VEDA

Multiple Criteria Vehicle, Driver and its Environment Neuro Analytics System (VEDA)

 

Highlights

 

  • VEDA analyzes biometric and physiological signals of drivers and passengers
  • VEDA creates affective and physiological maps of the road and its infrastructure, climate, weather and vehicle conditions, indoor thermal comfort, drivers and passengers
  • Based on such maps, VEDA can determine the investment value of vehicles
  • Based on such maps, VEDA gives tips (safe driving, stress management and driving productivity) to stakeholder groups
  • A neuro decision matrix will be used in the integrated analysis of the road and its infrastructure, climate, weather and vehicle conditions, indoor thermal comfort, drivers and passengers emotional and the biometric states
  • VEDA considers a driver’s emotional state and suggest to change the driving process accordingly
  • VEDA compiles many of required alternative texts and videos and selects the more appropriate ones
  • VEDA integrates affective computing

 

The next subsystems are included in the Multiple Criteria Vehicle, Driver and its Environment Neuro Analytics Method and System (VEDA): Sensor Network, Intelligent Database and its Management Subsystem, Model-base and its Management Subsystem and User Interface.

 

 The Sensor Network, Intelligent Database and its Management Subsystem

 

The Sensor network is applied for sensing the emotional and physiological state of drivers. This Sensor Network consists of a full set of equipment the VEDA needs to analyse vehicle driver’ nonverbal information and assess the neurobiological response to the driving process. The Sensor Network assists in acquiring a great deal of multimodal data. These data are stored in the Database and used henceforth for the calculations in the Model-base. The study involved capturing a range of raw biometric and affective vehicle drivers data delivered by the Sensor Network in various formats. The Sensor Network comprises remote (face emotions, heart rate, drivers and passengers composition by gender and age group (FaceReader 5.0), temperature (infrared camera FLIR A35SC), voice emotions analysis (QA5 SDK), ACL-600 human body electrostatic detection instrument, etc.) and contact (eye pupil (Mirametrix S2 Eye-Tracker), brain signal (Enobio Helmet), heart rate and pressure (iHealth Wireless Blood Pressure Monitor), etc.) biometric analysis devices. The Sensor Network is spatially distributed above autonomous sensors to monitor physiological and emotional and physiological states of vehicle drivers and passengers, such as emotions, valence, arousal, voice, temperature, heart rate, systolic and diastolic pressure and others, and to cooperatively pass this numerical data through the network to the Multiple Criteria Vehicle, Driver and its Environment Neuro Analytics System.

The Multiple Criteria Vehicle, Driver and its Environment Neuro Analytics System comprises the following databases:

  • data about drivers and passengers (age, gender, years of riving, education, gender, etc.),
  • behavioural and qualitative research data of different drivers and passengers (socio-demographic, occupancy, cultural, ethical, psychological, religious, ethnic, emotional and behaviour),
  • meteorological and climate data (temperature, relative humidity, precipitation, wind speed and direction, etc.),
  • road and its infrastructure database,
  • vehicle data (age, etc.),
  • indoor thermal comfort level (air temperature, humidity, ),
  • database of best global practices on actions to motivate changes in drivers and passengers’ sustainable safe driving behaviour (different types of data analysis techniques, such as data and text mining, benchmarking against first principles (theoretical analysis), regression analysis and pinch analysis; analysis of trends (seasonal, changing conditions); benchmarking against other sites or processes) for engaging drivers and passengers towards safe driving;
  • database of best global practices (safety, energy, economic, legal, health, technical, technological, innovative, microclimate, social, cultural, ethical, psychological, religious, ethnic, emotional and other indicators) on actions to motivate and providing support to changes in drivers’ sustainable safe driving behaviour;
  • database and a knowledge base of best practice benchmarks;
  • Europe-wide data and knowledge on the effective safe driving of new and used cars;
  • multi-variant alternatives designing tables for tailored interventions (individualized tips);
  • notification database; notification system is a combination of software and hardware that provides a means of delivering a message to a set of recipients.
  • alert database;
  • calibration database;
  • the database of vehicle and their components;
  • the remote sensor network database;
  • the contact sensor network database;
  • the correlation database;
  • the emotional database;
  • the historical statistics database;
  • the intelligent database engine;
  • databases management subsystem.

A brief overview of the databases follows.

The data obtained by analysing the vehicle drivers and passengers’s emotions, valence, arousal and physiological parameters (average vehicle drivers and passengers facial temperature, heart rate, 17 voice parameters, vehicle drivers and passengers composition by gender and age group, etc.) captured by the Remote Sensor Network go into the Remote Sensor Network Database.

The data obtained by analysing a driver’s emotions, valence, arousal and physiological parameters (pupil size and blinking, brain signals, heart rate, systolic and diastolic pressure, palm temperature and humidity, etc.) captured by the Contact Sensor Network go into the Contact and Remote Sensor Network Databases.

The Correlation Database contains established correlations that link the vehicle drivers and passengers’s emotions, valence, arousal and physiological parameters with the road and its infrastructure, climate, weather and vehicle conditions, indoor thermal comfort, drivers and passengers and efficiency of efficient safe driving, stress management and driving performance (Priority Pj; Utility Degree Nj).

The Historical Statistics Database stores historical data.

The Intelligent Database Engine consists of two main parts: (1) Text Analytics and (2) determination of the correlation between the vehicle drivers and passengers’s emotions, valence, arousal and physiological parameters and the efficiency of efficient safe driving, stress management and driving performance of vehicle.

All data in the Multiple Criteria Vehicle, Driver and its Environment Neuro Analytics Method and System database are stored in tables and organized as a relational database. They are then used in a typical, relational Intelligent Database Management Subsystem. The most important functions of the Intelligent Database Management Subsystem are designing the database structure; loading, populating and editing a database; reviewing, searching, sorting and otherwise arranging the data; creating applications and compiling reports, and applying the Intelligent Database Engine.

 

 

The Model-base and its Management Subsystem

 

The following models comprise the Model-base:

  • the Text and Data (Big Picture) Analytics designed to monitor efficiency of efficient safe driving, stress management and driving performance and comfort at the vehicle;
  • the Recommender Driver Behavioural Change Subsystem;
  • the Model for Multi-variant Designing of Recommendations for Tailored Driver Behavioural Change Interventions;
  • the Model for Multiple Criteria Analysis and Prioritisation of Recommendations for Individualized Driver Behavioural Change Tips;
  • the Model for Determination of the Utility Degree of Driver Behavioural Change Recommendations;
  • the Subsystem for Personalised Guidance and Providing Support to Driver Behavioural (efficiency of efficient safe driving, stress management and driving performance and comfort) Change;
  • the Driver Behavioural Change Education Subsystem;
  • the Real-time Relationship Management Subsystem;
  • the Notification Subsystem that provides a means of delivering a message to a set of drivers and passengers;
  • the Alert Subsystem (alerts to drivers and passengers and managing the actions they undertake until the issue is resolved);
  • the Information Driver Behavioural Change Visualisation Model. The Model data include both quantitative and qualitative numerical and non-numerical data, such as data, text and knowledge;
  • the Analytics Model;
  • the Model for Affective Mapping of affective and physiological maps of the road and its infrastructure, climate, weather and vehicle conditions, indoor thermal comfort, drivers and passengers;
  • the Multiple Criteria NeuroAnalytics Model for Vehicle and their Components;
  • the Modelling Model;
  • the Recommender Model;
  • the Correlations Model;
  • the Text Analytics;
  • the Model Bases Management System.

The models are described below.

The Analytics Model is designed for decoding the initial data that registers the emotional and physiological state of vehicle drivers and passengers in order to receive logical evidence about what is happening in the driving process. We use LOGIT, KNN and MBP techniques for data mining, in parallel with explanations delivered by specialists. The results produced by the Analytics Model were submitted to specialists and drivers and passengers in graphical form to assist them with the interpretation of raw data.

Using the results produced by the Analytics Model, the Remote Sensor Network, the Contact Sensor Network, the User Group Database, the Correlation Database and the Historical Statistics Database as input, the Model for Affective Mapping of Vehicle maps emotions for affective and physiological maps of the road and its infrastructure, climate, weather and vehicle conditions, indoor thermal comfort, drivers and passengers.

The Multiple Criteria NeuroAnalytics Model makes an integrated multiple criteria neuroanalysis of the vehicle, driver and its environment and offers tips on ways to make the safe and rational driving, stress management more efficient.  A neuro decision matrix will be used in analysis and simulations. The data contained in this neuro decision matrix (criteria, their values and weights) provide a detailed description of a vehicle, driver and its environment in question and the emotions, valence, arousal and physiological parameters of people present at that vehicle.

With VEDA’s Remote Sensor Network (FaceReader 5.0, infrared camera FLIR A35SC, QA5 SDK, H.264 Indoor Mini Dome IP Camera, etc.), emotional (emotions, valence, arousal) and physiological (average vehicle drivers and passengers facial temperature, heart rate, vehicle drivers and passengers composition by gender and age group, etc.) parameters of people present at vehicle can be mapped by the Affective and Physiological Mapping Model for Vehicle, and driver groups can get personalised tips with an aim to make the vehicle more efficient.

The Modelling Model is designed for simulating the emotional and physiological state of vehicle drivers and passengers. Specialists also deliver important input for simulations of emotional and physiological states of vehicle drivers and passengers.

The Correlations Model determines the statistical relationships between the vehicle drivers and passengers emotions, valence, arousal and physiological parameters and the efficiency of efficient safe driving, stress management and driving performance of the vehicle aj (Priority Pj; Utility Degree Nj). The above correlations are valuable since they can show a prognostic relationship that can be used to make vehicle more efficient. It may be determined, for instance, which vehicles are preferred by women and men by age, or an analysis carried out to determine which vehicles are preferred by twenty-year or seventy-year women. This data then can be used for superior event tailoring for specific groups. Specific vehicles could, likewise, be better tailored to serve people from characteristic clusters. The Correlations Model searches the Database of Vehicle and their Components, the Remote Sensor Network Database, the Contact Sensor Network Database and the User Group Database for various dependencies and trends. Any identified trends are included in correlation tables that show links.

The Text Analytics module will establish the most effective, personalized text for a specific user by employing the compilation of possible most interesting text alternatives, the system of keywords and their weights. There is a selection of the desired number of pages of material with the help of Text Analytics according to the system of keywords and their weights. Additionally a selection can be made of material for the user according to the time the user is able to devote for reading the material of interest.