The Drivaware system enables operators to perform virtually unlimited data analyses, elaborations and syntheses. As already mentioned in the global architecture overview, the system is able to analyze and synthesize data both on-line (using on-board probes) and off-line (automatically or manually).

The main goal of this approach is to allow to extract useful and valuable information from huge amounts of raw data, that wouldn't provide any interesting information per-se. Moreover, on-line data processing provides near-real-time information about vehicles' duty cycles, while off-line elaboration enables to deeply analyze the data and extract high-level synthesis results, useful for building valuable scientific knowledge.

The results of on-line analysis can be displayed as charts and maps, as shown in the data management section.

Edit

Automated Data Analysis

The server software can be configured so that it executes automated data analysis activities, such as computing the average on a channel's value, or the time a vehicle has been running.

Example of ROI Set-based Analysis

The analyses are strongly bound to geographic data. The system allows to create ROI Sets (Region of Interest), and apply analyses to them. The figure above, for example, shows an analysis that computed the cumulative amount of CO₂ that has been released during one day across a group of areas in the city of Lyon, France.

It's worth to note that automated analyses can be setup once and the system executes them automatically, updating the results with new data.

Edit

Data Synthesis

The key aspect of the platform is the ability to extract useful synthesis data from raw primary data.


From time-domain data, synthesis algorithms extract statistical information:

• Acquired data are read from the ECU
• Analysis characteristics and parameters are chosen before the synthesis starts
• Sampling frequency is imposed by the communication protocol
• All ECU-probe communication details must be known

Mathematical synthesis allows to obtain new information, such as:

• Average fuel consumption
• Polluting emissions (HC, CO, NOx)
• Mechanical components wear out models
• Route tipology identification
• ...

You can find more applications in the dedicated pages.