Demonstration of a WBAN testbed

Adaptation of the network architecture through on-demand data forwarding is an efficient mechanism to provide robustness against long outages in WBANs. We developed an experimental testbed that provides online observation of the network behavior for different data propagation approaches in WBANs. The demonstration shows how different approaches deal with special challenges in WBANs such as low quality of wireless links, topology variations due to posture changes, and mobility. Moreover, sensor nodes can be configured online to investigate how different protocols react in various situations.

Testbed setup

The testbed uses MyriaNed wireless nodes. These motes feature an ATMEGA128 microcontroller and a Nordic nRF24L01 radio chip as transceiver. The radio chip works in the 2.4GHz ISM band, using a data rate of 2Mbps and a 32 bytes fixed packet size. The radio can be set in RX, TX, or Standby modes. In TX mode, four transmit power levels (-18, -12, -6, 0dBm) can be adopted for the radio. This allows us to test each protocol in various network conditions.

11 wireless nodes (including the gateway) are deployed on different positions of the body. The gateway always uses the highest available transmit power (0dBm). It always listens to all nodes to receive their packets. Body nodes listen to the gateway in any round independent of the used protocol. This is the minimum listening activity of each node which is crucial for keeping nodes synchronized. Besides that, more listening to the other nodes can be performed according to the selected protocol. Here is a deployment example of our testbed.

The contol menu for parameter seeting in the GUI.

Configuring the WBAN parameters

Several parameters and commands can be set by the GUI to determine the behavioral mode of the WBAN. The following are the controllable parameters.

  • Protocol type
  • Three protocols are implemented and can be set through the computer GUI. Star topology, a full-gossip data propagation, and the ODLF mechanism are selectable. In the star network, every node only transmits its own data to the gateway. In the full-gossip protocol, every node listens to all other nodes and if it receives any packet, it propagates that as well as its own sensed data. In each frame, a node can at most transmit two data items in addition to its own sensed data. The ODLF approach works as explained above.
  • Transmit power
  • There are four transmit power levels in the sensor nodes. A uniform power level is set for all nodes except for the gateway that always uses the highest power. By performing experiments with different power levels, we can evaluate a protocol in networks with different connectivity.
  • Sampling period
  • The period of data generation by nodes can also be set. At the moment, a uniform sampling period is set for all nodes. This parameter is important to observe the data delivery provided by each protocol. If the sampling period is equal to the frame length, each data item has only one opportunity to be transmitted by the source node. In contrast, when the sampling rate is a multiple of the frame length, each data item has more than one opportunity to be sent. Such retransmissions can deal with intermittent disconnections that are very likely to happen in wireless communication.
  • Disconnection threshold
  • This parameter is only sent to the gateway and is used for maintaining the list of the nodes without a sufficiently good links to the gateway. In the demo, we use a uniform value for all nodes in the WBAN.

    Here is a snapshot of the parameter setting interface in the GUI.
    The contol menu for parameter seeting in the GUI.

    Online WBAN performance observation

    Several online plots show different performance metrics. The following information is presented.

  • Performance graphs:
  • The latency of arrived data items to the gateway, the packet reception and data delivery ratio, and the age of data items from each node in the WBAN at the gateway can be observed online. This information is extracted from the data that the gateway sends to the node connected to the laptop. This data includes the time stamp attached to the last data item received from each body sensor node.
  • Disconnected nodes
  • At each round the gateway announces the requested set which contains the nodes that do not have a proper link to the gateway. This is then shown by the GUI.
  • Radio energy consumption
  • Depending on the active protocol and the connectivity level of the network, nodes may perform different amounts of radio activity and so consume different amounts of energy. The radio energy consumed by the sensor nodes is shown online.

    Here is a snapshot of the plot for radio energy consumption, for instance.
    The Energy Consumption of plot in the GUI.

    The testbeb will be demonstrated during the IEEE SECON 2012 conference in Seoul, June 2012. Here is an abstract about the SECON12 Demo (pdf).

    In the case of any question about the WBAN testbed, please contact Majid Nabi or Marc geilen.