Identifying new concepts for innovative lighting-based interventions to influence movement and behaviours in train stations

The disorderly and disrupted movement of passengers within train stations are key concerns in rail transport, especially where there are increasing numbers of passengers, coupled with often out-dated, adapted station spaces. With careful planning and design, different characteristics of lighting can be employed to address problems relating to the movement and behaviour of passengers in railway environments. This study aims to offer an approach to identify new concepts for lighting-based interventions to influence passenger movement behaviours within train stations. Behaviourally orientated lighting literature was reviewed, providing the knowledge base to inform a series of engagement activities with transport stakeholders and lighting technologists, to understand problematic behaviours and how these might be resolved through targeted lighting design. In combining findings from the literature with insights from rail and transport-related industry stakeholders and lighting specialists, a number of potential opportunities for novel applications of lighting have been identified. Six scenarios are developed that illustrate these opportunities for potential lighting-based interventions to influence train passenger movement and behaviour. These scenarios can be used to inform the direction of further research and consideration of how different lighting characteristics can affect rail passenger behaviours.


Introduction
Lighting is an important factor in the design of any space occupied or accessible to people. However, the focus, when designing new lighting or assessing existing lighting, is typically performance related; for example, is there adequate lighting to support visual acuity and to enable a task or activity to be carried out effectively (i.e. visual and task performance 1 )? Health-and safety-related concerns are also relevant, from the potentially harmful effects of flicker (e.g. generalised malaise to more serious epileptic seizures 2 ), through to its contributory role in slips, trips and falls, particularly for the elderly who may suffer with reduced vision. 3 Such performance and health-and safetyrelated drivers are themselves supported by a number of lighting standards and guidelines (e.g. Literatures 4,5 ). These offer minimum or recommended lighting levels, primarily illuminance and luminance values, for performing a particular task or activity in a particular environment (often indoor).
While such objective requirements in lighting are critical, the role of more aesthetically driven, experience-based lighting is also recognised. So and Leung 6 referred to the design of lighting as 'a piece of work related to science and art' (p. 113), while Manav and Yener 7 comment on the aesthetical and emotional reactions that arise when observing a lighting system. Lighting design for the creation of atmosphere to encourage engagement and engender positive responses is of growing interest. Indeed, there is emerging research into both the psychological/behavioural effects and the physiological effects of light (including photobiology, human-centric lighting 8 ). de Kort and Veitch 9 argue that all classes of effects based on both non-image forming light (influencing circadian and acute effects) and image forming (influencing visual performance, visual experience and visual comfort) ultimately affect psychological functioning, including behaviours and experiences. However, with the exception of some specific environments/contexts, e.g. retail, 10 hospitals, 11 office spaces, 12 and the use of dynamically operated arrow-like stimuli in steering festival crowds, 13 there are few applications of lighting to influence behaviour and/or mood within real-world contexts. Multiple lighting characteristics (intensity, colour, flicker/flash rate) have typically been manipulated and examined in indoor laboratories or virtual reality scenarios and often with singular participants. Thus, ecological validity and generalisation of this research area to real-world contexts are challenged.
Aside from the recent studies on the use of blue lighting in suicide prevention [14][15][16] and emergency egress in road 17 and rail tunnels, 18 there still remains relatively little guidance relating to the applications of lighting in influencing behaviour in transport contexts, and more specifically rail locations, despite a number of behaviour-related challenges that lighting could potentially address. For example, there is evidence that changes in lighting can have impacts on walking speed, 19 mood and levels of activation 20 and route selection. 21,22 The movement and flow of passengers within train stations offer many challenges within the rail industry in Great Britain. Rail stations exist in a variety of sizes, with many comprising a mix of more modern and older or 'heritage' buildings, indoor and outdoor spaces (including canopies) and operating day and night, with significant fluctuations (e.g. at peak/off-peak travel times) and overall increases in passenger numbers. Such dynamic demands, combined with an often out-dated infrastructure, have served to create problems for passenger movement and flow, with crowding on platforms, bunching at intersections and blockages on stairways and bridges at different times of day.
Improvements in this area have the potential to deliver benefits of £258 m, by increasing capacity, enhancing customer experience and efficiency savings. 23 While these problems have often been addressed using physical interventions (e.g. barriers, use of loudhailers and messages via technological Customer Information Systems), these solutions are expensive, difficult to install and are often implemented without further evaluation to assess their relative impact or effectiveness. 24,25 This paper considers the prospect of using different characteristics of lighting to influence the movement and behaviours of passengers in a rail context, as a dynamic alternative to physical solutions.
The current research assumes that dynamic solutions to passenger movement issues could be realised with better integration and control of lighting (e.g. using new LED-based lighting, connected to Internet of Things control systems). The paper aims to offer a review of content from the literature and consultation with rail-and transport-related industry and lighting specialists, as part of a scenario-based approach to identify new concepts for innovative lighting in influencing passenger movement within train stations. The paper determines the current thinking in terms of research on lighting for this type of context, considering this in relation to the issues and experiences described by a wide range of aligned stakeholders. The added accessibility and applicability of the research findings from this study will be valuable to designers (lighting and infrastructure) and operational staff within rail and other transport or service industries.

Literature review
Thirty-six research papers and three review papers [26][27][28] have been reviewed to explore how lighting principles can be used to influence movement behaviours in stations. The literature within the three review papers, compiled over a series of studies that explored the potential for use of innovative lighting in rail transport and retail situations, was used as a starting point. The source materials within these were identified using several research databases, including NUSEARCH (hosted at the University of Nottingham), Scopus, Web of Science and Google Scholar. Search terms included combinations of the light* and the following contextual termspublic space; crowd control, behaviour, wayfinding, accessibility, affordance; dynamic; smart; mood; affect; feeling; emotion; atmosphere; visual and colour. To determine the suitability for this study, abstracts were read to identify relevance to the study aims. To be included, the paper needed to report on one or more characteristics of lighting or similar visual cues on aspects of behaviour, movements, performance or mood. The papers covered studies in laboratory situations, in virtual reality and real-world studies. Further searches were carried out to look for recent publications, in particular how lighting might influence general wayfinding, crowd behaviour and the design of architectural spaces. 29 These searches used similar keywords and the search terms linked to the main characteristics of lighting (e.g. intensity, colour, movement) that were emerging during the course of the review. Reference lists were searched and citation searches were conducted for selected papers. No dates were used to restrict the search for publications. The review focused primarily on the likely psychological/behavioural effects of lighting (as opposed to physiological), although as indicated earlier, there was a degree of overlap in how these have been considered and reported in previous work.
The review takes a diverse view of the literature, in order to appreciate how different characteristics of lighting can influence and impact behaviour, to establish particular lighting values (where specified) and clarify resultant behaviours. However, it is evident that the research literature in this area is not only dispersed but also often emergent, making it somewhat difficult to identify relevant sources. This is clearly demonstrated by de Kort and Veitch 9 who, following an initial search of The Journal of Environmental Psychology, as part of developing a special issue on light, lighting and behaviour, found 'only 6 papers with ''light'' or ''lighting'' as a keyword; in contrast, 281 used 'noise' or ''sound'''. Findings can also be inconclusive or conflicting, a possible symptom of the complex relationship between light and effect. 30 This complexity has been further compounded by the multifaceted nature of many lighting studies, which often include multiple lighting characteristics in different combinations, examined in relation to a diversity of outcomes (e.g. improved path detection through to increases in product buying) and within a range of contexts. In order to provide some clarity, this review is organised according to five key characteristics of lighting that have been identified from the literature: intensity; colour; flicker/flash rate; position/direction and source/operation. While there are potential overlaps, these characteristics offer useful qualities that can be referred to and potentially applied in the development of innovative, lighting installations within a transport context. Appendix 1 (uploaded as supplementary data) provides brief summaries of the key literature referred to in eliciting these lighting characteristics.

Intensity
Changing the intensity of lighting may influence the speed and choice of movement, e.g. direction/route and also the mood or subjective well-being of people. Pedersen and Johansson 19 examined the effects of reduced lighting levels upon walking speeds, reporting that participants walked significantly slower under dimmed (60%, 40% and 20% of maximum illuminance) lab-based, street lighting and the effect continued when such levels were increased back to 100%. People may not only be hesitant under reduced visibility, but also display generalised uncertainty in response to varying lighting levels. Contradictory results have, however, been produced, where darker pathways induce quicker walking speeds 31 and have a significant impact upon the decision to walk, 32 both of which can be largely attributed to reduced perceptions of personal safety. Furthermore, increases in sensation seeking have been observed under lower compared to higher illuminance. 33 Nevertheless, in station environments, where rushing and running are concerns, the use of lower lighting levels to induce slower walking speeds could be considered, although a minimum overall lighting level must be maintained for overall safety and performance (e.g. legibility of signage).
Increasing lighting levels in defined spaces (e.g. entrances to alternative tunnels/corridors) can impact on wayfinding and route selection. In navigation studies in virtual environments (e.g. an emergency evacuation scenario 21 and a hotel 22 ), although signage was generally found to be the strongest indicator of choice, a brighter, narrower (as opposed to a darker, wider) corridor was typically chosen by participants: a brighter, wider corridor (1.5 -m optimum), afforded 100% selection. 21 Within a train station context, this effect could be applied to encourage choice of an alternative or optimised route.
Changes in lighting levels have also been shown to support improvements in mood and subjective well-being by mimicking the effects of day-light, inducing levels of natural activation and relaxation. 34 Leppa¨ma¨ki et al. 35 reported on the benefits to (female) shift workers of brief, bright exposures to light (2500-4000 lx), combined with exercise. Hoffmann et al. 20 found that subjective mood could be improved in an office environment by constantly (over a 30-minute period) but imperceptibly changing illuminance (500-1800 lx) and correlated colour temperature (CCT) (4000-6500 k). Varying lighting levels in this way, over shorter periods or across a working day, could potentially benefit passengers having to wait for extended periods.

Colour
Colour within lighting broadly relates to either warmer or cooler white light, indicated by its CCT or specific colour hues, existing as pure monochromatic wavelengths or created by combining different wavelengths of light. As mentioned previously, there is a potential to influence activity levels and mood through differences in CCT. 20 Warm, white light Lighting to influence movement in train stations 979 (2700 k) is generally reported as more relaxing, approachable and pleasurable and as such often used in retail environments. 36 Alternatively, neutral CCT (4000 k) is preferable for orientation and wayfinding tasks, 37 whereas cool white light (5000 k) can support more attentive, focused behaviours. However, in a major study on the influence of warm, cool and artificial daylight lighting on selfreported mood and cognitive performance, there was no direct effect found of warm lighting on mood, though participants performed better in terms of short-term memory and problem solving. 38 Moreover, a warm CCT was found to increase levels of hostility compared to cool CCT. 33 While there is certainly potential to use CCT to create more relaxing versus engaging spaces within a station (e.g. ticket halls), further evidence is needed.
Changes in hue have also been widely examined. Elliot et al. 39 explained the notion of paired associations, i.e. 'both explicit and subtle pairings between colours and particular messages, concepts and experiences' (p. 155), which develop from infancy and directly influence effect, cognition and behaviours (e.g. red writing on research materials evoked avoidance motivation thus undermining performance on achievement tasks 39 ). However, understanding the meanings that people from different cultures attach to particular coloured lighting, in a particular context, is crucial in exploiting these colours effectively. Red lighting is thought to be more activating leading to impulsive behaviour, 40 aligning with its effectiveness in hazard warning signals 41 (see below). Green lighting (which is suggested to be less anxiety provoking than red illumination 42 ) has also been studied in relation to emergency egress, showing the positive effects of green, floor level LED track lighting upon walking speeds and movement paths in a VR-based rail tunnel evacuation. 18 Green (or white) flashing lights performed better than blue in promoting emergency exit usage. 17 Barlı et al. 43 demonstrated the positive impact of green lighting upon time spent in a retail store and subsequent increased product purchases. The effects of pulsating orange light as a distractor in stress recovery were inconclusive. 44 Finally, blue lighting has afforded considerable interest in recent years, for its alleged positive role in suicide prevention at rail stations 14,15 (though this was questioned by Ichikawa et al. 16 ). Blue is often described or associated with being a calming colour, 11,42 although its effect upon the circadian system of supressing the production of the hormone melatonin is likely to induce a more active, engaged and aware state over longer periods. 45 Chellappa et al. 46 have reported enhanced subjective alertness, well-being and faster reaction times in tasks requiring sustained attention in a cool, white, blue-shifted light (40 lx, 6500 K).
The use of coloured lighting within train stations to enhance passengers' levels of awareness, to support changes in mood or to influence the direction of movement has obvious potential. However, at a practical level, careful consideration is needed so as not to conflict with existing use of coloured light signals in rail operations.

Flicker/flash rate
Flicker/flash rate is typically examined in conjunction with intensity and/or colour. Indeed, flash rates (where the pulse duration is less than the inter-pulse interval length 47 ) have been widely researched in relation to emergency and safety contexts. Although values can vary, an optimum flash rate of 1-4 Hz is considered most effective in emergency exits 17 and a rate of 240 fpm (4 Hz) for conveying a sense of hazard. 41 While flashing could be effective in the current context for attracting or directing attention and influencing movement, it would be important not to conflict with existing emergency lighting and should also be considered in conjunction with other modes of communication, e.g. audio messages and alerts.
Flicker can also be used to induce positive effects in behaviour, such as heightened levels of activation when exposed to a high and imperceptible flicker rate of 32,000 Hz. 48 However, there are also some notable, negative health impacts of flicker. These include headaches and impaired vision from longer exposures to invisible flicker (typically below 165 Hz), through to photosensitive epileptic seizures incurred from brief exposures to visible flicker -typically between 3 Hz and 70 Hz, the greatest risk being between 15 Hz and 20 Hz. 2 Flicker can also have a significant and often detrimental effect on those with sensory sensitivities and neurodevelopmental conditions, such as autism, with bright or flickering lights often being painful 49 and potentially exacerbating repetitive behaviours. 50 Therefore, while a high, undetectable flicker rate could be effective in supporting levels of activation and awareness on a platform, rates need to be designed carefully and monitored to avoid any detrimental effects.
Linearly configured, sequential flashes can be used to indicate movement or direction. Cosma et al. 18 examined the use of continuous versus chasing (green) LED track lighting to support emergency egress, and while little significant difference was found between the two upon walking times, in comparison with no lighting intervention, participants tended to walk closer to the lights. In the absence of further research, it may still be useful to consider the use of chasing or travelling lights to indicate directional paths and also potentially influence pace, 51 particularly as such lights are increasingly used in other contexts (e.g. lane closures on highways).

Position/direction
Arrow-like lights influenced crowd decision making when walking through a walkway with a central obstacle, resulting in 27% more pedestrians selecting the indicated side than the non-indicated side of the obstacle. 13 Floor-mounted lighting tracks were also found to guide movement paths. 18 Wright et al. 52 reported wall-mounted lighting tracks (1000 mm above floor level), induced quicker walking speeds than standard high-mounted emergency lighting. Conversely, for older people walking along dimly lit corridors, ceiling mounted lights offering diffuse illuminance (as opposed to plug-in night lights or night lights with path edge lighting), best supported walking speeds and gait. This said, edge lighting did contribute to supporting movement paths by offering laser lines to follow. 53 The directionality of lighting is key within retail, particularly when combined with intensity and colour, to direct and maintain attention. 54 Overall, despite limited guidance on specific or optimum effects, the position and direction of lighting is crucial for maintaining overall visual acuity and performance levels, while also creating a focus. Moreover, the physical environment in which lighting is situated also plays a crucial role in influencing people's understanding and perception of a space and subsequent engagement with it, i.e. naturally affording certain behaviours. 55 Indeed, in a study of emergency egress from a lab-based, smoke filled tunnel, 56 91% of participants were reported to navigate using the physical walls (in addition to lit exit signs), for 75% of the 200 -m distance.

Source/operation
There have been several studies on the impact upon mood and performance of individuals having control of the lighting in their workplace. LED lighting offers increased efficiencies, a variety of visual effects and greater control in relation to more traditional lighting technologies. 10 Veitch and Newsham 57 found no significant improvements in mood or performance with additional control, while Moore et al. 12 concluded that although people preferred to Lighting to influence movement in train stations 981 have more influence, offices without such controls were more effectively lit (in terms of task illuminance/luminance levels). Despite such ambivalent results, lighting that is felt by passengers to be more directly responsive to their actions and the changing situation within a station could offer a more meaningful and engaging experience. Indeed, there is a growing body of research into the impact of interactive lighting installations in public, urban spaces, e.g. Alwani et al. 58 and Raby et al. 59 which suggest responses and feedback that are perceived to map directly to our actions, encourage our overall sense of engagement, empowerment and agency. 60

Conclusions from the literature
The relationship between light and behavioural effect is complex, 30 with research in the area sometimes being contradictory and inconclusive, making it difficult to generalise to wider contexts. Indeed, specifying a definitive set of lighting variables that can be applied to affect a particular behaviour or response in a situation may be challenging. However, using the five lighting characteristics as a working guide for researchers and lighting designers, it is possible to distinguish effects that could potentially translate to the rail context. A stakeholder engagement activity was carried out to establish the specific industry-based, behavioural and operational contexts, which might be relevant to the implementation of lighting-based interventions.

Methods
An investigation was carried out into the potential effects of different characteristics of lighting on behaviours, and the extent to which these might be employed to address problematic behaviours in train stations. This section outlines the methods applied in consulting with a wide range of rail and lighting stakeholders, to identify, contextualise and prioritise current issues and problematic behaviours associated with passenger movements in train stations. Insights from the literature review and subsequent stakeholder consultation activities are then combined, leading to the identification of a series of key scenarios to guide and inform the development of innovative lighting within train stations.

Stakeholder consultation
Specific engagement activities (interviews, a workshop and site visits) were carried out with 25 rail-and transport-related stakeholders, with a diversity of roles/backgrounds including station managers, safety and security officers, customer experience leads and building service engineers both from individual train operating companies (TOCs) and network rail (NR), in addition to academics in transport/userresearch and an officer from the British Transport Police (BTP). These activities were designed to draw upon the expert knowledge and experience of stakeholders, to understand the issues that impact movement and passenger behaviour in train stations. The intention was to identify a range of examples of problematic locations and associated behaviours, identify existing interventions to minimise such problems and explore opportunities where innovative lighting might be applied in further addressing these issues. The overall approach (particularly towards the workshop and site visits) was collaborative, with rail/transport specialists interfacing with lighting technologists and academic researchers (Human Factors) and served to generate a body of primarily qualitative data.

Stakeholder interviews
Eighteen one-hour telephone interviews were first conducted, to identify stakeholders' experiences. Interviews included questioning relating to job role, associated duties and length of service, with a Critical Incident Technique 61 used to elicit participants' personal experiences and observations of both more positive and also negative instances of movement in stations. Using this technique enabled participants to focus more directly on their own experiences, while also reflecting upon their relative meaning and significance. Participants were also asked to consider why, where and to what extent such movementrelated problems occur and how these are currently being addressed, in conjunction with any potential barriers to success. Finally, participants were asked to imagine how lighting might be used to better manage passenger movement and behaviours, what this might look like and any potential issues that may arise in implementation. The interviews also served to frame topics for discussion in the second engagement activity, i.e. a one-day workshop.

Stakeholder workshop
A one-day workshop was conducted involving 12 stakeholders (5 of whom had participated in the telephone interviews). Each participant completed a questionnaire, similar in format to the interview, capturing their job role, better versus worse experiences of moving around train stations and also a ranking of their top three movement-related problems in these contexts. A group discussion was used to highlight significant movement-related issues, and researchers recorded key points on a white board so that these were visible to all involved. Participants were then split into smaller groups (of three) and asked to consider the issues raised more specifically, using a set of preprepared resources. These included images (i.e. a station plan and photographs) of a sample station, which were used as visual prompts (i.e. Schock et al. 62 ) to encourage participants to elaborate on movement issues commonly linked to a physical location or feature. Participants annotated the images with their thoughts and ideas. A series of general questions helped to guide the participants through this exercise, covering questions including 'what are the main contributors/causes of movement problems in a train station?', 'where should a new lighting product be located for maximum effectiveness?' and 'what are the risks associated with such innovative lighting?' This activity was followed by a presentation of potential lighting products and an explanation of different lighting characteristics (drawn from the literature review). Participants were then requested to reappraise their annotated images from the earlier activity in light of this information. Finally, participants were asked to re-rank their top three movement-related problems, taking account of the likely severity of the problem, the feasibility of implementing a lighting solution and the potential impact of a lighting solution. Rankings were discussed across the whole group and movement-related problems subsequently prioritised in conjunction with potential recommendations for lighting solutions.

Site visits
Three site visits to a trial train station were conducted by both the researchers and lighting designers/technologists over a five-month period (spring/summer), during peak and offpeak hours. The visits offered the opportunity to take photographs of the infrastructure, and to envision how people might move around different parts of the station (e.g. entrance gates, stairways, bridges, walkways and platforms) and better understand the context of some of the key movement issues raised in the engagement activities. These visits also enabled identification of in situ opportunities for the use of innovative lighting to address these problems. The academic-researchers and designer-technologists were able to develop a situated, shared understanding of the issues and also begin to evolve a common language by which they could communicate as the project progressed.

Data analysis synthesis
The study activities served to produce a rich body of data. These included the Lighting to influence movement in train stations 983 following: outputs from the literature review; photographs from the site visits; and descriptive (textual) data from the various engagement activities (i.e. interviews, workshop and site visits), including researcher notes, questionnaire responses, annotated station plans/ images and problem/solution ranking grids. These data were subsequently analysed in three key stages. First, a general review of the interview and questionnaire responses was undertaken and a number of generic issues were identified, broadly aligning to questioning relating to movement problems, existing solutions and lighting opportunities in train stations. A conventional content analysis 63 was then applied to the data as a whole, i.e. including researcher and participant notes. This enabled categories to emerge and evolve directly and inductively from within the data. Indeed, through an iterative process of collating, comparing, coding/classifying and reflecting upon the data, an initial set of largely location-driven categories were developed. These included relevant location/physical feature, associated behavioural issue, solution required and opportunities for lighting, with the latter category being directly supported by evidence derived from the literature review. However, during this process, it became evident that while participants' insights and experiences were often intrinsic to their location, listed against each of these were a number of distinct yet commonly reoccurring behaviours (e.g. waiting, decision-making). The final stage of analysis, therefore, involved re-examining the data in relation to key behaviours, any of which could occur in a variety of locations at any one time (see Table 1, uploaded as supplementary data).

Results
A set of behaviourally driven categories emerged including the following: observable problematic behaviours; required solution(s); and potential lighting opportunities (see Table 1 in supplementary data). Based upon these categories and through further guidance from the lighting technologists, a number of behaviourally driven scenarios were then developed, which served to inform the subsequent development of the lighting-based interventions. These scenarios, based on theory (literature) and practice (stakeholder consultations and industry site visits) are outlined in Table 2 (uploaded as supplementary data).
Combining the findings from both the literature review and the stakeholder engagement activities, enabled a number of proposals to evolve for how innovative lighting could be applied to influence particular movement and behaviours in a train station context. These proposals were further guided by advice from stakeholders (particularly lighting technologists) on the practicability/ effectiveness of particular lighting within a particular scenario. The six scenarios and corresponding proposals for lighting interventions are detailed below (Table 2, uploaded as supplementary data).

Discussion
A review of the behaviourally orientated lighting literature was initially conducted, which served to inform a series of engagement activities with transport (including rail) and lighting stakeholders, to elicit and understand current movement and behaviour related issues in train stations and existing approaches for addressing these. These activities offered a valuable knowledge base to identify generic lighting effects, e.g. increasing the intensity of light in narrower versus wider corridors to influence route selection, 21,22 that could be applied to influence potentially problematic passenger behaviours. The knowledge of typical behaviours in stations and the understanding of these lighting effects were then considered in the development of the particular scenarios and proposals for innovative lighting solutions that need to be evaluated in future studies. In this paper, the potential to apply these effects has been considered in relation to a series of typical railway station settings, drawing on evidence from the literature and a series of consultation exercises with railway/lighting stakeholders. Such interventions have the potential to support movement behaviours in more intuitive, responsive and adaptive ways, while also avoiding another immutable layer of visual information and clutter from additional signage.
Although not specifically a focus of this paper, it would be useful to understand the benefits of lighting interventions over and above existing interventions, such as physical measures or increased staff support. It could be argued that the dynamic nature of many lighting products, along with their ability to convey a level of semantic meaning 64 in a simplified manner (compared with, for instance, signage) provides some of these benefits. The lighting interventions discussed here can be applied with some flexibility (i.e. with the ability to reposition, dim, turn on/off as required) and in a non-intrusive way (situated largely overhead), integrating effectively within the many physical, operational and financial constraints of a station environment. Moreover, they do not add to the visual clutter that is already apparent in railway stations. For example, not knowing where to wait (see Scenarios 1 and 2), particularly on a platform prior to boarding a train, can cause blockages in flow, potential uncertainty or stress for passengers and increased boarding and alighting times. Using lighting that can adapt to different lengths and configurations of trains, to indicate where passengers should congregate for access to a particular carriage, would mitigate these issues and encourage better distribution of passengers along a platform, without the need for additional signage, staff intervention or audio announcements. Similarly, knowing which direction to take having alighted a train (see Scenario 3) can help to reduce blockages arising from uncertainty and indecision. Typically, wayfinding information is displayed as textual signs, which may exclude people who are unfamiliar with the language, cannot see the signage or, for whatever reason, cannot comprehend the signage. Using lighting to indicate the optimum route to stairs, lifts and exits can help passengers to disperse quickly from the platform and maintain a continuous flow, without reliance on passengers being able to read/interpret text. Moreover, bunching, blockages or rushing in confined spaces such as stairways and bridges (see Scenario 4) can have serious safety implications (e.g. slips, trips and falls) and cause stress and frustration to other passengers. Using responsive lighting to separate passengers in these areas (e.g. up/ down; left/right and regulate flows and movement speeds), could help to reduce such dangers and conflicts, without the need for static, physical directions or for staff to direct passengers.
While these scenarios offer a potentially practicable approach for introducing innovative lighting to address movement-related issues in train stations, particularly in the absence of a clear consensus within the literature and few credible, real-world examples, there are certain limitations of the study which should be acknowledged. A search of the academic literature relating to lighting and behaviour has been conducted, though it is acknowledged that there is limited evidence from which to draw robust conclusions regarding the potential of lighting characteristics to affect behaviour. For this reason, wellconsidered, field-based evaluation studies of lighting interventions are needed to expand the body of knowledge in this area. Moreover, while a detailed review is outside of the scope of this paper, it is crucial to conduct an examination of evidence relating to crowd movements (e.g. Zhang et al. 65 ) and individual decision making within crowds (e.g. Moussaı¨d et al. 66 ), considering the Lighting to influence movement in train stations 985 particular ways in which people might respond to alternative lighting interventions in differing circumstances.
Developing a rigorous evaluation methodology that can be applied within the context of a real railway station is vital, but challenging. Existing frameworks for the evaluation of transport/rail safety interventions are available and can be adapted to fit the existing context. 25,67 This will require the selection of an appropriate trial station, taking account of the platforms and stairways, the trains to study (type, configuration, timings, stopping positions), obtaining the required permissions and access, and identifying the hardware and software required to capture, store and analyse passenger movement data in an ethical and secure way, throughout the duration of the study.
While the consultation workshop involved a diversity of key stakeholders, their insights may be limited to the participants' own experiences and preferences and potentially constrained by industry related policy and agendas. Wider consultation with a range of industry stakeholders, both within and outside the rail industry, may provide additional views of the possibilities for new lighting interventions.
Finally, while the scenarios and proposals for lighting interventions identified through the study are potentially relevant in any train station context, there will be a need to ensure that they translate effectively to any given railway station setting.

Conclusions
The current study offers an approach to developing new concepts for lighting-based interventions. This approach is informed by a review of the available academic literature relating to the five characteristics of lighting identified and consultation with rail and transport stakeholders, along with lighting specialists, to consider how passengers' movements and behaviours can be influenced in train stations. Insights and guidance from the study offer an invaluable resource to designers, developers and operational staff within railways and also other transport/ service-related industries where managing people movement is a factor. The identified scenarios provide researchers with a range of common, tangible, real-world movementrelated problems and potential solutions that need to be evaluated. Specifically, it is proposed that well-designed evaluations are conducted on the following and will be the focus of a subsequent study: 1) The use of zones of light to indicate optimal waiting positions for effective boarding/alighting. 2) The use of zones of light in order to encourage dispersion along the length of the platform to alleviate crowding issues.
3) The use of zoning and bathing specific facilities in light in order to indicate optimum routes into/through/out of stations and trains. 4) The use of pulsing or chasing lights to determine bi-directionality in crowd movements plus overhead lighting to maintain task-related functioning in crowds. 5) The use of pulsing or chasing of lights to indicate appropriate walking pace in different conditions. 6) Changes in properties to affect subjective mood/feelings of wellbeing and safety, as well as providing emergency assistance for wayfinding.

Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: DW Windsor has a financial interest in lighting products which may be developed based on the work described in this paper.

Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Innovate UK (grant number 103965).

Supplemental material
Supplemental material for this article is available online.