Easy development of monitoring applications for an autonomous float

Description

Autonomous floats have been used for decades to monitor physical properties of the oceans. More recently, these instruments have been used to record seismic signals in the oceans in order to improve tomographic images resolution at the global scale. A hydrophone is used to monitor the acoustic landscape but all the data cannot be sent through satellite communication since it is costly and power consuming. Therefore an algorithm is used to detect seismic signals. Monitoring the acoustic landscape of the oceans can also be used to study marine wildlife (whales, fishes, etc.) and the impact of human activities, or to study meteorology. For each of these monitoring activities specific algorithms must be developed and implemented on the instrument. But developing and implementing such algorithms on the float by traditional means requires much effort and expertise in embedded software.To overcome this, we have followed a Model Driven Engineering approach. This approach consists of a Domain Specific Language (DSL) - called MeLa - that targets the development of applications for the float. The language allows to write applications at a higher level of abstraction such that embedded software development skills is not required. The underlying model is used to compute several properties (e.g., battery lifetime, cost of satellite transmission), to merge several applications that have been defined separately, and generate code that can be deployed on the float. An experiment has been conduced to develop an improved version of the seismic detection application, and a whales detection application. The developers of these applications were a geoscientist and a biologist that are not experts in embedded software development. The results show that the developers were able to successfully develop applications by themselves and to take into account the limited capacities of the float. The applications have been tested on the float's electronic board with real data.The proposed approach allows scientists to develop monitoring applications for autonomous floats without the help of embedded software experts. We expect that it will enable the multidisciplinary monitoring of the oceans at the global scale.

Abstract

International audience

Additional details