Fuzzy Systems for Spectrum Access, Mobility and Management for Cognitive Radios

11.2 Soft Computing And Cognitive Cycle

CR was first coined by Mitola (1999) who represented its major functions to adapt the transmission parameters in changing environments through a cognitive cycle. This six stage cognition cycle is briefed as follows.

• •

  Observe: Know the information of operating environment through sensing and signaling mechanisms.

• •

  Orient: Evaluate this information to determine its significance and relevance.

• •

  Plan: Based on this evaluation, the radio determines its options or alternatives for resource optimization.

• •

  Decide: An alternative is chosen that evaluates more favorably than other options, including the current ongoing action.

• •

  Act: The radio implements the decisions taken for resource optimizations. These changes are then reflected in the interference profile presented by the cognitive radio in the outside world.

• •

  Learn: Throughout the process, the radio uses its observations and decisions to improve its own operation, creating new modeling states and alternatives.

Cognitive cycle as briefed above, forms the heart of a Cognitive Radio. Out of the six stages of cognitive cycle, the stages “decide” and “act” is important functionalities that may work efficiently using soft computing techniques. Thus, artificial intelligence and soft computing techniques form important enabling techniques for the success of CR. Some of the key functionalities (Mitola 1999), (Linda 2009) supported by these techniques are as follows:

• 1.

  Optimization of Resources:

  • •

    CR to make best use of use of resources it has been given or it has identified for itself while protecting the rights of others

  • •

    Auto reconfigurability of transmission profile and operating parameters and redirecting resources around the network e.g. “who gets what”

Thus the purpose is to evaluate all choices or solutions to find the one with highest rank and adapting to the “best” parameters. Thus, this amounts to the cross layer optimization largely focussing on physical (PHY) and medium access control (MAC) layers.

• 2.

  Signal Shaping: CR needs to represent, organise, store and analyse the collected knowledge so that appropriate optimization routines can be evoked. Some uniform mechanism for knowledge representation to be selected and then a suitable form of reasoning could be used. The operating parameters that need manageable decisions are: waveform type and antenna usage. The decisions are to be taken from the choices available as shown in Figure 5.1. Regulatory policies like any other knowledge need to be represented in a machine readable manner and the CR needs to reason about the policies and make decision.

• 3.

  Knowledge Gain Through Learning: Learning is crucial when dealing with unknown and unplanned scenarios and thus can be of great benefit to a CR in improving the overall performance. There are many learning techniques ranging from very simple memorising techniques to complex ones e.g. artificial intelligence and machine learning.