CSP4WSN

The subject of this proposal is the development of techniques and tools which permit new modes of cooperative working and self-reconfiguration in wireless sensor networks. The foundational methodology will be built on the formal methods of Communicating Sequential Processes. It addresses generic classes of applications, such as environmental monitoring and emergency response, where reliable, dynamic reconfigurability of networks of nodes is essential.

This proposal addresses two fundamental problems in wireless sensor networking. The first is that there is a need to develop a system paradigm which allows the WSN designer to plug-and-play "objects" in such a way that the system is truly composable, and that certain guarantees can be made about the liveliness of these objects and their communications. The design formalism should make no distinction between hardware and software, allowing heterogeneous "objects" to occupy the one design space.

The second is that systems need to be able to reconfigure themselves dynamically in the face of a changing environment. Such reconfiguration needs to result in a system which has exactly the same guarantees as the original one.

The theory of Communicating Sequential Processes (CSP) describes systems in terms of entities (processes) which interact using events (which can be thought of as messages). It is an important part of CSP that if you hide the interactions between a group of processes, but not the interactions between that group and the "rest of the world" (the environment), that you end up with another process. That is, processes compose. CSP permits the reliable construction of composable, scalable, provable networks whose overheads are very lightweight. Lightweight processing and communications is essential where nodes are remotely situated. The CSP formalism has 20 years of solid research behind it, in which a variety of rules have been discovered and proved, and the methodology has been packaged up into programming abstractions with bindings to commonly used programming languages such as C and Java. Industrial-strength proof tools are also available.

For the WSN designer, these qualities are highly desirable. Questions of scalability are at the heart of many WSN problems. When a system becomes large and complex, it is essential that the semantics of the system are composable. Objects and their interactions should be describable in a simple way. The use of formal methods as the basis for the proposed methodologies brings certain guarantees relating to the reliability of the behaviour of sensor networks. This is especially desirable in applications where the nodes are relatively inaccessible and/or demand a long maintenance interval. There will be a need to develop appropriate reliable communication protocols for dynamically changing systems;

The proposal is to develop a toolkit (with a formally proved foundation of CSP) which will allow WSN developers to (i) specify plug-and-play sensor node architectures and networks, with reliable and provable attributes and capabilities; (ii) experiment with what-if Hardware–Software partitioning; and (iii) design systems which self-reconfigure in a robust way.

This generic technology will be driven by, and tested in, two application areas.
(a) The first type of application relates to systems where a local change in conditions implies that the cluster of nodes should be able to reconfigure in order to better make the required measurement. This could include applications such as environmental monitoring, where a local change of some parameter requires a different sensor network configuration for optimal processing. In remote areas, the sensor network could be used to provide an alarm dependent on the early detection of forest fires, floods, chemical spills. This might also include security applications, where biosensors are being used to detect contaminants or other dangerous materials.
(b) The second application type is a more extreme example of the first, which might include the degradation of part of the sensor network itself (nodes and/or communications links). The remaining nodes need to reconfigure themselves in order that the system may carry on working. For example, in disaster management (eg. fire, terrorist incident, industrial accident, mine cave-in) where the sensor network would aid the emergency rescue operations on multiple levels: people inside the affected area could be guided to safety by the nodes; the emergency services could get a better idea of the situation; etc.

Wireless sensor networks
Reconfigurability
Hardware-software co-design
Formal methods
Communicating Sequential Processes
CSP
Cooperating objects
Environmental monitoring
Disaster management.

Experts in embedded systems, communications, control, formal methods, relevant sensor technologies, WSN lightweight node technology.

(1) Design and test of the sensor nodes. Or adaptation of existing nodes.
(2) Design of the process oriented architecture. Development of reliable software components based upon formal methods.
(3) Evaluation of wireless communication standards to identify suitable ones for these applications.
(4) Design and implementation of a network protocol which can adapt itself to a changing network topology.
(5) Field testing of the systems.

Industrial partners / end users, with a background of environmental monitoring or emergency systems. Or other applications with similar system requirements.

Research Groups in the areas of WSN, embedded systems, communications, control, and formal methods.

Yes