1 Introduction
ITS system |
Blind Spot Detection (BSD) |
Bicycle to Car Communication (B2V) |
Crossing Adaptive Lighting (CAL) |
Green Wave for Cyclists (GWC) |
Intersection Safety (INS) |
Intelligent Pedestrian Traffic Signal (IPT) |
Information on Vacancy on Bicycle racks (IVB) |
Pedestrian and Cyclist Detection System and Emergency Braking (PCDS + EBR) |
Co-operative (C-ITS) communications between Powered two wheelers (PTW) and Vehicle (PTW2V) |
VRU Beacon System (VBS) |
2 Methodology general recommendations
-
Ease of implementation: Easy, Medium, Challenging;
-
Time horizon: Short term (<5 years); Medium term (5–10 years) and Long term (>10 years);
-
Estimated effectiveness: Low (does not completely remove the barrier), Medium (barrier is removed in most cases), High (barrier is completely removed). This classification was only valid if the recommendation could be linked to a specific barrier;
-
Estimated costs: Low (hundred thousands of Euros); Medium (1–10 million Euros) and High (ten(s) of millions of Euros).
-
In addition, the relevant stakeholders were identified for each recommendation. A total of 213 barriers and 208 recommendations were identified. During the first analysis of recommendations, the recommendations were categorised in three different groups (recommendations regarding VRUs, vehicles and infrastructure). By means of grouping and additional analysis, the recommendations were merged and prioritised, resulting in a long list of 50 recommendations. A recommendation was included in the long list [8] if it scored high on the different criteria, such as having low implementation costs, being considered effective and concerning several motivated stakeholders.
-
Collection and processing the feedback of workshop participants. The recommendations were divided in three groups: recommendations regarding VRUs, vehicles and infrastructure. The workshop participants were divided into three groups and they participated in three different sessions, each covering one group of recommendations. The feedback collected during the workshop was used to update the contents and initial assessments of the presented recommendations.
-
Ranking of recommendations. At the end of each session, the workshop participants were asked to indicate the five main recommendations in each group which they considered the most important to facilitate the effective market introduction of the systems at a European level. The participants were asked to use a scale from 1 to 5 with five points for the most important recommendations of each group, four for the second most important, etc. Based on the overall scores given by the workshop participants, the recommendations with an overall score of less than 5 points were not considered for further analysis. This resulted in a shortlist of 38 measures.
-
Combination of similar measures. >In this phase, similar and/or overlapping recommendations were merged with each other. This resulted in a final list of 22 recommendations, which was verified by means of a questionnaire and sent to relevant stakeholders. The main purpose of the questionnaire was to outline the outcomes of the workshop and the selected recommendations. The respondents were asked to indicate if they agreed with the proposed recommendations, and if they saw any potential obstacles for the implementation of these recommendations. At the end, the respondents were requested to assess what they considered the most effective recommendation per topic.
-
Qualitative Analysis. An in-depth qualitative analysis of the 22 recommendations was performed on the criteria mentioned under ‘Step 1’. This resulted in a more detailed description of the recommendations, identification of main implementation issues, description of roles and required actions of main stakeholders involved, more information about barriers addressed and description of main cost categories.
-
Selection of 13 main recommendations. Based on the qualitative analysis together with the results of the workshop and the expert questionnaire, a list of 13 main recommendations was selected for a quantitative analysis of the potential benefits.
-
Benefit Analysis of 13 main recommendations. The method used for the Benefit Analysis was primarily based on the CBA framework used to assess the impacts of the 10 ITS systems in earlier phases of the VRUITS project [4]. The Benefit Analysis focussed on analysing the effects of carrying out the recommendation for each of these 10 ITS systems, compared to the situation where the recommendation would not be carried out. Each recommendation was considered to be effective for at least one of the 10 ITS systems. The effectiveness of each recommendation depended on the extent that the recommendation succeeds in removing the barrier. The recommendations are considered to directly enhance system performance in the areas of:
-
Safety, e.g. number of fatalities, injuries;
-
Comfort, e.g. comfort level;
-
Mobility, e.g. trip length, number of trips, trip duration or modal shift;
-
Penetration rates, e.g. percentage of potential users actually using the system; and
-
System costs i.e. the investment costs.
The impact of the recommendations on system effectiveness was qualitatively assessed, based on expert judgement of the partners involved in this study, for each of the above-mentioned areas. The assessment was made by rating the effectiveness as according to the scale: i) no effect, ii) very small, iii) small, iv) average or v) large. A distinction was made in relation to effects which are relatively short-term (2020), and medium/long term (2030).The main link between the qualitative assessment and the quantitative analysis is shown in Table 1. The table details how the different impacts on system performance can be converted to parameter figures. Large impacts on effectiveness were matched to the most positive situation in the sensitivity analysis and scenario analysis, following the sensitivity analysis and scenarios from earlier VRUITS work [4]. For smaller impacts, the quantification was based on an interpolation between large impacts and no impacts, as shown in Table 2.The assessment on safety, comfort and mobility was handled in the same way: an improvement in effectiveness resulted in higher benefits in the respective area, quantified by a percentage by which all effects related are multiplied. For example, if safety effectiveness is expected to have a small impact from the recommendation, the safety effectiveness parameter is 105%. This was then used in the CBA calculations, multiplying the corresponding safety benefits with 105%. Note that these percentages do not mean the system performs better than the theoretical maximum, but should be interpreted as an improvement to the theoretical maximum system performance. -
Effectiveness area | CBA parameter | Impact of effectiveness levels in removing barriers | ||||
---|---|---|---|---|---|---|
None | Very small | Small | Average | Large | ||
Safety | Safety Effectiveness | 100% | 102.5% | 105% | 110% | 120% |
Comfort | Comfort Effectiveness | 100% | 102.5% | 105% | 110% | 120% |
Mobility | Mobility Effectiveness | 100% | 102.5% | 105% | 110% | 120% |
Penetration rate | Penetration scenario | medium | 1/8 benefits of Large effectiveness | 1/4 benefits of Large effectiveness | 1/2 benefits of Large effectiveness | Difference high-medium scenario |
Costs | Cost reduction (2020, 2030) | (30%,70%) | (31.75%,72.75%) | (32.5%,75%) | (35%,80%) | (40%,90%) |
Quantified Effect: CBA inputs | ||
---|---|---|
Recommendation effects | 2020 | 2030 |
Safety | 100% | 102.5% |
Mobility | 100% | 102.5% |
Comfort (car drivers) | 100% | 102.5% |
Penetration rates | medium | ¼ of high penetration |
Costs | 30% | 75% |
Aspect/Indicator | Weight | Categories | Value |
---|---|---|---|
Ease of implementation | 1 | Easy | 3 |
Medium | 2 | ||
Challenging | 1 | ||
Time horizon | 1 | Short | 3 |
Medium | 2 | ||
Long | 1 | ||
Costs | 1 | Low | 3 |
Medium | 2 | ||
High | 1 | ||
Potential benefits | 2.25 | + | 1 |
++ | 2 | ||
+++ | 3 | ||
++++ | 4 |
ITS system | Number of barriers | Number of recommendations |
---|---|---|
Blind Spot Detection (BSD) | 22 | 25 |
Bicycle to Vehicle Communication (B2V) | 24 | 17 |
Crossing Adaptive Lighting (CAL) | 20 | 18 |
Green Wave for Cyclists (GWC) | 25 | 29 |
Intersection Safety (INS) | 18 | 11 |
Intelligent Pedestrian Traffic Signal (IPT) | 31 | 29 |
Information on Vacancy on Bicycle Racks (IVB) | 18 | 18 |
Pedestrian and Cyclist Detection System + Emergency Braking (PCDS + EBR) | 19 | 21 |
PTW Oncoming Vehicle Information System (PTW2V) | 13 | 16 |
VRU Beacon System (VBS) | 23 | 24 |
Total | 213 | 208 |
Potential barriers | Preliminary estimate on the ease to overcome the barrier | ||
---|---|---|---|
Easy | Medium | Challenging | |
Systems using camera may not work in poor lighting and adverse weather conditions | X | ||
Blind spot detection for Powered Two Wheelers has other requirements (other blind spots) than for non-motorised VRUs | X | ||
Currently there are different versions of the system on the market, some of them requiring equipment on bicycles | X | ||
Detection rate for pedestrians (and cyclists) is still quite low, at 77%. This needs to increase. | X |
Recommendation | Ease of Implement. | Time horizon | Estimated effectiveness | Estimated cost | Relevant stakeholders | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Easy | Medium | Challenging | Short | Medium | Long | Low | Medium | High | Low | Medium | High | ||
The automotive industry, with the aid of independent researchers where needed, should improve detection systems to detect both cyclists and pedestrians with low positive and negative false alarm rates. | X | X | X | X | OEMs | ||||||||
Inclusion of tests for BSD for non-motorised VRUs in Euro NCAP, including tests for night conditions. | X | X | X | X | Euro NCAP, Research institutes, OEMs | ||||||||
Promotion of BSD by European Commission (EC) for trucks and buses. | X | X | X | X | EC, Member States, OEMs | ||||||||
System should warn the driver if it is not working (failsafe) | X | X | X | X | OEMs |
No | Recommendations |
---|---|
1 | Research for the improvement of VRU detection accuracy and classification for in-vehicle and infrastructure systems |
2 | Research for improving location accuracy of devices for VRUs, e.g. where GPS functions poorly, such as smart phones or devices for VRU vehicles, e.g. through information fusion with other sensors such as cameras |
3 | Research on strategies for warning road users to make the system more efficient and less annoying |
4 | Development of reliable low-latency wireless communication components (ITS-G5 or LTE), which are suitable for integration in VRU vehicles (bicycles, motorcycles) and smart phones |
5 | Research on road users’ behaviour in traffic, especially of pedestrians (children) and prediction of intentions; including improved dynamics modelling of VRUs in traffic. |
6 | Research on optimal and efficient user interface |
7 | Reliability considerations (including harsh weather conditions and light pollution) should have a standard place in procurements of infrastructure related systems |
8 | Assessment of the possible effects of a potential mandatory use of the systems on long-term acceptance in different member states. |
9 | Testing of the systems in controlled situations and simulator testing to increase understanding of the interaction of different categories of VRUs with the systems |
10 | To design traffic light control, depending on the overall traffic demand (both vehicles and pedestrians) to minimise traffic delays for all road users while taking into account VRU needs and policy goals for safety, throughput, emissions and mobility |
11 | Standardisation of the functionalities of the systems, including VRU related cooperative systems and communication between VRUs and vehicles, that are required to enable interoperability |
12 | Creating field tests and system evaluations to provide more evidence on benefits, limitations and costs of ITS systems for VRU’s. |
13 | Starting implementation with local I2VRU systems providing benefits to early adopters. |
14 | Development of low-cost high-quality technical solutions for VRU sensing systems for both vehicles and infrastructure |
15 | 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 |
16 | Integration of different (cooperative) functionalities in a single device |
17 | Liability issues should be governed prior to deployment of the systems and further investigated |
18 | Guidelines/ regulations are needed on the use of mobile devices on bicycles to avoid distraction |
19 | Development of guidelines and standardisation regarding fail safe operation (informing road users of non- or mal-functioning of the system). |
20 | Regulation about signal timing changes needs to be updated, to accommodate sensor based systems. |
21 | Create legislation that all trucks and buses should have blind spot detection systems with sensors back and at both sides |
22 | Ensure that privacy of all road users is guaranteed through development and use of proper procedures for data exchange, data collection and storage, and use of collected data |
No | Recommendations | Systems affected |
---|---|---|
R1 | Research for the improvement of VRU detection accuracy and classification for in-vehicle and infrastructure systems | INS, BSD, PCDS + EBR, VBS |
R2 | Research on road users behaviour in traffic, especially of pedestrians (children) and prediction of intentions; including dynamics modelling of VRUs | INS, PCDS + EBR, VBS |
R3 | Research on strategies for warning road users to make the system more efficient and less annoying | BCD, PCDS, VBS, INS, B2V, PTW2V |
R4 | Research on optimal and efficient user interface design (e.g. information presentation, less distraction, reduced workload, integration of user interface in bicycle, helmet, vest etc.). | PTW2V, B2V, GWC, BSD, PCDS + EBR, INS, VBS |
R5 | Reliability considerations (including harsh weather conditions and light pollution) should have a standard place in procurements of infrastructure related systems | GWC, INS, CAL |
R6 | Testing of the systems in controlled situations and simulator testing to increase understanding of the interaction of different categories of VRUs with the systems | GWC, B2V, PTW2V, IPT, BSD, PCDS + EBR, VBS, INS |
R7 | To design traffic light control, depending on the overall traffic demand (both vehicles and pedestrians) to minimise traffic delays for all road users while taking into account VRU needs and policy goals for safety, throughput, emissions and mobility | IPT |
R8 | Standardisation of the functionalities of the systems, including VRU related cooperative systems and communication between VRUs and vehicles, that are required to enable interoperability | B2V, GWC, PTW2V, VBS, BSD, PCDS + EBR, CAL |
R9 | Creating field tests and system evaluations to provide more evidence on benefits, limitations and costs | BSD, PCDS + EBR, CAL, IPT, IVB |
R10 | 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 | IPT |
R11 | Integration of different (cooperative) functionalities in a single device | GWC, B2V, PTW2V, IVB |
R12 | Create legislation that all trucks and buses should have blind spot detection systems with sensors back and at both sides | BSD |
R13 | Ensure that privacy of all road users is guaranteed through development and use of proper procedures for data exchange, data collection and storage, and use of collected data | B2V, GWC, PTW2V, IPT, IVB, VBS |
Recommendation | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
System | R1 | R2 | R8 | R6 | R13 | R9 | R11 | R3 | R5 | R4 | R10 | R7 | R12 |
Total impact | ++++ | ++++ | +++ | +++ | +++ | +++ | +++ | +++ | ++ | ++ | ++ | ++ | + |
B2V | − | − | ++ | + | ++ | − | + | + | − | + | − | − | − |
BSD | + | − | + | ++ | − | + | − | + | − | + | − | − | + |
CAL | − | − | ++ | − | − | ++ | − | − | ++ | − | − | − | − |
GWC | − | − | + | 0 | 0 | − | 0 | − | 0 | 0 | − | − | − |
INS | ++++ | +++ | − | ++ | − | − | − | 0 | ++ | +/++ | − | − | − |
IPT | − | − | − | + | 0 | ++ | − | − | − | − | ++ | ++ | − |
IVB | − | − | − | − | 0 | 0 | 0 | − | − | − | − | − | − |
PCDS + EBR | ++++ | +++ | ++ | + | − | ++ | − | ++ | − | 0 | − | − | − |
PTW2V | − | − | ++ | +++ | ++ | − | +++ | + | − | ++ | − | − | − |
VBS | + | + | ++ | ++ | ++ | − | − | + | − | ++ | − | − | − |
Recommendation | Ease of implement. | Time Horizon | Estimated Costs | Potential Benefits | Overall score |
---|---|---|---|---|---|
R2 Research on road users behaviour | medium | medium | medium | ++++ | 15 |
R1 Research on improving VRU detection accuracy | medium | medium | high | ++++ | 14 |
R3 Research on warning strategies | medium | medium | medium | +++ | 12.75 |
R13 Guarantee privacy of all road users | medium | medium | medium | +++ | 12.75 |
R5 Reliability criteria in procurement | easy | short | medium | ++ | 12.5 |
R10 Traffic light control tailored to environment | medium | short | low | ++ | 12.5 |
R8 Standardisation of functionalities | medium | medium | high | +++ | 11.75 |
R11 Functionalities integration in one device | challenging | medium | medium | +++ | 11.75 |
R6 Testing of systems in controlled situations | challenging | medium | high | +++ | 10.75 |
R9 Field tests and system evaluations | challenging | medium | high | +++ | 10.75 |
R4 Design of optimal and efficient user interface | medium | medium | high | ++ | 9.5 |
R7 Traffic light control based on overall demand | medium | medium | high | ++ | 9.5 |
R12 Legislation of blind spot detection for trucks | medium | medium | low | + | 9.25 |
-
0: an estimated potential benefit of around 0–10 mln €
-
+: an estimated potential benefit of around 10–100 mln €
-
++: an estimated potential benefit of around 100 mln €–1 bn €
-
+++: an estimated potential benefit of around 1–2.5 bn €
-
++++: an estimated potential benefit more than 2.5 bn €An example of the calculations for the benefit analysis is provided in Box 1 below.
The recommendation ‘designing traffic light control depending on traffic demand, taking into account VRU needs’ (R7) is used as an example. The waiting times for all users are expected to decrease, through optimisation of the traffic light control. The recommendation will logically affect the Intelligent Pedestrian Traffic signal (IPT) system positively. This is confirmed by the participants of the workshop held by the VRUITS project, as it scored highest in the topic HMI & acceptance. In this case, IPT is the only ITS that is expected to be affected by the recommendation, however, most of the time, multiple ITS systems are affected by a recommendation. A qualitative assessment of the effects of the recommendation on the system has been made: | |
Safety: The number of red-light running accidents is expected to decrease as annoyance is decreased due to shorter waiting times. In addition, it indirectly decreases the number of annoyance related accidents. This all leads to a very small increase in safety.
Mobility: Is only marginally increased, due to optimised waiting times.
Comfort: Optimisation of algorithms increases comfort very little for all road users.
Penetration rate: Improvement of algorithms makes the system more mature and reliable, resulting in the ability to install these systems in a greater variety of (traffic) situations. Thus a small increase in penetration rate can be expected.
Cost: An accelerated deployment may reap economies of scale benefits, leading to a small increase in discount rate. Recommendation R7 has a long time horizon and for that reason the effects are estimated to only take place beyond 2020. Linking this to Table 2, the parameter values are thus easily determined (Table 3). The parameter values are then used in the CBAs. For Safety, Mobility and Comfort, this is done as described earlier, multiplying the respective benefits with the given percentage. The increase in penetration rates is calculated by using the figures for the medium-usage scenario in 2020, and the high-usage scenario in 2030. In the end, the difference with the original CBA output and the new CBA output is multiplied with ¼ to account for the corresponding “small” effect. The original CBA output is €160.9 mln, the new CBA output is €897.3 mln. Note that this new output seems fairly large in comparison with the original, but that’s because in this form it represents the benefits for a large increase in penetration rate. Multiplying the difference with ¼ gives the correct value (values in mln €): \( \frac{1}{4}\ast \left(897.3-160.9\right)=184.1 \). In this case, ¼ of the difference of the original CBA results and new one is between €100 mln and €1 bn, meaning the potential benefits are in the ‘++’ category. |
3 General recommendations
3.1 Analysis of the 10 its systems
3.2 Workshop and questionnaire
3.3 Further analysis of recommendations
-
Develop a more extensive description of the recommendation. What does it encompass? What needs to be taken into account when considering this recommendation?
-
Identify which systems are affected by this recommendation. Some recommendations can affect several systems, such as integration of several functions into a single device.
-
Identify which of the eight topics (technical performance, safety, mobility & comfort effects, human-machine interface (HMI), acceptance, business models, implementation issues, legal and standardisation issues and finally privacy, data storage, ethical and moral issues) this recommendation relates to.
-
Description of main implementation issues. Make the issues more specific for the involved stakeholders.
-
Time horizon. Provide main reasons for the estimated time horizon. Provide a more detailed overview of the actions needed for each stakeholder. Provide a high level action plan.
-
Identify the barriers addressed by the recommendation, and the effectiveness of the recommendation to address the barrier (low/medium/high). For each barrier, indicate the ease of overcoming the barrier.
-
Describe the main cost categories and cost-related barriers, the magnitude of the costs, and the relevant stakeholders
-
Development of detection sensors and algorithms, which detect all road users waiting and crossing the intersection, and potentially also of road users approaching the crossing.
-
Research on algorithm development, dependent on traffic demand by all streams, needs to take place. Representatives of the special user groups should be involved in the algorithm development. These developments need to take into account that some countries use detectors for dynamic traffic light signal cycles.
-
Traffic control suppliers need to be part of the development process and see a business case for implementation.
-
Road authorities and governments need to pay for the implementation of these algorithms for their traffic lights.
-
The EC: possibly involved in funding activities due to priority on Vulnerable Road Users and on Smart Cities (promote soft mode use, reducing emissions, promoting healthy lifestyles);
-
Road authorities and governments should be involved in testing activities and need to pay for the implementation of these algorithms for their traffic lights;
-
Research institutes: developing and critically evaluating the findings of the tests and ensure data collection during testing;
-
Technical partners, system suppliers, sensor developers: developing and providing equipment for the field tests; and
-
Representatives of user groups: ensure the needs of the specific groups are met by the innovation.
3.4 Final ranking of recommendations
4 Overview of the selected recommendations
5 Pros and cons of methodology
5.1 Pros of the methodology
-
The main advantage of the methodology is that it provides a filtering process that can be applied to shorten a long list of a very broad range of recommendations which have been defined based on qualitative information and to arrive at a short list of recommendations, by exploiting a combination of qualitative and quantitative information. The stepwise selection process of relevant recommendations ensured that most attention and in-depth analysis is focused to the most promising recommendations.
-
The methodology uses available results from earlier work done in the VRUITS project, such as the CBA calculations. Detailed interim results of the various steps provide relevant information which can be further used in the future work.
-
The adopted stepwise methodology included multiple validation checks with experts. Therefore, the methodology offered sufficient possibilities for feedback and review to improve the interim results. Furthermore, the workshop offered the possibility to discuss the interim results with the main stakeholders and to take their suggestions into account.
-
Quantification using classes of potential benefits provided an indication of the order of magnitude of the expected effects, despite the limited available quantitative information. The potential benefits were presented in this way in order to avoid a false expression of accuracy, which would be the case if we would present potential benefits in euros.
5.2 Cons of the methodology
-
It is a cumbersome methodology with a significant number of steps, which need to be carried out to define a final short list of recommendations.
-
A significant amount of time is needed to apply the methodology. Furthermore, the success of the execution of the methodology depends on the willingness of external experts to contribute to the process. The consulted experts must allocate sufficient time to the consultation and they need to be well prepared in order to fully understand the questions and provide relevant inputs.
-
There are some subjective elements in the methodology, such as quantifying the impacts using categories very small/small/average/large and when defining the weights in the MCA used for ranking the recommendations. The risk of subjective elements was minimised by involving stakeholders and experts in the process of validating the interim results.
-
Results of combining multiple recommendations during the benefit analysis of the main recommendations was not investigated in the VRUITS project. However the methodology is flexible enough to be adapted to take this into account, for example, by handling the combination of recommendations as a new recommendation and to carry out the same benefit analysis as has been used for a single recommendation.