@InProceedings{Sputh07,
  title =        "{A} {V}ersatile {H}ardware-{S}oftware {P}latform for {I}n-{S}itu {M}onitoring {S}ystems",
  author=        "Sputh,  Bernhard H.C. and Faust,  Oliver and Allen,  Alastair R.",
  editor=        "McEwan,  Alistair A. and Schneider,  Steve and Ifill,  Wilson and Welch, Peter H.",
  pages =        "299--312",
  booktitle=     "{C}ommunicating {P}rocess {A}rchitectures 2007",
  isbn=          "978-1-58603-767-3",
  year=          "2007",
  month=         "jul",
  abstract=      "In-Situ Monitoring systems measure and relay environmental
     parameters.
From a system design perspective such devices
     represent one node in a network. This paper aims to extend
     the networking idea from the system level towards the design
     level. We describe In-Situ Monitoring systems as network of
     components. In the proposed design these components can be
     implemented in either hardware or software. Therefore, we
     need a versatile hardware-software platform to accommodate
     the particular requirements of a wide range of In-Situ
     Monitoring systems. The ideal testing ground for such a
     versatile hardware-software platform are FPGAs (Field
     Programmable Gate Arrays) with embedded CPUs. The CPUs
     execute software processes which represent software
     components. The FPGA part can be used to implement hardware
     components in the form of hardware processes and it can be
     used to interface to other hardware components external to
     the processor. In effect this setup constitutes a network of
     communicating sequential processes within a chip. This paper
     presents a design flow based on the theory of CSP. The idea
     behind this design flow is to have a CSP model which is
     turned into a network of hardware and software components.
     With the proposed design flow we have extended the
     networking aspect of sensor networks towards the system
     design level. This allows us to treat In-Situ Measurement
     systems as sub-networks within a sensor network.
     Furthermore, the CSP based approach provides abstract models
     of the functionality which can be tested. This yields more
     reliable system designs."
}