Timer-Based Distributed Channel Access for Control Over Unknown Unreliable Time-Varying Communication Channels

Abstract

We consider the resource allocation problem for a system consisting of multiple control subsystems that share an unknown unreliable time-varying wireless channel. We propose a novel method for granting channel access deterministically without requiring information exchange between subsystems, based on local timers for prioritizing channel access with respect to a local cost that takes into account the (initially unknown) channel conditions. We propose a novel setup which has the capability of learning the parameters of imperfect communication links while ensuring distributed collision-free implementation. More specifically, we adopt learning algorithms for estimating the channel quality and, hence, define the local cost as a function of this estimation and control performance. The efficacy of this mechanism is evaluated via simulations.

Distributed Channel Access for Control Over Unknown Memoryless Communication Channels

“…We assume the actuators are collocated with the controllers, but the sensors need to transmit their data over a capacity-constrained time-slotted network. We consider the scenario in which the subsystems exchange no explicit information and they coordinate for channel access in a distributed manner 1 . The effects of state quantization and transmission delays are considered negligible and are thus ignored henceforth.…”

“…The effects of state quantization and transmission delays are considered negligible and are thus ignored henceforth. 1 The term decentralized control is often used to describe scenarios in which determining the control inputs requires no explicit information exchange between subsystems. In this work, we use the term distributed to describe the channel access problem in accordance with the WNCSs literature [12].…”

Power Allocation of Sensor Transmission for Remote Estimation Over an Unknown Gilbert-Elliott Channel

“…At any given time, the underlying channel parameters are learned with respect to the available state and observation history. These parameters are then adopted in (6) to determine the corresponding optimal schedule. By enforcing the determined policy for power management in the current step, a new observation is obtained which is used for learning the parameters more accurately and readjusting the policy accordingly.…”