Paper 69 (Research track)

Ham Radio Telemetry Data for the Semantic Web- Relaying APRS Messages to Linked Open Data

Author(s): Alex Stolz, Martin Hepp

Full text: submitted version

Abstract: The Automated Packet Reporting System (APRS) is a popular communication protocol that essentially operates over ham radio, a radio frequency transmission service. Global cross-linkage of local radio networks is achieved by extending APRS to a hybrid infrastructure consisting of radio frequency channels and the Internet. APRS allows to broadcast information related to real-world, physical objects such as positional data (e.g. of balloons or cars), telemetry data, or weather reports. The ability to capture and maintain this data for further processing would be very useful in many application scenarios. In this paper, we describe a method to relay APRS messages to Linked Open Data (LOD). In particular, we (1) develop a domain ontology for APRS, (2) propose a real-time architecture for conversion and delivery around existing APRS servers, and (3) describe use cases for APRS data available as Linked Open Data.

Keywords: APRS; Ham Radio; Linked Open Data; Telemetry

Decision: reject

Review 1 (by anonymous reviewer)

(RELEVANCE TO ESWC) Generally, the introduction of novel Linked Open Data sources is a valid contribution to the ESWC. However, the article does not contain significant research components which somewhat reduces the relevance to the research track where this paper was submitted to.
(NOVELTY OF THE PROPOSED SOLUTION) As far as I can tell, there is no vocabulary or system for integrating APRS data into the Linked Open Data world. However, the system does not solve any major challenges posed by the data and the vocabulary does not seem to be overly complex.
(CORRECTNESS AND COMPLETENESS OF THE PROPOSED SOLUTION) The system and the vocabulary seem to be robustly defined. However the formulation and discussion of the vocabulary is rather limited. Instead the authors focus on the software system which includes not surprising elements. Here it might have been better to elaborate on the challenges posed by the APRS data (if any have been encountered). Similarly, the challenges with regard to telemetry have not been described well and the solution has been posponed to the user. Also no comparsion to existing sensor ontologies was given. Finally, it is not evaluated if the system will be able to handle and serve data in real-time (as claimed by the paper) for arbitrary time intervals.
(EVALUATION OF THE STATE-OF-THE-ART) The paper cites important work with regard to enabling sensor data for linked open data. However, it does not give a good comparison of the proposed ontology with existing work (e.g., the semantic sensor web). That is: can the APRS data be modeled using such vocabularies?
(DEMONSTRATION AND DISCUSSION OF THE PROPERTIES OF THE PROPOSED APPROACH) The use cases are limited to simple queries from the endpoints. No use cases showing specific application scenarios in the context of linking data or gaining semantics have been conducted. Similarly, no in-depth study of the collected data has been conducted, which would serve well in demonstrating the value of the collected data as the proposed approach.
(REPRODUCIBILITY AND GENERALITY OF THE EXPERIMENTAL STUDY) No experimental study was conducted. Here are some suggestions for possible solutions making the proposed paper more fitting to the research track:
* Studies on the characteristics and content of the data: how many sources, frequencies, temporal properties, spatial distributions, source categories, etc.
* In-depth comparsion to existing ontologies
* Application scenarions where the collected data allows for novel insights (e.g., in combination with Flickr or Twitter)
(OVERALL SCORE) === Summary of the Paper
The paper introduces a system for collecting and providing APRS data (like telemetry or weather information; transmitted over amateur radio) by converting it into an RDF format. In particular, the paper develops a custom vocabulary for representing ADRF data and gives a detailed overview of the developed back- and frontend. A frontend is online and some data is available.   
=== Strong Points (SPs)
* well written paper
* detailed explanation of the scenario and the methodology
* interface and some data are available
=== Weak Points (WPs)
* missing contribution to the research track
* collected and available data is limited
* code does not seem to be available 
* use cases are very simple: no research questions are addressed which are made possible by the system
* no statistics about the data are given
=== Questions to the Authors (QAs)
* will the code be made available?
=== Overall
APRS data sounds like an interesting resource to study. However, no insights on spefici characteristics or its value for further research are given. As such the paper is limited to the introduction of an RDF vocabulary and a software system. Neither semms to handle any specific challenges. Thus, overall the paper (if more data as well as the implementation is made available) may be well suited to the resource track. However, since there are no research related results (e.g., no evaluation of the proposed concepts), this paper has to be rejected.


Review 2 (by Victor de Boer)

(RELEVANCE TO ESWC) The paper discusses Linked-datafication of an existing and potentially valuable data model and as such it is relevant to the ESWC community.
(NOVELTY OF THE PROPOSED SOLUTION) The resources itself are new the system architecture and web service are novel, but to what extent this extends beyond the state-of-the-art is not well described
(CORRECTNESS AND COMPLETENESS OF THE PROPOSED SOLUTION) Vocabulary and conversion process seem able to capture the broad spectrum of APRS messages. However, this has not been shown in an evaluation
(EVALUATION OF THE STATE-OF-THE-ART) A short comparison to other IOT / Semantic Web solutions is presented. The system architecture and web service are novel, but to what extent this extends beyond the state-of-the-art is not well described
(DEMONSTRATION AND DISCUSSION OF THE PROPERTIES OF THE PROPOSED APPROACH) The system is well-described, using three use cases.
(REPRODUCIBILITY AND GENERALITY OF THE EXPERIMENTAL STUDY) No evaluation of any kind is given.
(OVERALL SCORE) In this paper, the authors describe a method to convert APRS (ham radio) messages to Linked Data. The paper describes a custom RDFS vocabulary for APRS as well as an architecture to capture, convert and expose APRS messages as Linked Open Data through a web service.
The paper is well-written and describes the technology stack in appropriate detail. This is an interesting domain which extends the appeal of the combination of LOD and IOT. However, in its current form, the research contribution is unclear. I would suggest either considering this for the Resources track of this (or another) conference or clearly identifying the research contributions in the paper itself. As such, I think this is a borderline paper. 
Strong points
- Interesting and valuable extension of LOD: vocabulary, dataset and web service. 
- Sensible solution to specific challenges surrounding the translation of APRS to an RDF format
- Well-written paper with clear a technological report as well as clear examples.
- Online service at aprs2lod.org
Weak points
- The research contribution remains unclear. What do we learn from this work? What universal semantic web challenges are tackled and what are broader implications?
- No evaluation of any kind
- A minor issue is that there is no public sparql endpoint (or Linked Data Fragments of sorts), a custom API is defined to identify and retrieve resources.


Review 3 (by anonymous reviewer)

(RELEVANCE TO ESWC) The paper is an intersection of two domains, i.e. radio communication and Semantic Web. Therefore, it might be interest of ESWC.
(NOVELTY OF THE PROPOSED SOLUTION) The paper proposes an architecture collect information related to weather and telemetry, from radio communication in LOD. This information can be used further by other services.
(CORRECTNESS AND COMPLETENESS OF THE PROPOSED SOLUTION) The paper is written well and easy to follow, however, some more easy examples would help in making it more clear.
(EVALUATION OF THE STATE-OF-THE-ART) Although, an end to end evaluation of overall contributions could be difficult, it would be good to provide a use case study to make it easy to conclude.
(DEMONSTRATION AND DISCUSSION OF THE PROPERTIES OF THE PROPOSED APPROACH) It discusses the different aspect of the information collected.
(REPRODUCIBILITY AND GENERALITY OF THE EXPERIMENTAL STUDY) The paper provide an REST api to access the information collected.
(OVERALL SCORE) The paper proposes a real-time architecture for conversion and delivery around existing Automated Packet Reporting System (APRS) servers, where APRS is a communication protocol which is used to operate the radio frequency transmission service. This paper presets an ontology for APRS, and converts the data obtained from Ham radio into standard Json and RDF, to facilitate standardization between different call signs.
-  Strong Points (SPs)
1. Addresses an important problem of standardization of APRS messages
2. Provide a REST api to reuse the converted data in a standard format, i.e. Json and RDF
3. Added semantics in the data using ontology 
-  Weak Points (WPs)
1. The paper is intersection of two different domains, which makes it hard to follow.
2. The contributions of this paper would be more align with Semantic Web in Use or Resource track.
3. The problem is too specific, which might make it hard to conclude the novelty in compare to any existing work on converting such data into LOD.


Review 4 (by Caterina Caracciolo)

(RELEVANCE TO ESWC) The work presented aims at providing a LOD version of the radio telemetry data communicated through the APRS protocol. The topic is relevant to the conference.
(NOVELTY OF THE PROPOSED SOLUTION) The approach proposed consists in converting the APRS messages into RDF triples, made available through REST web services. RDF is generated live and also made available as daily archives. Messages are given URIs and available also as RSS feeds.  Content negotiation is also available.
(CORRECTNESS AND COMPLETENESS OF THE PROPOSED SOLUTION) The authors provide and essential diagram to illustrate the APRS vocabulary, while the whole vocabulary is public and available online. Messages in native APRS are first converted into JSON, then RDF. While data is stored as N-triples, it is then made available in other standard RDF serialization for consumption by applications. This all seems reasonable. The specific problem to be addressed seem to be the split of each telemetry message into smaller messages that need to be then put together. The technical solution proposed is to post-process the RDF graph and recompose the unity of the message through SPARQL CONTRUCT queries.
(EVALUATION OF THE STATE-OF-THE-ART) According to the authors, this work is the first of its type. I have no elements to contradict.
(DEMONSTRATION AND DISCUSSION OF THE PROPERTIES OF THE PROPOSED APPROACH) Much of the effort in the paper is devoted to the description of the proposed approach, little to its discussion. I would appreciate a discussion on the stability of the RDF & URIs so generated, compared with the original APRS data.
(REPRODUCIBILITY AND GENERALITY OF THE EXPERIMENTAL STUDY) All mentioned materials, including source data, vocabularies and API are available online, which should make possible to test the approach proposed.
(OVERALL SCORE) The work presented aims at bringing radio telemetry data into the LOD "world". The proposed and implemented solution consists in devising a vocabulary to render elements of APRS as RDF triples, and web services to access the produced RDF data. The work presented relies on stable technology, which is a good point. Also, vocabulary and web services are available online for testing and experimenting as needed. Weak: as use cases, I would have expected to see examples of the benefit of turning telemetry data into RDF, for example by joining it with other data sets, through the LOD technology. Instead, the use cases presented seem to aim at showing that no information is loss during the conversion. 
How "open" is this data from the legal point of view? What happens when moving data produced within an amateurs' setting into the LOD cloud?


Review 5 (by Hala Skaf)

(RELEVANCE TO ESWC) Semantic/Linked APRS radio messages -- produced at large scale by distributed and independent IoT devices — may rise several conceptual and technical issues, such as scalability, (near) real-time constraints, or stream-based processing for instance. These issues are not considered in this work.
(NOVELTY OF THE PROPOSED SOLUTION) Although related works are mentioned the authors do not provide any comparison of their approach to related works (W3C Semantic Sensor Networks) nor try to experimentally evaluate their approach. The scientific content of this work is thus very limited.
(CORRECTNESS AND COMPLETENESS OF THE PROPOSED SOLUTION) Although related works are mentioned the authors do not provide any comparison of their approach to related works (W3C Semantic Sensor Networks) nor try to experimentally evaluate their approach. The scientific content of this work is thus very limited. 
- Section 3.4 should revise existing ontologies / vocabularies, and clearly state the objectives of the modeling effort.
- 5 pages of implementation details is too much and is not really informative 
- at the end of section 3.5, the authors propose two technical implementations to handle incoming messages “at scale” .  The description the two options is too vague.  Authors have arbitrarily chosen  option 2 without any justification.
(EVALUATION OF THE STATE-OF-THE-ART) Although related works are mentioned the authors do not provide any comparison of their approach to related works (W3C Semantic Sensor Networks) nor try to experimentally evaluate their approach. The scientific content of this work is thus very limited.
(DEMONSTRATION AND DISCUSSION OF THE PROPERTIES OF THE PROPOSED APPROACH) At least an experiment should be conducted to evaluate the proposed solution, in particular its scalability.
(REPRODUCIBILITY AND GENERALITY OF THE EXPERIMENTAL STUDY) No experimental study.
(OVERALL SCORE) This paper proposes an application of Linked Data principles to share Automated Packet Reporting System (APRS ) radio messages through i) a vocabulary (a UML class diagram, and some RDF instances),  ii) a web service architecture iii) three use cases. 
Semantic/Linked APRS radio messages -- produced at large scale by distributed and independent IoT devices — may rise several conceptual and technical issues, such as scalability, (near) real-time constraints, or stream-based processing for instance. These issues are not considered in this work. 
Although related works are mentioned the authors do not provide any comparison of their approach to related works (W3C Semantic Sensor Networks) nor try to experimentally evaluate their approach. The scientific content of this work is thus very limited. However, the proposed application is interesting and could be submitted as a poster or a system demonstration. 
Detailed comments : 
- Section 3.4 should revise existing ontologies / vocabularies, and clearly state the objectives of the modeling effort.
- 5 pages of implementation details is too much and is not really informative 
- at the end of section 3.5, the authors propose two technical implementations to handle incoming messages “at scale” .  The description the two options is too vague.  Authors have arbitrarily chosen  option 2 without any justification. 
- Almost same RDF predicates are used in the  4 Turtle serialization. Only one serialization should be enough. 
- Use cases could be transformed as motivating scenarios to support the identified issues and the proposed approach. 
- At least an experiment should be conducted to evaluate the proposed solution.  
3 strong points: 
• Interesting IoT context, and use of ham radio
• Easy to read and follow
• Detailed background on APRS
3 weak points: 
• The research question is not well identified
• No experimental evaluation
• No positioning with related works


Metareview by Hala Skaf

This submission presents an approach for providing a LOD version of radio telemetry data communicated through APRS protocol.  The topic is generally relevant to ESWC. Unfortunately, the contribution of this work is questionable because it lacks a proper research contribution.
Reviewers agree that the current version of the paper does not fit to the  research track, the research contribution of the paper is unclear. Positing with respect to the state of the art and experimental study are missing.  
As a result, the submission does not fulfill the requirements for a ESCW publication in terms of maturity, technical depth, and quality.


Share on

Leave a Reply

Your email address will not be published. Required fields are marked *