Complexity and Usability of Voice-Enabled Alerting and Situational Reporting Decoupled Systems

Complexity and Usability of Voice-Enabled Alerting and Situational Reporting Decoupled Systems

Kasun Perera, Nuwan Waidyanatha, Tharaka Wilfred, Manoj Silva, Brenda Burrell, Tichafara Sigauke
DOI: 10.4018/jiscrm.2012100103
OnDemand:
(Individual Articles)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Telephone calls are the predominant telecommunication mode in Sri Lanka (Zainudeen et al). Leveraging low cost voice-based applications for disaster communication would be acceptable and sustainable. The findings in this paper are from an experiment concerning interactive voice for connecting community-based emergency field operatives with their central emergency coordination hub. Challenge was in interchanging the Freedom Fone (FF) Interactive Voice Response (IVR) generated, Sinhala and Tamil language speech data with the text-based ’Sahana’ disaster management system for analysis and decision support. Human interactions with the two decoupled software systems to accomplish the sequence of tasks, point to instabilities. Emergency Data Exchange Language (EDXL), a XML-based, interoperable content standard was adopted for mediation between the two disparate systems. Low quality voice data resulting in incomplete information was a barrier to automating transformations between the FF IVR and Sahana. Replacing these processes with human procedures significantly degrades the reliability. This paper discusses the complexities, usability and utility shortcomings discovered through controlled-exercises, in Sri Lanka.
Article Preview
Top

Introduction

Floods, cyclones, epidemics, and landslides are a common trend for the tropical island of Sri Lanka. Managing disasters is an immense challenge for the Government and other rescue and relief organizations in Sri Lanka.

Lanka Jathika Sarvodaya Shramadana Movement (Sarvodaya) is Sri Lanka’s largest community focused humanitarian organization operating in 15,000 of the 30,000 villages. Sarvodaya responds to all national and local -level disasters with rescue, food, shelter, and health related humanitarian assistance. Crisis information exchanges, for emergency coordination, are frequent between the Sarvodaya Community Emergency Response Team (CERT) members and their Hazard Information Hub (HIH). The HIH, namely the incident command center, is situated at Sarvodaya's head office, near the capitol city Colombo.

Wilfred and Waidyanatha (2011) realized through a survey that telephone calls were Sarvodaya's main emergency information exchange mode. This claim is independently justified by Zainudeen and Ratnadiwakara (2010), in their analysis of the Asians' preferred technology; thus, simple telephone calls, by a large margin, precede all other “beyond voice-based” applications. They also discussed the fact that increasing mobile phone proliferation has enabled over 90% of the households to have easy access to cheap telecommunication in Sri Lanka. The eight mobile and fixed telecommunications operators cover more than 95% of the Island. Sustainable general purpose technologies, especially mobile phones, can serve a dual role in supporting everyday activities in addition to emergency communications (Gow & Waidyanatha, 2011). Expanding Cutting-edge ICTs offer a variety of new options for enabling the humanitarian community to deliver services to the disaster victims efficiently and effectively, in hazard events.

Volunteers at Sarvodaya's HIH, referred to as HIH Operators (HIHO), would monitor hazard information through various websites and subscriptions (e.g., email bulletins or SMS alerts) for 'events of interest' that might be cause for concern (e.g., a cyclone originating in the Bay of Bengal). From here, relevant information bulletins would be relayed to selected individuals, typically Sarvodaya village leaders, in the communities. These individuals - referred to as CERT members – are responsible for activating local community emergency response plans (e.g. disseminating a local warning, producing incident reports, or engaging in relief activities).

Voice calls are, universally, easy for patrons. Simple telephone calls remove literacy and language barriers. It requires less training compared to communication through computers or high tech satellite devices. Since the mobile phone has become a part of life of people's lives, communication can be made any time of the day faster and more effectively. FF interactive voice-based application, being one such technology, presents itself as a tool that can be leveraged for emergency communication (Pant, 2011; Otieno, 2011).

Voice, as opposed to text, is an important element for enabling field level operatives to communicate disaster incident information. However, it is equally important that the emergency information is presented to central crisis managers in categorical data formats for rapid analyses and decision support. A challenge faced by the research team was developing the software interfaces and operating procedures for data interchange between the voice and text information management systems.

Perera et al (2011), Waidyanatha et al (2012-a) and Waidyanatha et al (2012-b) show that “large vocabulary continuous speech” voice data, received over the telecommunications networks, are significantly degraded by more than 13%. The noisy data cannot be subject to any kind of Automatic Speech Recognition (ASR). Laboratory research carried out by Weerasinghe et al (2007) and Bandara et al (2010) points to Sinhala Text-to-Speech software performing only with 71% accuracy. Given these shortcomings, it is impossible to apply any kind of automation software to transform the information between FF voice and Sahana text disparate systems. Therefore, human intermediation is necessary.

Complete Article List

Search this Journal:
Reset
Open Access Articles
Volume 11: 2 Issues (2019)
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing