Evaluating a Location-Based Application: A Hybrid Test and Simulation Environment

Abstract

L ocation-based applications are an important field of pervasive computing. In fact, Mark Weiser described location as one way to modify application behavior. 1 Examples of current location-based applications include visitor information systems, shopping assistants, interactive citywide games, and pollution-monitoring systems. One significant issue that affects developing and deploying such applications is the need for complex and expensive testing and evaluation. In particular, researchers must evaluate most location-based applications in their target environment with a suitable infrastructure and a volunteer user community. Such testing is costly and doesn't scale well to large numbers of users or applications. Consequently, we need a different approach to testing and evaluating location-based applications.In response to this need, we created a test environment that supports the evaluation of key aspects of location-based applications without the extensive resource investment necessary for a full application implementation and deployment. Although developers must evaluate many different aspects of their location-based applications (such as user interface issues, system and network related issues, and physical device considerations), here we report on system-and network-related issues (see the "Related Work" sidebar for other approaches). Using our test environment, we evaluated a location-based, mobile, remote heart-andlung health-monitoring application, focusing on component interaction, networking, location changes, and multiple component instances. So, our concern here isn't improving the application but the test environment and how we can use it as part of the test and evaluation process for this and other, similar applications.Because location awareness and, more generally, context awareness clearly distinguish ubiquitous computing applications from other mobile applications, evaluating how location-based applications react to context changes is critical in a test environment for such applications. Therefore, our test environment architecture distributes context events generated by simulators to applications with context-dependent behavior. Changes in simulated context can thus trigger changes in application behavior. This approach leverages the integration of third-party context simulators in the test environment. Health-monitoring applicationIn a typical location-based application scenario, users move about with a location-aware device that interacts with the supporting infrastructure. Users are provided with location-dependent information and interact with other users or infraComplex and expensive testing and evaluation can inhibit cost-effective development of ubiquitous computing applications. A new environment for testing and evaluating system-and network-related issues in location-based applications strives to evaluate these applications prior to deployment.

Requirements for an ubiquitous computing simulation and emulation environment

“…The Lancaster simulator [13] attempts to provide a "new environment for testing and evaluating system and networkrelated issues in location-based applications". Developed after Ubiwise, this work is more experimentally focussed.…”

Survey of Context Provisioning Middleware

“…A Hybrid Test and Simulation Environment [118] has been developed by researchers at Lancaster University. In contrast to UbiWise, their approach does not provide a three-dimensional GUI interface.…”

Towards Comparable Simulations of Cooperating Objects and Wireless Sensor Networks

“…As a humanin-the-loop simulator, virtual devices can execute real application code, interact with external services, and accept human input providing a powerful tool for simulating pervasive computing environments. The Lancaster [31] simulator supports the integration of third party simulators to evaluate location-based applications. Network packets from applications under test are intercepted in a modified kernel and redirected through the NS-2 thus providing the network simulation functionality, whilst a web services interface allows third party simulators, such as mobility models to be integrated unto a unified simulation.…”