data management in ambient assisted living
platforms approaching iot: a case study
Susanna Spinsante, Ennio Gambi, Laura Montanini, Laura Raffaeli December 6, 2015
Dipartimento di Ingegneria dell’Informazione Universita’ Politecnica delle Marche Ancona, ITALY
table of contents
1. Introduction
2. SoA: Data Handling and Management in AAL 3. Data Management in AAL: a Case Study 4. Results and Discussion
5. Conclusion
data-related issues in aal platforms
• Heterogeneity (coarse classification):
• behavioral (or habits-related) [1, 2] • physiological [3, 4]
• environmental (or ambient-related) [5, 6] • healthcare [7, 8]
• Volume • Velocity
aal, big data, and iot
• Variety, volume, and velocity: commonly mentioned features in the field ofBig Data[9]
• Big Data commonly referred to in conjunction withInternet of Things(IoT) paradigm
• Remarkable affinity between AAL and IoT,but
• AAL exhibits specific requirements and constraints, about:
• data processing (local / remote)
• data delivery format (raw data / aggregated information) • data sharing / access options
aims of the work
• To investigate the emerging approaches to data handling and management in IoT-oriented AAL platforms, by resorting to the state of the art literature
• To discuss the findings with respect to a practical AAL implementation, assumed as a case study
• To highlight the peculiarities expressed by AAL in data handling, thus contributing to the discussion around IoT reference models and architectures
soa: data handling and
manage-ment in aal
data collection and communication
• Aims of IoT in AAL: to facilitate data collection (e.g. RFID, NFC), to make communication establishment easier and more effective [10]
• IoT frameworks for AAL shall support bidirectional flows of information (Closed Loop Principle)
• Communication protocols: trade-off amonglifetime,security level, andpre-processingof sensors-generated data
• Data analysis capabilities needed to enable anomaly detection at the sensor level [11]
• Anytime + anywhere access to sensor data: constant network connectivity required
• Need to reduce the amount of data to be transmitted:
event-based data transmission [12], processing tasks to aggregate data for a compact representation of relevant information [13], learning from user’s decisions [14]
data processing
• Cloud-based big data management frameworks:
• data mining + machine learning algorithms: from sensors data to knowledge
• real-time processing of large data volumes: clustering, classification, feature selection [15]
• heterogeneous data collected by AAL platforms cannot be efficiently supported by flat and data-agnostic management frameworks, but demand for differentiated processing, to ensure service-specific requirements
• Local data processing:
• collected data analysis and validation [16]
• interoperability of AAL system with other platforms
• algorithms and protocols for data privacy, integrity, access control • security mechanisms to protect locally stored data [17]
data management in aal: a case
study
system architecture - i
• Several sub-systems / functional domains managed by alocal server[18]:
• home automation
• sensors for behavior detection • health monitoring
• human-system interfaces
• Heterogeneous devices (environmental sensors, home appliances,virtual sensors,smart objects)
• Heterogeneous transmission technologies and gateways: CAN-bus, ETH, WiFi, Bluetooth, NFC, SubGHz technologies • Different interfaces tailored to users’ needs: Smart TV,
touchscreen devices, NFC-enabled devices (tap-based
interaction), depth sensors • Local server:
• relational DB management, aggregation tasks (REST services) • report tasks (JSON objects), connection to a remote cloud-based
platform
system architecture - ii
living lab deployment
sensor-generated data and aggregated information - i
• The local server hosts a unified DB including:
• tables populated by each sensor / device
• aggregated tables generated byaggregation tasks
• Sensor-populated tables:
• power consumptiontable⇔power metering sub-system • sensorstable⇔environmental and AAL sub-systems • user statetable⇔smart object (smart shoes) • reminderstable⇔health operator / caregiver
• body sensor network concentrator(BSNC) table⇔smart object (wearable device)
• Aggregated tables generated by REST services:
• report_consumption: hourly power consumption in a day • report_activities: % amount of daily activity by monitored subject • report_cooking: daily report on time spent cooking
• report_presence: daily report on time spent at home / out of home
• Aggregated tables transferred daily to the remote platform as
sensor-generated data and aggregated information - ii
Table 1:Records in the local DB tables hosting sensor-generated datapower consumption user state
Field Data Type Field Data Type
id int(11) id int(11)
id_meter int(5) shoe_type varchar(20) bnode varchar(20) state varchar(20)
name varchar(50) date Date
type varchar(50) time Time
level int(2) timestamp bigint(20)
date Date date Date
time Time
timestamp int(13) power int(11)
reminders BSNC
Field Data Type Field Data Type
drug varchar(30) id int(11)
quantity Text date / time Date / Time
time Time family varchar(32)
device varchar(32) code varchar(16) description varchar(64)
date Date
sensor-generated data and aggregated information - iii
Table 2:Records in the local DB tables hosting aggregated datareport_consumption report_activity Field Data Type Field Data Type
id int(11) id int(11)
name Text name Text
level int(1) state varchar(20)
date Date date Date
hourly_consumptions Text activities Text report_cooking report_presence Field Data Type Field Data Type
id int(11) id int(11)
name Text state varchar(20)
date Date date Date
events Text time Time
data management policies
• Bidirectional communications between local server and sub-systems within the AAL platform
• Data exchanges originated upon user’s requests:
• Software applications running on physical devices / interfaces issue requests to the local server by HTTP protocol
• Local server replies in the form of JSON objects
• Data exchanges originated automatically (e.g. based on time triggers):
• Applications poll REST service on the local server
• REST service checks current time against pre-defined time triggers • Match found: requested data transferred as JSON object to
applications
• Unidirectional data flow from local server to remote
cloud-based platform (JSON objects generated by REST services)
examples of local data exchanges
• 3 types of requests may be issued to the server:
• configure: get information on sub-system configuration
• state: get information on current state of specific sensors within sub-system
• command: issue commands to loads controlled by the sub-system (single or multiple loads)
• HTTP request format:http://[IP address: port number]/*
• wildcard*may take different values:/domotic/configure/bnode,
/domotic/state/meter/current, or/domotic/command?type=[...]
• Other parameters depend on specific command to issue
raw sensor data
• Experimental evaluation of data volumes locally generated by AAL platform and remotely transferred
• Figures referred to living-lab deployment
Table 3:Daily and average amount of DB rows generated by AAL platform. Dates chosen randomly over four months.
Sample # DB rows # DB rows Total daily
Date (sensors data) (power meters data) # DB rows
18/02/2015 5455 75 5530 26/03/2015 3061 555 3616 31/03/2015 3669 281 3950 08/04/2015 4122 44 4166 14/04/2015 6166 375 6541 21/04/2015 5546 6 5552 07/05/2015 3996 465 4461 12/05/2015 3743 39 3782 22/05/2015 2607 40 2647 03/06/2015 6911 487 7398 16/06/2015 3349 35 3384 22/06/2015 3618 26 3644 Average 4353 202 4555 20
aggregated data
• Each night server applications executeaggregation tasks
• Aggregated tables and so-calledreports(JSON objects) are generated
• Reports are transferred to remote cloud-based platform:
• Activity: daily time % of sleeping, sitting, cooking, walking, stasis (≈1.3 KB)
• Presence: % of total daily time spent indoor / outdoor (<1 KB) • Energy: min, max, avg hourly power consumption for each device
(72 values per meter) (≈42 KB)
• Graphical reports and statistics generated on remote cloud-based platform, accessible by web
remote platform view
discussion
• Different data management policies:
• Massive and unobtrusive data collection for behavior analysis, prediction, anomalies detection (e.g. worsening health conditions) • Heterogeneous data fusion and synchronization issues
• Differentiated flows of operations:
• Almost static data (environmental sensors)⇔local decisions (e.g. open / close windows, switch on / off lights...), daily basis
processing, remote collection for long-term observation • Time-sensitive data⇔prompt local action (e.g. falls, alarming
events)
• Interoperability issues:
• Heterogeneous nature of AAL data
• Required formatting operations performed before data transferring to common remote platform
• Ensure integrity and standard rules compliance of single sub-systems data (e.g. certified biomedical devices or telemedicine systems)
conclusion
• AAL platforms have to cope with quitedifferentdata
requirementsand policies about transmission, storage
capabilities, and data processing to create structured relations (long term behavioral analysis)
• Orchestration and data validation by a local serverbetter than
agnostic data management framework (vertical applications of IoT)
• Experimental case study presented
• Contribution to the definition of anAAL data management profilein IoT
references I
[1] A. Aztiria, G. Farhadi, and H. Aghajan, “User behavior shift detection in ambient assisted living environments,”JMIR Mhealth Uhealth, vol. 1, no. 1, Jun 2013.
[2] K. Park, Y. Lin, V. Metsis, Z. Le, and F. Makedon, “Abnormal human behavioral pattern detection in assisted living environments,” in
Proceedings of the 3rd International Conference on PErvasive Technologies Related to Assistive Environments, ser. PETRA ’10. New York, NY, USA: ACM, 2010, pp. 9:1–9:8.
[3] O. Gama and R. Simoes, “A platform to emulate ambient assisted living environments,” ine-Health Networking,
Applications Services (Healthcom), 2013 IEEE 15th International Conference on, Oct 2013, pp. 46–50.
references II
[4] F. Palumbo, J. Ullberg, A. Stimec, F. Furfari, L. Karlsson, and S. Coradeschi, “Sensor network infrastructure for a home care monitoring system,”Sensors, vol. 14, no. 3, p. 3833, 2014. [5] C. Nugent, L. Galway, L. Chen, M. Donnelly, S. McClean, S. Zhang,
B. Scotney, and G. Parr, “Managing sensor data in ambient assisted living,”Journal of Computer Science and Engineering, vol. 5, no. 3, pp. 237–245, 2011.
[6] F. Corno and F. Razzak, “Real-time monitoring of high-level states in smart environments,”Journal of Ambient Intelligence and Smart Environments, vol. 7, no. 2, pp. 133–153, 2015.
[7] M. M., S. Wagner, C. Pedersen, F. Beevi, and F. Hansen, “Ambient assisted living healthcare frameworks, platforms, standards, and quality attributes,”Sensors, vol. 14, no. 3, pp. 4312–4341, 2014.
references III
[8] D. Rodrigues, E. Horta, B. Silva, F. Guedes, and J. Rodrigues, “A mobile healthcare solution for ambient assisted living environments,” ine-Health Networking, Applications and Services (Healthcom), 2014 IEEE 16th International Conference on, Oct 2014, pp. 170–175.
[9] A. Sheth, “Transforming big data into smart data: Deriving value via harnessing volume, variety, and velocity using semantic techniques and technologies,” inData Engineering (ICDE), 2014 IEEE 30th International Conference on, March 2014.
[10] A. Dohr, R. Modre-Opsrian, M. Drobics, D. Hayn, and G. Schreier, “The internet of things for ambient assisted living,” in
Information Technology: New Generations (ITNG), 2010 Seventh International Conference on, April 2010, pp. 804–809.
references IV
[11] A. Jara, M. Zamora, and A. Skarmeta, “Knowledge acquisition and management architecture for mobile and personal health environments based on the internet of things,” inTrust, Security and Privacy in Computing and Communications (TrustCom), 2012 IEEE 11th International Conference on, June 2012, pp. 1811–1818. [12] M. Kiran, P. Rajalakshmi, K. Bharadwaj, and A. Acharyya,
“Adaptive rule engine based iot enabled remote health care data acquisition and smart transmission system,” inInternet of Things (WF-IoT), 2014 IEEE World Forum on, March 2014, pp. 253–258.
[13] C. Tsirmpas, A. Anastasiou, P. Bountris, and D. Koutsouris, “A new method for profile generation in an internet of things
environment: An application in ambient assisted living,”Internet of Things Journal, IEEE, vol. PP, no. 99, pp. 1–1, 2015.
references V
[14] J. Lloret, A. Canovas, S. Sendra, and L. Parra, “A smart communication architecture for ambient assisted living,”
Communications Magazine, IEEE, vol. 53, no. 1, pp. 26–33, January 2015.
[15] H. Alemdar, “Multi-resident human behaviour identification in ambient assisted living environments,” inProceedings of the 16th International Conference on Multimodal Interaction, ser. ICMI ’14. New York, NY, USA: ACM, 2014, pp. 369–373.
[16] W. Kun, S. Yun, S. Lei, H. Guangjie, and Z. Chunsheng, “Ldpa: a local data processing architecture in ambient assisted living communications,”Communications Magazine, IEEE, vol. 53, no. 1, pp. 56–63, January 2015.
references VI
[17] Q. Tan, N. El-Bendary, F. Pivot, and A. Lam, “Elderly healthcare data protection application for ambient assisted living,” in
Advances in Security of Information and Communication Networks, ser. Communications in Computer and Information Science, A. Awad, A. Hassanien, and K. Baba, Eds. Springer Berlin Heidelberg, 2013, vol. 381, pp. 196–203.
[18] S. Spinsante, E. Cippitelli, A. De Santis, E. Gambi, S. Gasparrini, L. Montanini, and L. Raffaeli, “Multimodal interaction in a elderly-friendly smart home: A case study,” inMobile Networks and Management, ser. Lecture Notes of the Institute for
Computer Sciences, Social Informatics and Telecommunications Engineering, R. Agüero, T. Zinner, R. Goleva, A. Timm-Giel, and P. Tran-Gia, Eds., vol. 141. Springer International Publishing, 2015, pp. 373–386.
