in Victoria (Creswick) and South Australia (Monarto) under real train conditions, but without signals being visible to the public (Jordan, P et al., 2006).
Automated wayside horn
Wayside horns can be used where train horn is a problem. Traditionally a 'whistle board' is placed at a certain distance from the crossing, and the train drivers blow the horn making to give road users sufficient warning. With a wayside horn, the train driver or an automated system sends a signal to the wayside horn when the train reaches the whistle board, so that same amount of warning is given. The advantage is that wayside horns can be directed more accurately towards vehicles approaching the crossing, and therefore can be lower in volume and less disrupting the rest of the neighbourhood. There is also a counter design in this type of system such as if the signal from wayside horn is failed, the train driver can detect and can blow the train horn.
Wayside horns do not provide any information on the direction of the approaching train, and thus do not provide any clues to drivers attempting to localise the train. However the main purpose of horns is to warn the road users at the crossing. But there are no studies found about the driver's response to these detection systems.
Countermeasures for pedestrian safety
Pedestrians represent about half of the injuries and fatalities due to level crossing accidents (Berry, J.G., Harrison, J et al., 2008; Hoy, M et al., 2005). In pedestrian-train crashes the pedestrians are worst affected.
They may cross unsafely for a variety of reasons (Pitsopoulos, J et al., 2007). It varies from age to age where the young people find difficult to decide the required action, youth and males take risk such as racing, and old people need more time to cross, thus making critical. Similarly like vehicle drivers failed to see the activated warning systems not expecting a train and they assume that if a train has crossed or passed they don't wait and see for the second train, they just simply cross while the warnings are active.
There is also a problem with the disabled people.
New standards for disability access at pedestrian crossings were introduced in 2002. These include 1:14 gradient approach ramps, wider escape areas and emergency gates to accommodate Gophers/Scooter mobility devices, and changes to the crossing surface (Spicer, T et al., 2008).
The first priority to ensure pedestrian safety is to investigate the causes or reasons which are responsible for these crashes and they have to be implemented in the design process.
From this data there is a possibility of determining the following research.
- The effectiveness of second train warnings;
- Crossings where there are more pedestrian violations because of long waiting;
- The effect of recent education campaign whether it is successful in developing the awareness of the people.
The simplest solution is to erect barriers or over bridges.
Strategies that enable choice of other countermeasures
Operational methods for assessing the risk
The best method to choose at crossings is first to assess the risk and then determine the factors contributing it and select the proper countermeasure. There are about seven different approaches to assess the risk. The United Kingdom Rail Safety and Standards Board commissioned a report that examined approaches to the problem worldwide (Little, A.D et al., 2007b). The report describes 23 models from twelve countries, and groups them into four categories:
- Parameter Gate approaches (India, Japan, Russia, Spain and Sweden) generate the appropriate control type/treatment from the parameters given without calculating risk;
- Simple Weighted Factor Models (Australia, Northern Ireland, and New Zealand) weight parameters by their relative contribution to risk.
- Complex Weighted Factor Models (Great Britain, Ireland) use more complicated algorithms such as fault/event trees or simulation to derive parameter weightings; and
- Statistically Driven approaches (Australia, Canada, New Zealand, and USA) use statistical analysis of accident history to determine correlations between parameters and accident occurrence.
In Australia, the relevant Standard (AS1742-7, et al., 2007) specifies what signage is required once the type of control at a level crossing has been determined, but does not give any guidance as to how to determine what type of control is required. Instead it refers to 'risk assessment models such as ALCAM'.
ALCAM is the Australian Level Crossing Assessment Model. It comes under Simple Weighted Factor Models. This is first introduced by Queensland Government in 1999 for risk analysis. It assesses the characteristics of a particular crossing which has impact on the occurrence of accidents. The main advantage in this method is the identification of the source of risk, thus providing the most cost-effective solution. Saccomanno, Park and Fu (2007) noted that the effect of a particular countermeasure can vary depending on the characteristics of the crossing at which it is to be implemented, and therefore modelling of countermeasure effects should take into account the characteristics of the specific crossing rather than assuming a similar effect at all crossings.
National adoption of this system was recommended at the 2002 Symposium (Ford, G., Matthews, A et al., 2002). Still after that the refining and updating of the model has continued. Use of this system has been urged by several government inquiries into level crossing safety (Neville, P et al., 2004; STAYSAFE Committee, et al., 2004; Truss, W et al., 2005). Victorian implementation began with field surveys of crossings in 2006-07.
The main disadvantage of this model is it does not include all the factors causing crashes. For example, ALCAM assessed a level crossing at Kerang as 'needing no attention', before a fatal crash at the crossing in June 2007 (Ross, A et al., 2008).
And also this model does not take in to account of the road users expected behaviour which might affect the characteristics of the similar crossings in same route.
There is a lack of data about the crashes at level crossings. In Victoria, Public Transport Safety keeps the database of accidents but is not accessible to public. The Department of Transport contains the number of level crossings but these are not related to the accidents data.
The main problem with this ALCAM is it doesn't record the crash history thus making it very difficult to determine the common issues at crossings that have a higher collision rate.