ecoDriver – this project targets a 20% reduction of CO2 emissions and fuel consumption by encouraging adoption of dedicated multimodal human machine interface (HMI), which supports the driver in conserving energy and reducing emissions.

While in-trip, drivers receive eco-driving recommendations like optimum speeds to save fuel and avoiding unnecessary braking at particular locations & instead to cruise at recommended speeds which are calculated based on driving style, terrain and vehicle characteristics for optimum utilization of engine energy.

To motivate & encourage the drivers, they are informed about their performance after every trip and receive next level of advises for improvements.

Aim

The aim of this project is to save abnormal engine energy consumption by improving the driving skill through effective vehicle-based real-time advisory system (during driving as well as post-driving) in order to optimize the energy utilization. This creates a closed loop between car and the driver.

Introduction

This project has been divided into 5 sub-projects:

SP1: Driver Support strategies for Eco Driving – this project is led by TNO, to formulate most optimum and effective eco-driving feedback mechanisms for a very wide range of scenarios and based on this, eco-driving applications are developed [2].

SP2: Real-time Calculation of Energy Use and Emissions - also led by TNO, quite a lot of reverse engineering has been done in this sub-project where special powertrain models have been derived in correspondence with CO2 emissions, and validation of these algorithms through real world trials, which is another sub-project.

SP3: Real World Trials - coordinated by CTAG, trials conducted across a range of driving scenarios (which were finalized in SP1), powertrains & vehicle models (finalized in SP2). Extensive data are gathered to validate the CO2 emission reduction.

The trials have been conducted in total at 9 sites across 7 countries, with a total of 57 vehicles (including trucks, buses and passenger cars)

SP4: Evaluation of Effectiveness - led by IFSTTAR, this is the phase where collected data from SP3 are subjected to comprehensive evaluation with specific aspects / criteria like fuel consumption, CO2 emissions, driver acceptance etc. & comparison of effectiveness of nomadic & integrated devices.

SP5 - Scaling Up and Future Casting: led by VTI (Sweden), estimates potential impacts of future technological advancements & lifestyle scenarios on green driving. Also, evaluating the benefits of ecoDriver systems keeping in mind the present and future market trends and fluctuations.

System Construction / Architecture

Coming to the control system / algorithm working behind all these services, overall basic architecture behind the working of ecoDriver can be seen in figure below:

ecoDriver system is divided into 3 different modules -

First module will be continuously studying the rapid changes in the driving situations and vehicle surroundings in real-time and keeps the track of all those that may affect vehicle safety and ecology that need sudden attention.
Second module takes all that is done by the first module, analyses the scenario outside the vehicle and decides / calculates what the driver should do to have safe and eco-friendly drive.
Third Module takes up the calculated advices from the second module and presents to the driver through different human-machine interfaces (HMI) like dashboard / head up display etc.

Out of the two feedback loops, the outer one plays the major role as this continuously runs to always give the best advice at all situations including the scenarios where the driving situations change rapidly. The purpose is first module is informed with what HMI is advising the drivers and compares with the vehicle surrounding at that instant. If it doesn’t match with the conditions, new set of constraints are generated and second module calculates accordingly.

As these are active systems, lots of sensors are involved for the success of the system. Sensor Structure has been constructed as required for system architecture (discussed in Fig. 3.) and is shown in figure below

Comparing these two figures

First module is handled by the algorithm called VE3 - Vehicle Energy and Environment Estimator which does the job of “What is going on” around the car.

Second Module is handled by the algorithm called RSG - Reference Signal Generator which does the job of “What should driver do”.

High range Radar, High Definition Monocular Cameras, CAN-bus and GPS are equipped into the car to sense the changes in the surroundings of the car.

Inputs

Following were the 6 types of systems which were developed for 7 different vehicles.

FeDS (Full ecoDriver System): with gear shift assist for manual transmissions and event-based advice / feedback
OEM: BMW ecoDriver system: This system was developed by BMW for its cars, which has Head up display to show avg. CO2 emissions; another unique feature is - it includes auto-coasting gearbox which supports the reduction in CO2 emissions. Fig. 5 shows the display system used by BMW.
OEM: CRF ecoDriver system: CRF (Fiat) had designed three HMIs for 3 of their cars - Alfa Romeo, Giulietta & Lancia - with the option of a haptic accelerator pedal (also called as smart haptic pedal - where the pedal gives drivers an intuitive signal right at their feet like either by knocking or applying soft counterpressure or vibrates allowing them to remove their legs off from the pedal just in case if the fuel is burnt more than required - which is a big leap towards saving fuel)
OEM: Daimler ecoDriver system: This was designed by Daimler for a truck and included a haptic pedal.
OEM: TomTom ecoDriver system: This was developed on a TomTom device platform enriched with ecoDriver features.
Nomadic ecoDriver system: designed based on FeDS design with basic functions to be used in any car as a plug-play system.

As there are many systems and tested on 60 vehicles, it was obvious that analyzing the loads of data was challenging. Three different data were logged: Driver behaviour, ecoDriver HMIs & Vehicle Performance. Therefore, a few pre-processing steps were constructed in order to handle unwanted data and prepare the right set of data.

For the on-road tests, two drive types were conducted: one is “Controlled” drive - in which vehicles were driven along a fixed route, and the other one is “Naturalistic” drive - in which vehicles were driven like normal daily use. During the controlled drive tests, observers were used in all the experimental cars (in both with & without systems) to collect driving behaviour data such as red light violations or overtakings with the help of dedicated application.

Also as mentioned earlier, 6 different vehicle makes were chosen for the tests which are listed below:

VW Passat CC (Petrol)
Nissan Leaf (electric)
BMW 535i (Petrol)
Lancia Musa (diesel)
Mercedes Benz Actros (diesel truck)
Renault Clio III (Petrol)
VW Touran (diesel)

Various sensors like RADAR, Camera, CAN-bus, OBD, GPS are also required for the vehicles with ecoDriver systems.

Outputs

Results have been categorized to 7 main aspects:

Energy & emissions
Driver workload & attention
Driver speed
Time headway
driver behaviour in events
Four Golden rules
acceleration & deceleration

MAIN FINDINGS: Energy & Emission

Positive:

Prioritizing safety over energy savings and thus making ecoDriver never to advice to drive over the speed limit and to encourage speed reduction at events like sharp curves
On an average, 4.2% reduction in fuel consumption (5.76% in urban, 2.2% in highways & 5.8% in rural) & 3.4% reduction in NOx emissions were recorded overall (5.1% in rural)
Embedded systems did a very good job compared to apps like from TomTom in saving fuel up to 6% and reducing NOx up to 5.7%.
FeDS had showcased significant results with an avg. savings of fuel & NOx reduction up to 1.5% & 3.2% respectively.
Haptic systems like haptic accelerator pedals reduced fuel consumption up to 3% compared to purely visual system.

Negative:

This project had aimed to achieve 20% reduction in energy usage which was not achieved.
System should advice to drive @ 100 kmph on the motorways with speed limit higher than 100 kmph, rather than following green speed technique, which may be less than 100 kmph. So, it gives us the hint that sometimes speeds calculated using green speed feature may not impress the drivers especially on motorways which may lead to dissatisfaction / negative opinion on using ecoDriver.

MAIN FINDINGS: DRIVER WORKLOAD & ATTENTION

Positive:

Regarding Driver Workload on using the systems, there was small increase in total workload because every now and then the driver had to interact with the system for the feedback. But according to the team, as the exposure towards the system usage increases, workload eventually decreases.

Negative:

By giving continuous feedback to the driver through visual user interface, driver gets distracted quite often.

Scope for improvement:

Glance pattern of the drivers in highway driving showed that drivers were anticipating feedback from FeDS, which indicates that human machine interface could be still improved to reduce this workload.

MAIN FINDINGS: Driver speed

Positive:

With eCoDriver system, while cruising average velocity will be lower.
It will restrict the speed automatically before it reaches the low speed limit region.
It reduces the speed well in advance of any intersections, sharp curves and zebra crossings in rural conditions.
It resulted in speed reduction on rural roads , urban roads and highways by 4%, 3.3% and 3.4% respectively.
Cruising speed reduction of 1.5% to 3.5% in all conditions and around 10% in urban areas was observed with the aid of embedded systems and additional 3.6% reduction with haptic systems.

Negative:

The cruising speed reduction with the App was not significant enough.

MAIN FINDINGS - Time Headway

POSITIVE:

The systems were more effective at the road intersections, sharp curves & during speed limit changes, showing 18% increase on average in time headway. Major contribution was from FeDS system.
Also, Time headway increased / decreased based on the road type which is a very good feature in the system, that it adapts to different road conditions automatically.

NEGATIVE:

The systems would not show any effect / changes on time headway if the cars are provided with haptic systems. This leads to huge dependency on haptic systems which are expensive.

MAIN FINDINGS - Driver Behavior in events

Positive: The system could not have control on red & amber light violations during the controlled trials. Sources say that it was proved difficult to observe in a reliable way.

Negative: Using Embedded type of ecoDriver systems, system could able to reduce the overtakings.

MAIN FINDINGS: Four Golden Rules

Before having the summary of Pros & Cons, it is good to have a look at Four Golden Rules -

Shift up as soon as possible: i.e., Shift up between 2,000 and 2,500 rpm
Maintain a steady speed: Use the highest gear possible and drive with low engine RPM
Anticipate the traffic flow as far as possible
Decelerate smoothly.

Based on the extent to which the golden rules were followed, some main findings have been listed out below -

POSITIVE:

The Engine brake usage improved (or reduced) in rural areas.
ecoDriver App shows nearly 100% effectiveness in compliance with all the rules except for the use of engine brake

NEGATIVE:

Engine Brake usage could not be regulated in urban areas and motorways.
Across all types of systems, 3% of avg. rpm reduction was registered when shifting up at right time in naturalistic drives
Based on the road types, system showed weaker feedbacks for highways.
Use of engine brake increased with both FeDS & the App by 5.1% and 6.4% respectively under controlled drive

MAIN FINDINGS: Acceleration & Deceleration

POSITIVE:

Embedded type of systems were able to soften the high accelerations and hard decelerations in both rural roads & motorways.
Acceleration of vehicle from rest will become less aggressive in all of the ecoDriver systems (i.e., engine load when starting from stationary will reduce thereby registering considerable reduction in energy consumption.)

NEGATIVE:

Haptic systems & App were not able to reduce the high accelerations & Decelerations.

Gap Analysis

Compared to the objectives and targets taken at the beginning of the project, there are some Gaps between them which have been identified as below:

On an average, 5% of energy reduction have been registered against the target of 20%.
According to their objective, they had to formulate feedback and advice strategies for the ecoDriver systems on 3 different powertrain - IC Engines, Hybrid & Purely Electric. But the results of the project focus more on strategies related to IC engines. Also they have not tested any of the ecoDriver system on Electric vehicles.
Comparing the number of features in common over the systems vary with Car Make (i.e., between BMW & Fiat, VW & Renault etc). This is due the sensors available on the vehicles. Not all sensors are available on all the vehicles. This is has led to difficulty in transferring skills between different ecoDriver systems.
This project has developed systems for the drivers who are willing to change their behaviour and learn eco-Driving. But for those who are not willing to change their behaviour / driving style, no approaches have been considered in the system, which may fail as the no. of drivers who are not willing to change are more compared to who are willing. To make the system more flexible and robust, this aspect of driver behaviour also has to be included in the system.

Documents:

Project Evaluation Form - ecoDriver