Toward a Semantic Web of Vehicles
1 like681 views
The document discusses the integration of semantic web technologies in smart vehicles, focusing on knowledge extraction for advanced driving assistance and diagnostics. It outlines a framework for real-time monitoring and data fusion using various sensors, highlighting the importance of semantic matchmaking for risk detection and decision-making. Additionally, it presents a prototype called iDrivesafe and emphasizes the use of ontologies for better understanding of driving contexts and safety measures.
![Toward a Semantic Web of Vehicles
Michele Ruta, Floriano Scioscia, Filippo Gramegna, Saverio Ieva,
Giuseppe Loseto, Agnese Pinto, Eugenio Di Sciascio
Polytechnic University of Bari, Via E. Orabona, 4, I-70125 Bari, Italy
[name.surname]@poliba.it
Knowledge Extraction for the Web of Things (KE4WoT) Challenge
co-located with The Web Conference 2018 - Lyon, France, April 23-27, 2018
Semantic Web of Things
Semantic Matchmaking for
Discovery and Detection
Smart Cars
Knowledge-based system for on-board
diagnostics and driving assistance
Knowledge-based real-time
car monitoring
iDriveSafe prototype for iOS
MAFALDA (MAtchmaking Features for
mAchine Learning Data Analysis)
Ontology-based model training in SWoT MAFALDA for smart cars
Semantic-enhanced VANETs
Emerging ICT vision: Internet of Things + Semantic
Web approaches
Embed intelligence into everyday objects and
locations
Take into account pervasive computing constraints
Optimize queries and reasoning
Ubiquitous Knowledge Bases (u-KBs)
Exploiting Concept Abduction and Concept Contraction
non-standard inferences [Ruta et al., Web Intelligence and Agent
Systems, 9(3), 2011]
Approximate matches and explanation of outcomes
Mini-ME reasoner
Mobile and embedded computing target architectures
SEMANTIC CHARACTERIZATION AND MATCHMAKING
DATA MANIPULATION
DATA GATHERING
Accelerometer
Magnetometer
GPS
ODB-II
Parameters
Google Maps
CloudMade
Google Weather
Integration of multiple information sources
Data fusion algorithms to identify high-level events
and conditions
Matchmaking to discover possible risk factors
Vehicle parameters (e.g., emissions, gear setting, fuel
consumption) annotated w.r.t. an ontology for driving safety
Vehicle Status Driver and ContextOBD-II Data
Urban Route
Even Pace
Urban Route
Imprudent
Extra-Urban Route
Imprudent
Semantic-enhanced data mining on sensor streams
ML classification treated as a resource discovery
based on non-standard reasoning
Outputs endowed with machine-understandable
descriptions
Data corpuses translated to OWL-based datasets
Reusing LOV vocabularies: M3-lite and Traffic Danger
Feature to concept mapping
Dynamic annotation of streaming samples
Output model includes most relevant axioms
(significance threshold)
Road surface, traffic congestion, driving style outputs
Camera viewfinder: classification outputs without
looking away from the road
Dataset publicly available on github.com/sisinflab-
swot/mafalda
Vehicular Ad-hoc Networks (VANETs) Concept Integration Problem (CIP)
for knowledge fusion
Semantic-enhanced VANETs
V2
V3
V1
Broadcast range Data propagation
V4
V1
V5V3
V4
Generated data Relayed data
(a) First BP (b) Second BP
1
2
3
4
5
Settings1
M’ (My) elements
(N and not others)
2
C’ (Confirmed)
elements (N and
others)
3
X’ (Clash)
observations (N
inconsistent with
others)
4
E’ (External)
elements (others
but not N)
5
managing incompleteness
reconciliating inconsistencies
Moving vehicles (OBUs) and fixed stations (RSUs)
Gathered environmental information to detect
dangers
Bidirectional dissemination scheme for data
generated and relayed by a OBU - Broadcast Period
(BP) time span
Efficient dissemination of annotations, compliance
with VANET standards
Annotations of a vehicle in a Semantic Packet (SP)
C X M E
P1 T T T
P2 T T T
P3 T T T
P1
’ T
P4 T T T
P5 T T
P4
’ T
Multiple sensors and Electronic Control Units (ECUs)
connected on Controller Area Network (CAN) bus.
On-Board Diagnostics (OBD-II) protocol and connector
real-time access to vehicle parameters
wired (RS-232, USB) and wireless
(Bluetooth, IEEE 802.11) scan tools
Augmented Machine Learning on low-level data for
Advanced Driver Assistance Systems (ADAS) and
autonomous driving
Knowledge Representation and Reasoning for high-
level information extraction and processing
quick adaptating
robustness w.r.t inaccuracies](https://image.slidesharecdn.com/rutaetalke4wotwww18posterv4-180502144732/85/Toward-a-Semantic-Web-of-Vehicles-1-320.jpg)
Toward a Semantic Web of Vehicles
- 1. Toward a Semantic Web of Vehicles Michele Ruta, Floriano Scioscia, Filippo Gramegna, Saverio Ieva, Giuseppe Loseto, Agnese Pinto, Eugenio Di Sciascio Polytechnic University of Bari, Via E. Orabona, 4, I-70125 Bari, Italy [name.surname]@poliba.it Knowledge Extraction for the Web of Things (KE4WoT) Challenge co-located with The Web Conference 2018 - Lyon, France, April 23-27, 2018 Semantic Web of Things Semantic Matchmaking for Discovery and Detection Smart Cars Knowledge-based system for on-board diagnostics and driving assistance Knowledge-based real-time car monitoring iDriveSafe prototype for iOS MAFALDA (MAtchmaking Features for mAchine Learning Data Analysis) Ontology-based model training in SWoT MAFALDA for smart cars Semantic-enhanced VANETs Emerging ICT vision: Internet of Things + Semantic Web approaches Embed intelligence into everyday objects and locations Take into account pervasive computing constraints Optimize queries and reasoning Ubiquitous Knowledge Bases (u-KBs) Exploiting Concept Abduction and Concept Contraction non-standard inferences [Ruta et al., Web Intelligence and Agent Systems, 9(3), 2011] Approximate matches and explanation of outcomes Mini-ME reasoner Mobile and embedded computing target architectures SEMANTIC CHARACTERIZATION AND MATCHMAKING DATA MANIPULATION DATA GATHERING Accelerometer Magnetometer GPS ODB-II Parameters Google Maps CloudMade Google Weather Integration of multiple information sources Data fusion algorithms to identify high-level events and conditions Matchmaking to discover possible risk factors Vehicle parameters (e.g., emissions, gear setting, fuel consumption) annotated w.r.t. an ontology for driving safety Vehicle Status Driver and ContextOBD-II Data Urban Route Even Pace Urban Route Imprudent Extra-Urban Route Imprudent Semantic-enhanced data mining on sensor streams ML classification treated as a resource discovery based on non-standard reasoning Outputs endowed with machine-understandable descriptions Data corpuses translated to OWL-based datasets Reusing LOV vocabularies: M3-lite and Traffic Danger Feature to concept mapping Dynamic annotation of streaming samples Output model includes most relevant axioms (significance threshold) Road surface, traffic congestion, driving style outputs Camera viewfinder: classification outputs without looking away from the road Dataset publicly available on github.com/sisinflab- swot/mafalda Vehicular Ad-hoc Networks (VANETs) Concept Integration Problem (CIP) for knowledge fusion Semantic-enhanced VANETs V2 V3 V1 Broadcast range Data propagation V4 V1 V5V3 V4 Generated data Relayed data (a) First BP (b) Second BP 1 2 3 4 5 Settings1 M’ (My) elements (N and not others) 2 C’ (Confirmed) elements (N and others) 3 X’ (Clash) observations (N inconsistent with others) 4 E’ (External) elements (others but not N) 5 managing incompleteness reconciliating inconsistencies Moving vehicles (OBUs) and fixed stations (RSUs) Gathered environmental information to detect dangers Bidirectional dissemination scheme for data generated and relayed by a OBU - Broadcast Period (BP) time span Efficient dissemination of annotations, compliance with VANET standards Annotations of a vehicle in a Semantic Packet (SP) C X M E P1 T T T P2 T T T P3 T T T P1 ’ T P4 T T T P5 T T P4 ’ T Multiple sensors and Electronic Control Units (ECUs) connected on Controller Area Network (CAN) bus. On-Board Diagnostics (OBD-II) protocol and connector real-time access to vehicle parameters wired (RS-232, USB) and wireless (Bluetooth, IEEE 802.11) scan tools Augmented Machine Learning on low-level data for Advanced Driver Assistance Systems (ADAS) and autonomous driving Knowledge Representation and Reasoning for high- level information extraction and processing quick adaptating robustness w.r.t inaccuracies