Deploying Spatial-Stream Query Answering in C-ITS Scenarios

Tracking #: 2194-3407

Authors: 
Patrik Schneider
Thomas Eiter
Ryutaro Ichise
Josiane Parreira

Responsible editor: 
Guest Editors EKAW 2018

Submission type: 
Full Paper
Abstract: 
Cooperative Intelligent Transport Systems (C-ITS) play an important role for providing the means to collect and exchange spatio- temporal data via V2X between vehicles and the infrastructure, which will be used for the deployment of (semi)-autonomous vehicles. The Local Dynamic Map (LDM) is a key concept for integrating static and streamed data in a spatial context. The LDM has been semantically enhanced to allow for an elaborate domain model that is captured by a mobility ontology, and for queries over data streams that cater for semantic concepts and spatial relationships. We show how this approach can be extended to address a wider range of use cases in the three C-ITS scenarios traffic statistics, events detection, and advanced driving assistance systems. We define for them requirements derived from necessary domain-specific features and report, based on them, on the extension of our query language with temporal relations, delaying, numeric predictions and trajectory predictions. An experimental evaluation of queries that reflect the requirements, using the real-world traffic simulation tool provides evidence for the feasibility/efficiency of our approach in the new scenarios.
Full PDF Version: 
Tags: 
Reviewed

Decision/Status: 
Major Revision

Solicited Reviews:
Click to Expand/Collapse
Review #1
Anonymous submitted on 14/Jun/2019
Suggestion:
Accept
Review Comment:

This is an interesting paper both from a theoretical and a practical point of view.

My recommendation is to accept it as it is.

The paper is on the topic of cooperative intelligent traffic systems and makes the following contributions:

1. It defines the domain of C-ITS and 3 relevant use cases

2. It contacts interviews with experts to evaluate the proposed technologies such as ontologies, query languages etc.

3. It presents a data model, query language and query evaluation strategy based on spatial stream OBDA using conjunctive queries in description logic DL-Lite_A

4. It implements a prototype of the spatial stream OBDA approach using the PipelineDB system

5. The prototype is evaluated with respect to the requirements of the use cases expressed as queries.

6. A detailed discussion of related work is given.

The paper is well written and the authors have managed to mix theory and practice in a nice way.

Review #2
By Maxime Lefrançois submitted on 18/Jul/2019
Suggestion:
Major Revision
Review Comment:

This paper is intended to be a revised and extended version of a paper presented at EKAW 2018. It describes a system to query streams of spatio-temporal ontology streams, designed in the context of a project on Cooperative Intelligent Transport Systems (C-ITS).

The paper should clearly indicate to what extent this article revises and extend the EKAW paper. I had to read EKAW 2018 and this article side by side to identify the differences. Some content is copy-pasted from another article of the authors, without reference. Such practices are wasting the time of reviewers, and would pollute the research space if published. The authors must cite these previous work, and clearly state what is the novel contribution of the paper and what is already published (and where).

For this reason, I recommend Major Revision. The rest of this review provides guidelines on how to also improve the next version of the article.

The introduction is almost the same as the EKAW 2018 paper.
Sections 2 and 3 correspond to sections 2 and 3 in the EKAW paper.
in Section 2.1, 2.2, 2.3, only the bullet points are slightly extended. Only the last paragraph is new (introducing a slightly different notation.)
Section 3 is almost the same, with some additional references to RCC5 RCC8 OWL RL,... (manual eye-diff)
It ends with stating that F6 F7 F8 are entirely new features, which is not true as they are published at EKAW.

in F4: L2 is not described.
Table 1 is not the same, but no justification is given for the differences. Why are there stared features? what does it mean?
There is a lack for proper justification for the required levels.
The column P9 contains references to very precise entailment regimes, but no justification is given why they are necessary and sufficient

Section 4 is new.
It contains a summary of the interview with four experts.
Although the point of view of each of the experts is interesting, in its current state this section it messy and raises important questions: how methodologically correct is it to interview the experts a posteriori? to what extent did their feedback get taken into account in subsequent development of the solution? For example, they mention the existence of messages of type CPMs, and that SOSA/SSN would be important to consider in the LDM ontology. Why weren't these suggestions integrated instead of barely mentioned?

The formalization p10 and the examples are very slightly extended.
- It is not clear what it means to apply function v to Fcam
- TBox and ABox do not contain the same type of axioms. Some belong to the ABox, some to the TBox.
- it is not clear how the definition of a spatial-stream knowledge base and the LDM are connected.
- why is it important that spatial objects in A have a spatial extension in Sa and ?
- it is not clear why limited form of disjunction would move the language beyond CQs "in general", but not here.

First 6 lines of section 5.3 replace first 3 lines of section Query Reqriting with Spatio-Temporal Relations in EKAW paper.
First and second paragraph of From timestamps to intervals seems to be new.
The second paragraph uses undefined setobject with new notations. This is not self contained.

Section 5.5: it is not clear to me how the system behaves when joins need to be done between the different streams?
The last paragraph of p12 that runs to p13 is new. It describes the general idea of the query evaluation strategy. To me this is really what would require expansion in this article. I expect algorithms; semantics; completeness/correctness proofs.

apart from this paragraph, I see almost no difference with the content in EKAW 2018. Again, it would have been easier if the authors provided a clear explanation of what content is new.

Section 6 Implementation:
only the last paragraph seems to be new. It describes how trajectory projections are made and lists potential future work.

Section 7 Evaluation:
Almost identical to EKAW 2018, only one new paragraph 7.5 summary of expert evaluatino.
This paragraph mainly condenses the suggestions made by the experts in Section 4. It would make sense to actually implement part of these suggestions as this is meant to be an extended version of the EKAW paper.

Section 8 describes related work (8.1) and compares the system to them (8.2).
- Section 8.1 is exactly the related work section of the ESWC paper published in 2017, title Spatial Ontology-Mediated Query Answering over Mobility Streams. I had to look up google to figure out.
- After all this, hard to tell if the Section 8.2 is genuine. If so, then it does represent 3.5 pages of new and interesting content to the paper.
I cleary recommend Major Revision, and I expect this section to also be extended in the next version of the article.

Minor comments

p2:1:10: rephrase
p3:1:39-41: Figure?
p3:2:10-11: rephrase
p3:2:39: hasLoc returns the geometry? or the location?
p4:2:34: on a sensor data -> rephrase
p5:1:17: on task of -> rephrase
p5:2:10-12: the features Fx haven't been introduced yet, and we don't know at this point that they are introduced below...
p5:2:24-25: why isn't OWL Time used?
p5:2:35: typo
p5:2:41-44: L2 is not described.
p5:2:49-51: not clear
p6:1:45-51: this is not a new contribution. It was a new contribution is EKAW 2018 paper
p7:Table 1:differences with EKAW 2018 are not explained. What are the started headers?
p7:1:44: typo
p8:1:17: typo
p8:2:6: typo (the map do not build something. they are built)
p8:2:19: are not crisp?
p9:1:51: typo
p10:1:16-18: probably a typo, one E should be replaced by Uc
p11:1:9: [agr,b] is the aggregate of last or next b -> specify (depends if positive/negative)
p11:1:11: what is a tuple of Fj? (notation of the EKAW 2018 changed in this article)
p13:1:16-17 first v is a line then it's an interval?
p14:1:2: typo
p15:Figure 3: Why are SpatialRelationEvaluator and SpatialObjectMatcher isolated?*
p17:1:43: typo
p24:1:14: typo

Review #3
Anonymous submitted on 22/Jul/2019
Suggestion:
Major Revision
Review Comment:

The paper tackles the problem of query processing against Cooperative Intelligent Transport Systems and proposes spatial-stream query processing techniques for efficiently solving this problem. The proposed approach resorts to the LDM ontology to capture the universe of discourse. LDM is represented in DL-LiteA and enables the representation of the domain in four levels. Conjunctive queries enhance with building spatial-stream predicates allow for the representation of numerical and spatial aggregations, and time intervals. Additionally, domain-specific predicates to represent numerical and trajectory predictions, as well as spatial matching and advanced reasoning are included in the language. The results of a user study where experts evaluate the features of the system from different perspectives, e.g., expressiveness of the LDM, data representation, system requirements. The query engine of A spatial-stream QA resorts to query rewriting techniques that enable the transformation of spatial and temporal relations while the axioms that encode the semantics of the building operators, e.g., during(.,.) is respected. The evaluation techniques named Hypetree Decomposition implement the execution of a query against the different relations of the knowledge base. The quality and performance of the approach are empirically evaluated over a data set of traffic data composed of streams and continuous views. The benchmark of data was generated using the microscopic traffic simulation tool, PTV VISSIM; this enables the simulation of realistic driving and traffic light behavior, as well as the possibility to create unexpected events like accidents. The empirical evaluation in conjunction with the evaluation conducted by the users provide evidence of the suitability of the approach.

This is an extension of the previous work of the authors published in EKAW 2018; this new version of the paper includes: a user study to evaluate the scenarios, use cases, and query features; additionally, the related work has been extended and query evaluation on top of the Hypertree decomposition is presented. With these extensions, the authors are showing the applicability of the proposed approach and the positioning of this work with respect to the state of the art. Although the results of the user evaluation are very valuable, it is neither clear the criteria followed to select the experts nor the questionaries followed to conduct the interviews/studies with the experts. Regarding the experiments, the studied use cases are explained in detail, and the reported results show the appropriateness of the proposed techniques. Nevertheless, the parameters that impact on the good performance of the approach are still not clear. Another important issue to highlight, it is the lack of analysis of the time complexity of the proposed approach, as well as potential query optimization strategies to ensure that the rewritings are efficient.

Positive Points
a) The problem of query processing over Cooperative Intelligent Transport Systems is formalized as ontology-based query answering over DL-LiteA; conjunctive queries (CQs) extended with spatial-stream capabilities are used for expressing queries. 
b) An extensive user study that includes experts in the domain and the utilized technologies
c) An extensive empirical evaluation of the behavior of the proposed approach.
d) A good analysis of the state of the art

Negative Points
a) The extensions of this paper with respect to the paper published in EKAW 2018 are not discussed.
b) Criteria followed to select and interview the experts are not described.
c) The parameters that impact on the behavior of the query processing techniques are not analyzed.
d) The time complexity of the query execution and query rewriting processes are not analyzed
e) Optimization opportunities at the level of query processing and rewriting are not mentioned.

Recommendation:
1) The EKAW 2018 paper needs to be cited in the introduction and the contributions of this new version of the paper need to be clearly listed in the introduction.
2) A subsection discussing the time complexity the query execution and query rewriting processes should be included in Section 5.
3) A subsection discussing the parameters that impact on the behavior of the proposed query processing and rewriting techniques should be included as part of section 7.
4) Subsection 5.4 should be extended with the optimization strategies to ensure that the query rewriting is efficient.