CO2 emissions reduction is a top element of transport policy agenda. Among other mitigation policy measures, eco-driving techniques have proven to be effective in reducing fuel consumption and CO2 emissions. The aim of this paper is to compare the impacts of adopting eco-driving in different cities, road segments, traffic, and driver features. It intends to gain an insight into how city size and driving characteristics can reduce fuel consumption and CO2 emissions in order to develop specific eco-driving strategies. Field trials were conducted in two Spanish cities (Madrid and Caceres). 24 drivers, with different driving experiences, drove two different vehicles (petrol and diesel) along roads with different characteristics. The experiment was divided into two periods of 2 weeks; after the first one, drivers received an eco-driving training course. The impacts of eco-driving were measured comparing before and after results. They showed that eco-driving is highly effective in reducing fuel consumption and CO2 emissions in both, large-congested and small, cities. Savings between 5% and 12% were achieved. The efficiency increases with road capacity and decreased with city size. Eco-driving appears to be more effective in small, uncongested cities. In addition, limiting speeds on high capacity roads has proven to be a good energy saving measure.
One of the challenges in the developed world is to promote low-carbon mobility models in terms of social equity and fair distribution of wealth. This is, in short, the challenge sustainability poses. As a result, sustainable mobility means ensuring our transport systems respond to economic, social, and environmental needs, thereby minimizing any detrimental impact they have [
According to the European Environmental Agency, greenhouse gas (GHG) emissions from transport have increased every year since 2014, being 29% above 1990 levels in 2018 as a consequence of the growing demand in passenger and freight road transport [
Hence, to advance in our missions and accelerate the implementation of the Paris Agreement, the 25th session of the Conference of Parties of the United Nations Framework Convention on Climate Change, COP25 took place in Madrid in 2019. At this Conference, 73 nations (including Spain) have committed to become carbon neutral by 2050 [
The EU member states and the United Kingdom anticipate significant emission reductions in the transport sector from 2018 to 2030, to be achieved through a wide range of measures such as vehicle efficiency, low-carbon fuels, electric vehicles, and modal shift [
In this context, countries may act in several key areas in order to reduce GHG emissions from transport. According to a comprehensive transport-sector, GHG reduction strategy should at least address the following key policies [ Reducing demand for transport: controlling land use to avoid car dependency and increasing occupancy rates. Mode share: measures aimed at promoting more environmentally friendly modes such as public transport and nonmotorised modes. Fuel choice: measures aimed at using technologies for alternative fuels and new energy sources other than petrol. Fuel efficiency: promoting efficient technologies for vehicles and traffic management, traffic congestion abatement measures, and eco-driving.
These four policies to reduce GHG can only be achieved with a global societal compromise. It should be fostered and sustained through awareness actions, targeted to all sectors. The fourth policy depends particularly on individual travel decisions and on driving styles. Eco-driving aims to improve driving efficiency by limiting velocity, showing fuel consumption, and CO2 reductions around 10% depending on corridor parameters (including traffic volume and speed) [
Within this framework, eco-driving has emerged as the operational decision of drivers to maximize fuel efficiency and, consequently, reduce GHG emissions [
The most commonly applied method for individual drivers to implement eco-driving consists of training programmes that teach drivers theoretical and practical trainings. At a practical driving level, efficient fuel consumption implies smooth acceleration and braking, gear changes at low engine revolutions, and maintaining a constant speed avoiding sudden braking and acceleration. Fuel savings obtained after attendance at training programmes are heterogeneous, depending on the design of the training programme and the individuals’ performance [
Most of the research carried out so far have focused on measuring very specific impacts from particular types of car or in specific cities, e.g., Seedam et al. [
The aim of this research is to compare eco-driving global effects by adopting this technique in two different cities as different as Madrid and Caceres (both located in Spain). It intends to gain an insight into how city size and driving characteristics affect eco-driving. Therefore, its goal is to compare general changes in fuel consumption, CO2 emissions, and driving patterns according to different types of city in order to develop specific eco-driving strategies. To obtain unbiased global objective data, it will be necessary to carry out the test with different roads and vehicle types, driving under various weather conditions with drivers of different characteristics.
After this introduction, Section
This paper evaluates the short-term impacts of an eco-driving training programme. Field trials were conducted in both Madrid and Caceres, which are in the centre and west of Spain. 24 drivers, who were male and female, with different driving experiences, drove vehicles, one of which was powered by petrol and the other by diesel along different types of roads at various periods of time for one month. By comparing scenarios (before and after the eco-driving training), fuel savings and reductions in CO2 emissions were analysed according to road sections and the cities under consideration. Figure
Research framework.
As mentioned above, this research was based on a case study in which data was collected through a real data collection campaign. The methodological framework consisted in three steps extensively described in the following sections, which were data acquisition, dataset creation through data processing and filtering, and results analysis by evaluating impacts before and after training for eco-driving.
Firstly, two eco-driving tests were carried out, one in Madrid, the other in Caceres. Note, Madrid, with a population of 6.6 million, is the capital of Spain, while Caceres, with only 96,000 inhabitants, is relatively small. Thus, these cities had different road characteristics and traffic flows.
After the data collection campaign, by implementing an energy consumption model, instantaneous data recorded were processed in order to determine specific eco-driving strategies according to different types of city and roads (see Sections
The experiment consisted in a field trial in Madrid and Caceres which took place from April-May 2017. Each of the 24 drivers involved in the experiment drove along pre-established routes. The test was initially performed for 2 weeks and drivers drove as they normally did (first period). Then, after they attended theoretical and practical training for eco-driving (see Appendix
12 drivers were recruited for each city. Driving shifts lasted 4 hours; each one covered by a pair of drivers who alternated driving every 2 hours. 3 driving shifts per day of 4 hours each, from 8.00 am to 8.00 pm (only workdays), enabled us to collect a large amount of real data along different stretches of road under different traffic conditions. Data with regards to GPS position speed and engine parameters, as well as operational conditions, were collected by the second through a preinstalled device in each vehicle (see Section
The experiment is carried out with different vehicles, drivers, and road types to obtain global results. It took place in April-May, which in Spain is spring, with mild temperatures around 20°C. Driving was done mostly in sunny weather conditions. Exceptionally, moments of fog and rain were recorded but did not produce significant changes in the experiment results.
During the data collection campaign, 24 drivers drove diesel or gasoline vehicles along pre-established routes, before and after attending training for eco-driving. Ten different itineraries, which covered several alignments and types of road (e.g., local street, urban collector, major arterial, and motorway), were selected for the field test—six in Madrid and four in Caceres. Figure
Location of the field trials in Caceres and Madrid, Spain.
As the main goal of this research consisted in creating different eco-driving strategies according to city size, for purposes of consistency, the whole itineraries covered during the field trials have not been considered. In fact, in our analysis, we only looked at the same types of road in both Madrid and Caceres, which are local street and major arterial [
Information on driving urban routes tested in both cities.
City | Road section | Lanes | Speed limit (km/h) | Distance (km) | Driving time (min) ( | Average slope (%) I/II ( |
---|---|---|---|---|---|---|
Madrid | Local street | 1 × 1 lanes | 50 | 12.00 | 17–30 | 3.0/3.1 |
Major arterial | 2 × 2 lanes and barrier | 30–50 | 10.83 | 17–32 | 3.1/3.2 | |
Caceres | Local street | 1 × 1 lanes | 50 | 6.10 | 15 | 2.9/2.9 |
Major arterial | 2 × 2 lanes and barrier | 40–80 | 10.30 | 12 | 2.0/2.0 |
(
Madrid is the capital of Spain. Located in the heart of the Iberian Peninsula, the metropolitan area of Madrid covers 8,022 km2 of built-up land and has a population of 6.6 million. Three orbital motorways encircle the city (the M-30, M-40, and M-50) which are accessed from seven radial motorways [
The data collection campaign focused on the connection between Madrid and two interurban municipalities of the capital city, Pozuelo and Majadahonda. Thus, two itineraries, both located in the northwest of the city with different stretches of road and alignments, were selected to ensure a variety of driving and traffic characteristics in the sample. Both itineraries were characterized by having gentle slopes and connected the civil engineering school of “Universidad Politécnica de Madrid” (UPM) with two municipalities in the Madrid Metropolitan Area (Pozuelo and Majadahonda), where 92% of daily trips are made by car [
In Madrid, 12 drivers were involved in the data collection campaign. The sample was made up of seven males and five females aged between 23 and 50, with different driving experiences, between April 17th and May 19th, 2017. During the data collection campaign, 7,288 km of road were recorded in Madrid. 3,691 km during the first driving period and 3,597 km during the second one as eco-driving was implemented.
The routes covered were composed of different stretches of road, local roads or motorways, with different longitudinal and transverse profiles and characterized by having different speed limits. Thus, when comparing Caceres and Madrid, we only considered those stretches of the itineraries with the same types of road in both cities when extracting our results; i.e., local street and major arterial.
It is a small city, which can be crossed in under 15 minutes by any itinerary. Two alternative routes (local street and major arterial) with different road capacities were chosen for comparing results with Madrid. Both routes, which started at the university campus and ended at the train station, followed roads with different characteristics. The two routes with similar characteristics to those in Madrid are described below.
Local street: it is made up of urban roads that cross the city centre. The route is 6.1 km long and the speed limit is 50 km/h. Travel time is 15 minutes at an average speed of 23.6 km/h. There are several traffic lights, pedestrian crossings, and single-lane roads in both directions.
Major arterial: it is a recently built bypass which encircles the city. It has a section of urban motorway with level intersections with roundabouts. It is 10.3 km long and speed limits vary from 40 to 80 km/h. Travel time is 12 minutes at an average speed of 39.4 km/h.
In Caceres, the experiment took place from May 2nd to May 26th, 2017. During the data collection campaign, 2,472 km of road were recorded. The first 1,234 km driven was noneco (first period) and the next 1,238 km was eco-driving (second period). The sample was composed of 12 drivers, 8 males, and 4 females aged between 21 and 44 years [
Eco-driving was tested under real traffic conditions (free floating itineraries). During the data collection campaign, data were recorded by the second through an on-board logging, preinstalled device in each vehicle [
Data acquisition equipment.
Once all data were recorded, the VSP-Vehicle Specific Power model [
Itineraries covered during the data collection campaign were composed of several sections of road, i.e., urban motorway, collector, major arterial, and local street. The purpose of this research was to gain an insight into how city size and driving characteristics affected fuel consumption and CO2 emissions. In addition, in order to develop specific eco-driving strategies, we only took into account parts of the itineraries in which there were sections of road common to both cities. Thus, we only selected data on local street and major arterial roads.
The sample selected was composed of 1,022 trips, or 9,760 km of road (4,925 km in period 1 and 4,835 km in period 2), each one measured by the second with 128 different variables.
During the data collection campaign, data on GPS positions and instantaneous speed were recorded with an OBD-key preinstalled in each vehicle. The consumption values obtained with this device turned out to be different from the real consumption recorded on the petrol station receipts. Therefore, after an initial data processing stage with R programming, instantaneous fuel consumption was calculated on the basis of the VSP-Vehicle Specific Power model [
VSP depends on speed, acceleration, and grade. As shown in Figure
VSP for different slopes, speeds, and acceleration.
VSP (W/Kg) was calculated each second of driving and it was associated to one “VSP mode,” which corresponded to a certain interval of power requirements (Table
Correlation between VSP (power requirements in W/Kg) and “VSP mode.”
Vehicle specific power VSP (W/kg) | VSP mode | Driving and road conditions |
---|---|---|
<−2 | 1 | Negative slope, no acceleration |
[−2, 0) | 2 | |
[0, 1) | 3 | Low acceleration for increasing speed smoothly |
[1, 4) | 4 | |
[4, 7) | 5 | Rapid accelerations to reach higher speed |
[7, 10) | 6 | |
[10, 13) | 7 | |
[13, 16) | 8 | |
[16, 19) | 9 | |
[19, 23) | 10 | |
[23, 28) | 11 |
VSP mode sets driving ranges from low to high energy consumption. VSP mode 1 and 2 mean negative VSP values. These values are obtained when the car is running through negative slope and the accelerator is not pressed. Performance of these VSP categories are not stable delivering fuzzy results. VSP mode 3 means small positive values, which rarely happens since small speeds are usually linked to large accelerations/decelerations. VSP mode 4 is one of the modes that occur most frequently since it deals with common relationship between speed and acceleration. Finally, VSP mode 5 to 11 involve higher accelerations and speeds, so they typically occur less frequently, when starting a section with slope, or when overtaking another car or lorry.
Once the corresponding VSP mode was obtained by the second, each of them could be correlated with energy consumption according to vehicle engine and segmentation [
Correlation between VSP mode, energy consumption, fuel consumption, and CO2 emissions.
VSP mode | Energy [ | Instantaneous fuel consumption (10−4 l/s) | CO2 emissions (g/s) [ | ||
---|---|---|---|---|---|
Petrol | Diesel | Petrol | Diesel | ||
1 | 4.0 | 1.244 | 1.116 | 0.63 ( | 0.21 ( |
2 | 6.0 | 1.866 | 1.674 | 1.05 | 0.61 |
3 | 6.6 | 2.053 | 1.841 | 1.02 | 0.73 |
4 | 20.0 | 6.220 | 5.580 | 2.07 | 1.50 |
5 | 27.0 | 8.397 | 7.533 | 2.79 | 2.34 |
6 | 37.0 | 11.507 | 10.323 | 3.47 | 3.29 |
7 | 46.0 | 14.306 | 12.834 | 4.31 | 4.20 |
8 | 54.0 | 16.794 | 15.066 | 5.19 | 4.94 |
9 | 64.0 | 19.904 | 17.856 | 5.81 | 5.57 |
10 | 73.0 | 22.703 | 20.367 | 6.43 | 6.26 |
11 | 88.0 | 27.368 | 24.552 | 7.37 | 7.40 |
(
The consumption and emissions data are not very consistent with each other, since they come from different studies. In our research, we have given it as valid since our goal is not to accurately estimate the CO2 emissions produced by each car at any given time. We are looking for a comparison between efficiency of eco-driving in different cities in percentages and not in absolute terms.
By using this methodology and taking advantage of the high resolution of real driving data outputs (collected every second), we could evaluate how driving behaviour influences fuel consumption and emissions. This was fundamental to our research and enabled us to associate emissions with the different routes involved, as well as fuel consumption and eco-driving efficiency according to city size.
The evaluation of the impacts of eco-driving focused on the differences in fuel consumption and CO2 emissions between period 2 (after eco-driving training) and period 1 (before eco-driving training). It must be stressed that the aim of this research was not to obtain exact values for consumption or emissions, but to compare the efficiency of eco-driving according to types of routes or cities.
The effects of the training programme in terms of changes in driving patterns and reductions in fuel consumption are first checked according to vehicles, drivers, and types of road in both cities. Fuel consumption depends on driving style and traffic context on each road segment. An exploratory factor analysis was carried out to identify the main underlying factors related to driving elements, grouping the variables with high factor loadings [
Table
Description of the parameters selected for measuring effects of eco-driving.
Parameter type | Description | Code | Unit |
---|---|---|---|
Driving performance (OBD-key) | Average speed | avg_speed | km/h |
Average rpm | avg_rpm | rpm | |
Time with acceleration over 0.83 m/s2 | acc | s | |
Time with deceleration under −0.83 m/s2 | dec | s | |
Fuel consumption and emissions (OBD-Key + VSP) | Average fuel consumption | avg_fc | L/s |
Average CO2 emissions | avg_CO2 | g/s |
The analysis of the outputs shows how the different elements interact with each other: eco-driving skills, type of road, and city characteristics, to change energy consumption and consequently CO2 emissions. Firstly, energy efficiency (VSP modes and energy consumed) is analysed by the type of route and city characteristics. Secondly, the overall impact of eco-driving training regardless of the city is explored, along with changes in driving performance. Finally, it shows specific effects according to the two types of road selected in Madrid and Caceres in order to research the combined influence of city size and eco-driving in different city contexts.
Experimental statistics of distances driven by route, vehicle, and driving period are shown in Table
Km driven per vehicle, road type, and test period.
Road type | Non eco-driving (period 1) | Eco-driving (period 2) | ||||
---|---|---|---|---|---|---|
Diesel | Petrol | Total km | Diesel | Petrol | Total km | |
Local street | 1,592 | 1,802 | 3,394 | 1,698 | 1,616 | 3,314 |
Major arterial | 758 | 773 | 1,531 | 829 | 692 | 1,521 |
Total km per period | 4,925 | 4,835 | ||||
Total km driven | 9,760 |
The 9,760 km travelled have been distributed in a homogeneous way between eco and noneco-driving styles. However, this balance is not maintained in the type of road travelled since 69% has been done on local street and 31% on major arterial. This difference is due to the fact that Madrid has a bigger share of urban highways than Caceres, which is a small city, where most of the itineraries are local. For the shake of correct comparison, we have only selected the road sections which are common to both cities to compare the effectiveness of eco-driving between the two cities.
Figure
VSP mode distribution for different types of road and cities.
Caceres major arterial shows different behaviours to the other routes, since there is no marked peak at VSP mode 4 and there are bigger percentages of time at high VSP modes (bands 7–11). This difference is due to the speed limits on this route. Recorded maximum speeds are higher than speed limit (80 km/h). Enforcement is not so effective being a small city. Therefore, there are higher energy consumption on many stretches.
Figure
Average energy consumption (MJ/km) and average speed (km/h) for each city and type of road.
Figure
Caceres major arterial shows a high percentage of time (almost 30%) in VSP mode 1, so the final average values obtained for energy consumption are lower than those for the two local routes in Madrid and Caceres. Nevertheless, it showed more energy consumption than that recorded for the major arterial in Madrid where the speed limit was 50 km/h. Therefore, from an energy-savings perspective, restricting speeds on high capacity roads seems to be a sensible measure to take.
Figure
Effects of eco-driving training on traffic and emissions parameters.
Fuel consumption and CO2 emissions were always lower when eco-driving, and savings were bigger on higher capacity roads (major arterial). Savings values were between 5% and 12%, which were similar to values from other studies. Rolim et al. [
Eco-driving led to important reductions in all the driving parameters analysed in this research. The highest savings were once again linked to higher road capacity sections, being “major arterial” road types which achieve more reduction values and, on the contrary, “local street” type of sections account for the lowest.
Accelerations and decelerations were the parameters which change most when drivers do eco-driving (36–52%) while average speeds reductions were less marked (3–7%). These values were similar to other ones in which reducing CO2 meant decreasing travel speed and increasing travel time [
Figure
Impact of eco-driving in fuel consumption and CO2 emissions.
Madrid and Caceres showed reductions in consumption and CO2 emissions for all types of road, and the higher the road capacity was, the more savings there were. This suggests that eco-driving is less effective in congestion. Other studies such as those by García-Castro and Monzon [
The impacts of eco-driving on the other performance parameters are included in Figures
Impact of eco-driving on average speed.
Impact of eco-driving on average rpm.
Impact of eco-driving in accelerations and deceleration higher than 0.83 m/s2.
Figure
The range of legal speed limits on major arterial roads was higher than on local street. This produced a greater variation in noneco- and eco-speeds and hence greater dispersion in the error bars. In this respect, the values for Madrid were singular, since on major arterial roads average speeds did not fall with eco-driving. Average reduction values were obtained although these were very low compared to the other routes. On motorways, Barth and Boriboonsomsin [
From Figure
Finally, we analyse the impact of eco-driving on driving style (Figure
One of the priority goals of the political transport agenda, particularly in developed countries, is to mitigate GHG produced by the large number of CO2. Eco-driving techniques can contribute to reduce these effects, with a modest but consistent reduction.
This research studies the effects of eco-driving on fuel consumption, CO2 emissions, and driving parameters in two very different Spanish cities, Madrid and Caceres. The results showed that roads with less congestion (higher average speeds) obtained better result for average energy consumption. The effectiveness of eco-driving techniques is clear in both, in large-congested cities and in small ones. The saving values achieved in both cities ranged between 5% and 12%. Eco-driving was more efficient when driving on higher capacity roads and decreased with city size. Performing eco-drive in low capacity roads was less efficient because they usually are more congested.
These techniques were most effective in the small-uncongested city (Caceres). Their effectiveness in a large/congested city like Madrid was less clear. Eco-driving was found to be no efficient on local, one-lane streets. However, on these roads, when the lane was duplicated (major arterial), there was a big reduction on consumption and emissions. Regarding the effects of fuel saving on higher capacity roads, they were similar to those obtained in a small city such as Caceres.
Figure
Eco-driving impacts in different types of city.
Eco-driving is not a policy measure advisable for any road and urban context. In roads suffering chronic congestion problems, eco-driving could get worse the problem. However, in noncongested areas, it is clearly beneficial. This happens normally in cities like Caceres, where wider spread of eco-driving practice is recommended on any type of road, with substantial fuel savings and to a lesser extend CO2 emissions. In those cities, savings could be achieved in all type of roads and itineraries. On the contrary, in big cities, with more congestion levels, eco-driving policies should be customized; limiting speed on high capacity roads has proven to be a good measure to energy savings. In local roads, the problem is their capacity limit for not reaching the congestion threshold. Measures looking at solving bottlenecks and adding capacity such as traffic lights and parking regulation, dual carriageways, and bus priority lanes are recommended wherever possible, since they help to lower traffic congestion, facilitating also the efficient application of eco-driving techniques. What is universal—both for small and large cities—is that major arterial roads are the optimum road sections for recommending eco-driving. Therefore, public authorities should encourage training eco-driving at driving schools and awareness campaigns, as well as nonaggressive driving patterns [
Other papers have analysed eco-driving performance according to gender and driving experience [
The course has two parts.
Technical operation of car engines, different regimes, and their impacts on fuel consumption.
Concept of eco-driving s and how to apply this technique in practice.
Each driver performs two driving sessions with a supervisor. First, each participant drives with his own style and the supervisor indicates what is wrong. Then, he starts to apply the lessons learned in Part 1, while the supervisor corrects every time according to the traffic situations.
Finally, a final discussion on how to perform eco-driving in each situation is presented in a practical way.
The data used to support the findings of this study may be released upon application to the “Centro de Investigación del transporte,” who can be contacted at Centro de Investigación del Transporte Universidad Politécnica de Madrid Escuela de Ingenieros de Caminos, Canales y Puertos c/Profesor Aranguren, 3 28040 Madrid 91 336 66 56 (
There are no conflicts of interest.
This work was partly funded by the national R&D programme (Ministerio de Economía y Competitividad) under the Eco-Traffic Project “