Welcome to a new post in this series covering the basis to a successful EDMS. It is time to talk about efficient and safe driving indicators. We already know the importance of the context and how to establish base lines to monitor the progression of the indicators. Let’s talk about indicators themselves.

ED indicators can be broad and varied. However, since the ultimate goal of improving driving efficiency will be to reduce fuel consumption, a key indicator will be exactly the amount of fuel used. This metric gives a clear reference of how efficient the driving has been on a particular route. However, when we think about evaluating how a person has executed a process such as driving, fuel consumption presents a series of important problems. The first of these is that the amount of fuel consumed is not a direct metric of the excellence in driving. Fuel consumption is affected by another series of issues such as, for example: 1) the nature of the service that may involve discontinuous routes with stops which will condition consumption; 2) the condition of the vehicle and its maintenance, such as: the condition of the tires, the orography, the weight of the load or the use of air conditioning, etc. These elements are not attributable to better or worse driving on the part of the vehicle driver. Furthermore, fuel consumption also largely depends on the context in which the driving takes place.

A driver applying her best efficiency techniques could have a higher consumption in one trip than another made previously if she is stuck in a traffic jam. The fact of not being able to advance and being immobilized while consuming fuel would make driving efficiency worse without being able to avoid it, even with the start-stop system activated.

Therefore, either we are able to isolate the context conditions perfectly or we will be misleading the evaluation by considering fuel consumption as the unique indicator. Given the inconsistencies that using fuel economy as a single KPI can lead to, it is important to develop specific metrics tailored to what drivers are expected to do. An obvious way to create these indicators would be to extract them from the efficient driving techniques that are indicated to drivers as a reference for the improvement process.

Let’s analyze some possible indicators. An example could be simply counting the number of times a driver uses the brakes. Braking causes the car to lose its kinetic energy and, from an efficiency point of view, it is negative. However, there are traffic situations where this might be necessary. A good way to improve efficiency would be to increase the distance to the vehicles in front, avoiding most of the braking. This is commonly known by the term anticipation. So if a driver anticipates traffic situations and road conditions he hardly needs to brake.

This occurs in all situations except for total vehicle stops in which at least a small final braking is necessary. In other cases, the driver can with a small touch of the brake correct small excesses of speed, which, although they could have been controlled in advance, would require maximum dexterity. As we can see, the selection of braking as an indicator of efficiency requires a parameterization in terms of the moments in which its actuation will be taken into account and a calibration of the threshold values ​​from which a negative action will be considered. Once this is done, it will be necessary to determine the indicator itself. That is, if you are going to use, for example, the number of times the brake is pressed under certain conditions when circulating at 100 km per hour or, for example, with some multiplicative factor depending on the intensity.

Following this philosophy, indicators can be defined. Here are some possible definitions:

Idle time

Minimum speed of revolutions per minute (revolutions or revolutions per minute) to which an internal combustion engine is adjusted to remain in stable operation without the need to actuate an acceleration mechanism or fuel inlet. This situation is detected by observing if the speed is zero and the rpms greater than zero.

Two levels are established from which the period is considered: level 1 (2 minute), level 2 (3 minutes).

KPIs:

max_idle: duration (seconds) of the highest idle that occurred during the route.

avg_idle: average duration (seconds) of idle periods that occurred during the route. It is an arithmetic mean.

level1_idle: sum of the durations (seconds) of the level 1 idle alarms.

level2_idle: sum of the durations (seconds) of the level 2 idle alarms.

Inertia

Time (seconds) elapsed with speed greater than zero and instantaneous consumption (fr) less than or equal to 1 l / hKPIs:

inertia: total duration (seconds) of all moments of inertia that occurred during the route.

These types of indicators that have just been listed are what we will call main indicators. They basically analyze an action at a given moment. However, more complex indicators can be defined. Since a driver’s performance is conditioned by what happens both before and after performing it, there is a way of evaluating that, instead of evaluating specific actions, it analyzes longer time sections. For example, while braking is bad in itself since it generates a loss of kinetic energy that previously had to be generated, it is much worse if braking is followed by reacceleration. On the one hand, the energy from the braking would have been lost and then energy would have been used again to recover the initial speed. We will refer this way of studying the driver’s actions as analysis of behavior patterns and they go beyond the analysis of specific actions. During this analysis period, the driver will chain a series of actions and their coherence can be studied from the point of view of efficiency.

An example of this type of pattern-based metric is as follows:

(Acceleration – Brake) Pattern

The behavior pattern of the driver that locates an acceleration followed by the activation of the brake seeks to detect inefficiencies mainly due to situations of lack of anticipation.

Initially, the driver accelerates the vehicle and then goes on to brake. Better anticipation would have reduced initial acceleration to avoid later braking. In this process part of the kinetic energy of the vehicle is wasted.

The pattern is detected by measuring positive acceleration followed by brake activation. To avoid false positives due to the slope of the terrain, it will be necessary to detect the variation in height through the Z component of the GPS. When the slope is negative the detections of the pattern will be discarded. In the figure we can see the scheme of the pattern.

KPIs:

In the case of the Acceleration – Brake pattern, the KPI is the number of times the pattern appears every 100km traveled by the driver.

𝐾𝑃𝐼𝐴(𝑒𝑣.𝐴𝐹100𝑘𝑚⁄)=𝑛º 𝑒𝑣ents AF total distance (𝑘𝑚)∙100(𝑘𝑚)1(100𝑘𝑚)

As we have just introduced,  ED indicators can go from simple to very complex ones to detect other kind of driving behaviours to act on. In any case the context where any indicator takes place must be always considered.