Recommendations for actions concerning supporting ITS developments for VRUs

Intelligent Transport Systems (ITS) are defined by [1] as advanced applications which, without embodying intelligence as such, aim to provide innovative systems relating to different modes of transport and traffic management and enable various users to be better informed and make safer, more coordinated and ‘smarter’ use of transport networks.

In recent years, both technological developments and research activities in the fields of ITS have primarily focussed on motorised transport to improve safety, mobility and ecological (environmental) impacts. In regard to safety, the uptake of ITS has led to the decrease of road traffic fatalities, particularly amongst passenger car occupants. However, Vulnerable Road Users (VRUs), such as pedestrians, cyclists, motorcyclists and moped riders have not experienced the same decrease in fatalities: inside European urban areas, pedestrian fatalities represent 39% of all fatalities [2].

The VRUITS project, which was co-sponsored by the European Commission through the FP7-programme, has identified [3] and analysed [4] a set of ITS systems, which have the potential to improve the safety and/or the mobility and comfort of VRUs. The selection of the systems was performed in two workshops with stakeholders. In the first workshop, the applications which were seen as having the most potential to improve safety and/or mobility and comfort were selected [3]. The assessment was performed in two steps: a qualitative and a quantitative step. It was based on the method developed by Kulmala [5] for vehicle safety systems, which uses a set of 9 mechanisms, via which ITS can affect road user behaviour and thereby road safety, which was adapted for VRUs [6]. A qualitative assessment of the impact on safety, mobility & comfort was performed for a set of 23 systems [4]. In the second workshop with the stakeholders, the qualitative assessments were presented and a score was given by the participants to different aspects of the ITS systems (benefits, costs, deployment, user aspects). A multi-criteria analysis has been used to rank the different ITS systems. After that, a portfolio check was carried out to determine whether all important aspects were covered, for example, that systems addressed important vulnerable road user groups. This resulted in 10 ITS systems which were selected for a quantitative safety, mobility and comfort and socio-economic assessment. The 10 systems selected for quantitative analysis are presented in Table 1.

Building on the results of the safety, mobility & comfort assessment for 10 different ITS systems [4], and from the results of field trials on selected ITS systems [7], the VRUITS project has developed a set of recommendations for actions at the EU level regarding ITS systems for vulnerable road users and intelligent transportation systems,.

This paper presents the methodology to generate the general recommendations in Section 2. Section 3 presents the results for selecting and assessment of the general recommendations, while Section 4 presents the selected recommendations in more detail. The conclusions are outlined in Section 6.

Important sources of information for this paper are the VRUITS project deliverables. All these deliverables can be found at the VRUITS project website: www.​vruits.​eu.

The methodology for the identification and ranking of the main recommendations to enable the maximum deployment of selected ITS systems in Europe, consisted of four main steps. These steps are presented in Fig. 1.

Each of these steps is described in more detail in the following:

The impact on the penetration rate was calculated differently. The original CBAs were based on penetration rate scenarios – scenarios that contain figures on the usage of the system. Higher usage of the system was assumed to increase the safety benefits. In addition, effects were also anticipated in mobility and comfort due to changes in the penetration rate. The scenarios were used to calculate the impact of the recommendation by taking the difference of the high scenario CBA results and the medium scenario CBA results. This difference was then multiplied by 1/8, 1/4, 1/2 or 1, depending on the assessment and Table 2. The estimation of the recommendation benefits is the sum of all the additional benefits to the systems for which the recommendation is applicable.

The benefit analysis did not aim to provide an accurate estimate of the added benefits of the different recommendations. It was rather used to provide an indication of the order of magnitude of additional potential benefits of the ITS due to the recommendations. The following scale was used to rate the order of magnitude of additional benefits:

The potential benefits criterion has four categories valued 1 to 4 in the MCA, based on the “+” to “++++” scores. These values are multiplied by 2.25 in the overall score, to resemble that 50% of the scores come from potential benefits. In this way a maximum score on potential benefits (4 is the maximum value, multiplied by a weight of 2.25 equals 9) equals a maximum score on the remaining criteria (3 criteria with maximum value 3 also equals 9).

For example, imagine a recommendation with ease of implementation (medium), time horizon (long), costs (low) and potential benefits (++). The weighted sum is then 2*1 + 1*1 + 3*1 + 2*2.25 = 10.5. The weighted sum of these criteria was then taken to be the overall score, ranking the first 13 recommendations. The various criteria scores and overall scores used in MCA are shown in Table 4.

This section describes the recommendations with the highest scores (Table 11) in more detail.

Several of the recommendations are related to research to improve the performance of the systems, i.e. to determine reliably that there is a risk for an imminent collision and to decide to warn the driver or take a corrective action. In order to achieve a high performance, a high accuracy of the detection of the VRU and correct classification (R1) is needed, as well as an accurate prediction of the behaviour of the VRUs and their trajectories (R2), and as efficient warning strategies (R3) and efficient user interfaces (R4) to achieve the desired reaction of the road user.

Significant R&D has been performed on detecting pedestrians by vehicle sensors [9], much less research has been conducted regarding cyclists [10] and other VRUs. Work on accurate prediction of the behaviour of VRUs and their trajectories has been performed by Brunsmann et al. [11] and Köhler et al. [12]. The Euro NCAP rating system includes AEB systems for specific scenarios for pedestrians since 2016 and from 2018 onwards, will include both cyclists as well as night and obscure lighting conditions [13], which will stimulate development and deployment in this field. Additional funding is required in order to stimulate the development of systems addressing a wide range of road user types in challenging urban settings and environmental conditions.

User interface design, together with the comfort levels of the user, are crucial in influencing the use of the systems and the compliance with the advice provided by the system. The user interface should be designed to receive a correct and timely response of the road users addressing aspects such as comprehensible information presentation, minimum distraction and a reduced workload. Guidelines have been developed for user interfaces for drivers, e.g. [14, 15]. Aspects which require further research include integration of ITS and warnings to bicycles, helmets, vests, and the optimal way to provide warnings to the different road users.

The system warnings should initiate the desired reaction by the road user without causing annoyance, which can occur in case of too frequent low risk warnings. For vehicle drivers, research on warning strategies has been performed for ADAS systems [16, 17]. As the amount of assisting systems increases and hence potential warnings, guidelines are required for developers on the optimal timing and warning modes for drivers and other road users, to achieve an efficient reaction and increased system acceptance. Efficient warning strategies will have a positive effect on safety, and result in a small increase of penetration rate.

Recommendations R6 and R9 are both related to testing. Testing both in controlled situations and in a simulator testing is needed in order to increase the understanding of the interaction of different categories of VRUs - including elderly persons, people with limited mobility and vehicle drivers - with the systems (R6). Large scale field tests and system evaluations are needed to provide more evidence on benefits, limitations and costs of the systems (R9), as well as provide information on the long-term effects of the systems on the behaviour of both VRUs and vehicle drivers. This will produce a more concrete understanding of the feasibility of installing an ITS. The costs and complexity of field tests and system evaluations are often too large for investors to bear alone, and therefore require collaboration between stakeholders from different sectors and domains.

Two recommendations relate to signalised intersections: “Design of traffic light control systems which take the whole traffic demand of both vehicles and VRUs into account” (R7) and “Traffic signal systems for intersections should be able to adapt to different environments in order to accommodate for different traffic management schemes and different environmental conditions” (R10). Technologies such as big data and IoT allow for a real-time view of all road users approaching, waiting and crossing the intersection. This allows for the optimisation of traffic light phases for all road users, taking into account the VRU needs and policy goals for throughput, emissions and mobility, as well as avoiding unnecessary waiting times for road users.

Standardisation of the functionalities of the systems, including VRU related cooperative systems and communication between VRUs and vehicles, is required to enable interoperability (R8). For C-ITS systems, current message implementations, such as CAM and DENM do not take VRUs into account in a structured way. Recently, work on this subject has started in ETSI with the study on VRU use cases and standardisation perspectives (TR 103300). Further progress of this standardisation work for ITS systems for VRUs and underlying message sets and communication networks needs to take place in the upcoming years based on new insights, applications, and technologies.

Standardisation is a prerequisite to make the integration of different (cooperative) functionalities in a single device (R11) possible. The ITS applications for VRUs can either be deployed as apps in smartphones or on dedicated devices, which are integrated in the VRU vehicles, such as electric bikes, mopeds and motorcycles. The integration of communication equipment in VRU vehicles provides challenges, including limited cost, size and power consumption of the devices. Research and development has to address the physical integration of the devices in the vehicles, including the antenna design and radiation patterns for optimal performance of the devices. In case smartphones are used, the smartphone has to sense the context of the user, e.g. is the user walking outdoors or indoors, sitting in a bus, etc.

Privacy is critical in ITS systems, and is a major barrier to the deployment of systems. This is especially the case for C-ITS systems [18], where users broadcast data on their location. The C-ITS Platform addresses privacy and security, mainly for vehicle systems, but needs to be further developed to include VRU applications. Proper procedures for data exchange, data collection and storage, and the use of collected data (R13) have to be developed to guarantee the privacy of all road users.

Recommendation R12 relates to the blind-spot detection system. Legislation has to be implemented enforcing all trucks and buses to have blind spot detection systems with sensors at the back and at both sides of the vehicle.

For all infrastructure related systems, reliability considerations (including harsh weather conditions and light pollution) should have be accounted for in procurement procedures (R5). Governments should therefore include a comprehensive set of rules regarding a wide range of aspects ensuring that the deployed system works reliably in all environmental conditions and mitigates negative external effects (e.g. light pollution).

An overview of the pros and cons of the methodology for identifying and ranking recommendations for actions supporting ITS developments for VRUs can be provided by using the experiences in the VRUITS project. In addition, some guidelines will be provided for exploiting the methodology in future projects.

In order to achieve the full potential of the ITS systems at EU level, a wide range of barriers have to be overcome, including technical issues (detection inaccuracy, uncertainty in road user trajectories), user acceptance issues (high level of unnecessary warnings, lack of suitable user interface), business models (high costs, chicken-egg problems) and legal issues (privacy, current legislation). In order to address these barriers, a large set of recommendations has been identified, covering a wide range of potential measures, including research activities to technical aspects, field tests and simulation, behaviour and usage aspects. Cooperation of a wide range of stakeholders is critical for implementing the recommendations: European Commission, national authorities, various industries, user organisations and research organisations all play a role. Market related aspects and business models should also be investigated.

Based on the results of the analysis carried out, the focus should be on implementing the thirteen main recommendations. The recommendations regarding procurement (R5) and system design of traffic control systems (R10) can almost be directly taken forward. The technical recommendations regarding road user modelling (R1) and location accuracy (R2) are critical, as they allow the systems to more accurately predict potential collisions. Improved technical performance increases user acceptance by reducing “nuisance” alerts and enables increased efficiency in warning the driver or taking corrective actions. Major issues are the costs of the systems and privacy related issues. Field tests and large pilots which study how road users interact with the systems and demonstrate the benefits of the systems are needed. In order to implement the suggested recommendations successfully, all stakeholders have to be involved. The implementation and time planning of the other recommendations should be further discussed with the main stakeholders. A relevant aspect when discussing implementation is also the stability and safety of each ITS system, which can be addressed by looking at the possibility of malfunctions, consequences of ITS failure and potential maintenance costs. OEMs and infrastructure system providers should be stimulated to implement the sensors system and collision prevention tools. Authorities should be stimulated to deploy the systems, and road users encouraged to use the devices and systems which have been designed to improve their safety.

Most of the recommendations address several ITS systems, including all in-vehicle systems. Recommendations affecting multiple ITS systems can have large potential benefits. In the Benefit Analysis, calculations of five factors have been taken into account: safety, mobility, comfort effectivity, penetration rates and costs. Having done this on multiple systems for 13 recommendations, it can be concluded that especially increasing the penetration rate (i.e. number of users European-wide) can greatly contribute to the total benefits.

Most of the recommendations include challenges at a technical level or requiring the collaboration of a wide set of stakeholders, which cannot be resolved within a short time horizon (shorter than 5 years). A similar conclusion is made in relation to costs: only a few solutions have low costs, and the most effective recommendations are less likely to have low costs.