• Introduction
• Participants
• Objectives
• Background
• Research & Development
• Progress
• Benefits
• Sponsors
• Contact us

Background

Wireless sensor networks enhance Hong Kong development

One of the strategic directions of Hong Kong is to become a world-class services centre for the benefit of all stakeholders including local people, tourists as well as the supply chain industry. A key competitive advantage of the vendors in Hong Kong is their fast response to the business environment.

Wireless sensor networks (WSN) are an enabling technology in ubiquitous computing to support the next-generation digital society. They capture data from sensors and actuators. Applications atop WSN thus collect the pervasive data, iron out inconsistencies and resolve them so that application services can acquire accurate data instead of confusing or contradicting data.

However, it is difficult to develop embedded systems for WSN correctly, even though their typical program sizes are small. Since low-quality software requires frequent and high-cost maintenance, it would unavoidably obstruct the underlying digital infrastructure from effective and efficient responses to the changing business environment.

We show the key points in developing WSN programs as follows:

• The majority of embedded WSN programs are control software, thus, the testing of control predicates is crucial.
• It is difficult, time-consuming and error-prone to write test scripts in the embedded system languages, because such embedded programs are hard to develop in the first place.
• The number of test cases required to test WSN software is large.
• The length of the development cycle should be shortened, in order to improve the potential return on investment and avoid missing the time to market.

A testing toolset enhances WSN development

Software testing remains the most practical approach for quality assurance in the IT industry. To improve the quality of WSN programs, we plan to use our experience in several previous projects that lay the foundation of our current proposal:

1. In an RGC-funded project "TACCLE" by T.H. Tse and S.C. Cheung, we have developed a rigorous technique to test object-oriented programs and transferred the technique to develop an integrated testing toolset for the embedded software of their bonding machines.
2. In a RGC-funded project "FAST_SALE" by Y.T. Yu and T.H. Tse, we have analyzed the theoretical relationships amongst important classes of faults and developed a novel hierarchy of faults. We have also developed a testing strategy to test decision-making statements.
3. We also have collaborative research and management experience in other RGC-funded projects. We have eight joint publications on various issues related to software quality and testing.

For the testing of WSN programs, based on the fault hierarchy, we can reason analytically whether or not certain faults are subsumed by others. Advantages to develop a toolset are shown as follows:

• The techniques can now be opened up to the local industry. Our technology-transfer experience will be very helpful in developing the testing toolset for embedded programs for the local industry.

• The testing toolset will be integrated with the popular Eclipse development platform so that software engineers can use familiar interfaces to develop and test their embedded software.

• We also envisage that software managers can better plan their test budgets and shorten the testing time if analytical results and test statistics can be fed back to them via a high-usability toolset.