SIMUTOOLS 2015, the Eighth EAI International Conference on Simulation Tools and Techniques held at the end of August in Athens, was an occasion to discuss and update about the status-of-the-art of simulation methods and software. Among the high-quality papers presented at the event, the one by Daniel Maier, Steffen Moser, and Frank Slomka from the Ulm University in Germany, won the Best Paper Award.
‘Deterministic Models of the Physical Layer through Signal Simulation’ is the title of the study introduced by Daniel Maier to the audience, which focuses of a novel approach for the simulation of the physical layer by utilizing existing software-defined radio implementations to create signals, to calculate interference and to decode signals.
State-of-the-art Current wireless network simulators provide very detailed deterministic models of the network protocol layers. The models simulate the flow of data packets from a transmitter node to one or more receiver nodes. The physical layer (PHY), which is the interface between analog signals and data packets, is usually simulated stochastically by using a bit error rate (BER) based model, where the BER derives from the signal-to-noise ratio (SNR). The result is an extensive simplification of the PHY layer, which hinders a detailed studying and optimization of the PHY itself in the context of the entire communication stack.
A new model In the last years, there have been some attempts to fix the disregard of the physical layer in network simulators. Most of them are hardware-based and therefore tied to the existence of hardware prototypes, or consist of an implementation of a specific protocol. The team from ULM presents a new model taking advantage from the emerging software-defined radios (SDR) and from the availability of ready-made implementations of wireless communication protocols.
Signal simulation: a deterministic model of the physical layer A deterministic model replicates the behavior of a system in a non-random way by simulating the process in the system, so it doesn’t need an existing world counterpart. It comes that some specific assumptions must be considered to set up a deterministic simulation of a physical layer:
- The physical layer generates analog signals based on given digital data packets
- The physical layer reconstructs digital data packets from received analog signals
- The radio channel between transmitter and receiver alters the signal
- Additional interference can occur due to transmitters transmitting at the same time
- Additional thermal noise can occur and alter the signal
In order to meet these assumptions, four elements are necessary:
- A transmitter that creates signals that are sent
- A channel model to delay and attenuate the signal
- A mechanism that rebuilds the state at the receiver’s antenna consisting of different delayed and attenuated signals and thermal noise
- A receiver which can decode signals and verify their correct reception
Results and future perspectives Starting from the implementation of a prototype for the simulation of IEEE 802.11a/g/p and an added integration in the Veins VANET simulator, the papers shows two important outcomes: firstly, stochastic modeling is a valid approximation for collision-avoiding PHYs like CSMA/CA; secondly, the PHY implementation works well and delivers reasonable results. Furthermore, according to the authors of the study, an adaption for other network simulators, such as ns-3, is possible as well as other PHY protocols and other SDR software. Future applications may also include simulation of CDMA for real-time media access and simulation of MIMO-based systems.
The full research will be available online.
