Keywords

1 Road Infrastructure Assessment

The road infrastructure assessment represents the initial building block of the proposed fuel efficiency assessment methodology. Figure 11.1 visualises the overall assessment approach and specifies required input as well as output data; e.g. the assessment of the road infrastructure (i) requires data on given platooning restrictions as well as the underlying road network and (ii) generates a risk-rated road network as output. The assessment aspects are detailed in the subsequent sections. As reported in Chap. 3, previous studies already investigated potential opportunities and risks of truck platooning as well as required technical and infrastructural conditions. The assessment of the road infrastructure with respect to specific truck platooning regulations represents a vital starting point for subsequently assessing fuel efficiency in certain regulation scenarios. A list of truck platooning regulation scenarios is presented in Fig. 3 in Chap. 3 based on related work and results gained by the Austrian council for road safety.

Fig. 11.1
figure 1

Fuel efficiency assessment methodology—overview

Full size image

In the Connecting Austria project, a software solution has been developed, which allows to generate a “risk-rated map” for a platooning regulation scenario comprising different constraints, e.g. restriction of platooning in tunnels or on bridges. The software solution allows to asses different scenarios and generates a road graph consisting of platooning-eligible and non-eligible road segments for a defined route. A visualisation of an example result is depicted in Fig. 11.2. In the following section, the software application that was used to generate the results presented in Chap. 12 will be discussed in further detail.

1.1 Risk-Rated Map

A given road network may consist of several possible zones of higher risk that are restricted for platooning. Given the assumption that a platoon needs to extend the distance between vehicles in a platoon and fully dissolve before it reaches an area that is restricted for platooning, one might ask the question “How many net kilometres will be left for platooning once all potential locations of higher risk and tolerance areas around them are excluded?”. In order to answer this question with respect to underlying assumptions, ANDATA Artificial Intelligence LabsFootnote 1 developed a solution strategy and implemented a software application. In the following, the assumptions, the solution strategy and a test example for the generation of a risk-rated map for the Austrian motorway are sketched.

Assumptions

  • Vehicles in a platoon are obliged to extend the intra-vehicle distance to the normal safety distance (\(\ge 50 \text {m}\) intra-vehicle distance) before reaching a zone of higher risk that is restricted for platooning.

  • Segments that are non-eligible for platooning can be categorised into:

    • Static higher risk segments, e.g. entries, exits, bridges, tunnels.

    • Quasi-static higher risk segments, e.g. long-term building sites.

    • Dynamic higher risk segments, e.g. road traffic accidents and patch of fog.

Solution Strategy

The road infrastructure assessment depends on the given road network and the defined platooning restrictions. Since the restrictions might be modified with regard to different platooning situations, a generic solution supporting a flexible configuration of restrictions based on an input file has been developed. The implemented software application may generate a (colour-coded) road segment classification depending on an input file. Doing so, routes may be analysed, e.g. with respect to the amount of platooning-eligible kilometres or restricted road segments in a given configuration.

The application generates a data processing pipeline based on available geographic information databases (e.g. GIP and OpenStreetMap), which calculates all possible routes that remain free for platooning, depending on the parameters defined in the input configuration file. The underlying architecture of this data processing pipeline is flexible in the sense that dynamically changing streams of data (e.g. traffic/weather information) could be fed into it. The outputs of the application are:

  • A map showing all road segments specified in the input configuration file, where areas that are eligible for platooning are labelled green and areas that are unsafe for platooning are labelled red (risk-rated map).

  • Statistical data, e.g. the total number of zones that are deemed safe for platooning and the distribution of lengths of distances between platooning and non-platooning zones.

Test Data Set D

In order to demonstrate the functionality of the application, a set of parameters listed in Table 11.1 was tested (matching the “play safe” scenario in Chap. 3) for the motorway A1.

Table 11.1 Input file parameters of test data set D
Full size table
Fig. 11.2
figure 2

Risk-rated map of a segment of the A1 for the direction west to east: test data set D

Full size image

A segment of the resulting risk-rated map of the motorway A1 in direction west to east is depicted in Fig. 11.2. Resulting characteristic numbers, e.g. the number of kilometres available for platooning given the input parameters, are listed in Table 11.2. A more detailed view of the distribution of lengths of segments that are eligible for platooning is displayed in the form of an empirical cumulative distribution function plot in Fig. 11.3 on the left-hand side. On the right-hand side, the distribution of the distances between these segments is depicted. According to Fig. 11.3, the maximum length of road segments that are available for platooning is below 8 km and about \(90 \%\) of all segments eligible for platooning are below 6 km of length for this input parameter configuration.

Fig. 11.3
figure 3

CDFs of the distribution of lengths of and distances between segments that are eligible for platooning on the motorway A1 in direction from west to east for the input parameters of scenario  D

Full size image
Table 11.2 Results for test set D on the motorway A1 for the direction west to east
Full size table

2 Driving Behaviour Assessment

The formation and dissolution of truck platoons may build upon different driving behaviours and strategies. Within the Connecting Austria project studies, different strategies for forming and dissolving truck platoons were defined and simulated on a microscopic level.

Regarding platoon formation, a collaborative strategy was defined, where all of the trucks adapt their speeds in order to close the gap between them and drive as a platoon. In general, the compression of the platoon occurs primarily by braking and accelerating the front platoon vehicles. Figure 11.4 shows some possible trajectories for solving the platoon formation problem.

Fig. 11.4
figure 4

Example of trajectories for platoon compression

Full size image

Regarding ad hoc platoon formation, that is, cases where there is an already formed platoon, and a truck that intends to join the platoon, the following possibilities were considered:

  • The truck is behind the platoon. The platoon does not act collaboratively; the truck has to accelerate in order to reach the platoon.

  • The truck drives in front of the platoon. The platoon does not act collaboratively; the truck has to reduce its speed in order to reach the platoon.

  • The platoon and the free truck collaborate, each one of them adapting the speed in order for the truck to join the platoon.

Regarding platoon dissolution, a collaborative strategy was defined, where all of the platoon trucks adapt their speeds. At the end of the dissolution manoeuvres, it is assumed that the trucks drive at their original speeds.

The fuel consumption for manoeuvres was calculated based on the vehicle speed, the intra-platoon distance and the formation/dissolution strategy. The additional fuel consumption was clustered according to intra-platoon distance, manoeuvre lengths for formation and dissolution manoeuvres, and the dedicated trucks in a platoon. Clusters for manoeuvre lengths were defined as follows:

  • 0–499m manoeuvre length \( \rightarrow \) fast formation/dissolution strategy.

  • 500–1199m manoeuvre length \( \rightarrow \) medium formation/dissolution strategy.

  • 1200–2500m manoeuvre length \( \rightarrow \) slow formation/dissolution strategy.

In order to derive fuel consumption models, a database of several 10000 truck kilometres driven on Austrian motorways and the dedicated fuel consumption has been evaluated. A visualisation of the distribution of truck routes in a typical data set within this database is shown in Fig. 11.5.

Fig. 11.5
figure 5

Visualisation of truck routes in a typical data set. The colours highlight several of the longer continuous paths contained within the data set

Full size image

In the course of obtaining suitable training data sets for a fuel consumption prediction model, data preprocessing routines were implemented. Trajectories which proved to be unsuitable for the purpose of predicting fuel consumption were excluded. For example, a filter that yielded only trajectories on motorways inside a speed range between 60 and 90 km/h was applied. Furthermore, data sets with faulty or implausible speed sensor readings, e.g. instant velocity drops or surges outside a certain boundary, were excluded. A machine learning model was then calibrated to predict the fuel consumption of a typical truck inside the speed range of interest. By means of an elaborated feature engineering process, the most suitable parameters and combinations thereof were derived for the prediction of the total fuel consumption on a given trajectory.

The results for additional fuel consumption for the formation and dissolution of a truck platoon depending on the intra-platoon distances 0.5, 1, 1.5 and 2s and the different driving behaviours (fast, medium, slow) are summarised in Tables 11.3 and 11.4 and visualised in Fig. 11.6. These results serve as basis to assess fuel efficiency and economic viability (see Sect. 11.3) of transport routes.

Table 11.3 Median fuel consumption and manoeuvre length for platoon formation and dissolution at a target speed of 80 km/h (1)
Full size table
Table 11.4 Median fuel consumption and manoeuvre length for platoon formation and dissolution at a target speed of 80 km/h (2)
Full size table
Fig. 11.6
figure 6

Median values for additional fuel consumption at intra-platoon distances 0.5, 1, 1.5, 2s

Full size image

3 Efficiency Assessment

In this section, the efficiency assessment approach developed and applied in the Connecting Austria project is described. Relevant input data for the assessment represents (compare Fig. 11.1):

  • Risk-rated road network—generated on defined scenarios for a certain route, comprising information on platooning ability of road segments.

  • Manoeuvre lengths for platoon formation and dissolving—depending on different strategies to form and dissolve a platoon.

  • Fuel consumption models—for solo trucks and trucks in a platoon, depending on intra-platoon distance and speed.

  • CO\(_2\) emission factor for fuel consumption—to estimate CO\(_2\) emissions.

Based on the risk-rated road network, net platooning distances may be calculated as illustrated in Fig. 11.7 and formula 11.1 for each road segment in the network. The assessment may consider two aspects:

  1. 1.

    The assessment of the general platooning feasibility for a given route.

  2. 2.

    The assessment of economically viable platooning operations for a given route.

The two aspects are detailed within Sect. 11.3.1 and 11.3.2.

$$\begin{aligned} d_{\text {net\_platooning}} = d_{\text {platooning\_allowed}} - d_{\text {platoon\_formation}} - d_{\text {platoon\_dissolution}} \end{aligned}$$
(11.1)
Fig. 11.7
figure 7

Net platooning distance illustration

Full size image

3.1 General Feasibility of Platooning on a Road Segment

The assessment of the general feasibility of platooning on a road segment is detailed in Algorithm 11.1. Basically, road segments where platooning is allowed and the length of the road segment is greater than the distance required to form and dissolve a platoon are accumulated to the net platoon length of a given route. This net platooning length may serve as basis to assess fuel and emission savings based on fuel efficiency effects of truck platooning. However, costs for forming and dissolving a platoon may decrease saving potentials. Therefore, a further investigation of economically viable net platooning distances is considered appropriate (cf. Sect. 11.3.2).

figure a

3.2 Economic Viability of Platooning on a Road Segment

The economic viability takes into account additional fuel consumption for platoon formation and dissolution. To assess these additional costs, reference values from microscopic traffic simulations based on diverse platoon formation and dissolution strategies (cf. Sect. 11.2) may be used. The approach for this assessment is detailed in Algorithms 11.2 and 11.3. In case the accumulated solo consumption of the number trucks in a platoon is greater than the consumption of the trucks driving in a platoon (including additional costs for formation and dissolution), a route segment is considered economically viable.

figure b
figure c

The resulting net platooning distance rated with a truck platooning fuel efficiency saving model that takes into account savings in relation to intra-platoon distances plus the additional fuel consumption for platoon formation and dissolution allow to asses fuel efficiency for certain use cases. Applications of the described procedure will be presented in the following Chap. 12.

4 Conclusion

This chapter described the methodological approach taken in the Connecting Austria project to assess fuel efficiency. Doing so, the approaches for the three aspects road infrastructure, driving behaviour and efficiency were detailed. Within the assessment of the road infrastructure, the suitability of road segments for truck platooning is rated based on configurable platooning regulations that allow a scenario-based assessment of fuel efficiency considering diverse regulations. Furthermore, the assessment of driving behaviour and strategies for truck platoon formation and dissolution provided a required basis to evaluate the economic viability of platooning-eligible road segments (see Sect. 11.3). The resulting segments suitable for platooning serve as basis to assess fuel savings for certain use application (see Chap. 12).