Innovations-based priority assignment for control over CAN-like networks

Abstract

Innovations-based priority assignment for control over CAN-like networks.In: 54th IEEE Conference on Abstract-We present an innovations-based prioritization mechanism to efficiently use network resources for data gathering, without compromising the real-time decision making capability of the control systems. In the envisioned protocol, each sensor assigns the Value of Information (VoI) contained in its current observations for the network as the priority. Tournaments are used to compare priorities and assign transmission slots, like in the CAN bus protocol. By using a rollout strategy, we derive feasible algorithms for computing the VoI-based priorities for the case of coupled and decoupled systems. In the case of decoupled systems, performance guarantees with regard to the control cost of the VoI-based strategy are identified. We illustrate the efficiency of the proposed approach on a platooning example in which the vehicles receive measurements from multiple sensors.

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

“…To capture the dynamic nature of WNCSs and allow for flexibility, dynamic schedulers have been proposed where, at each instance, a scheduling unit grants channel access based on evaluation of some objective function. For instance, [5], [6] use the discrepancy in state estimates as the objective, [7]- [9] base their evaluation on the growth of the standard quadratic T. Farjam cost (penalizing the error from the deviation from the desired state and the magnitude of the control action), while [10], [11] consider the transmission power consumption in calculations as well.…”

Optimal self-driven sampling for estimation based on value of information

“…This is closely connected to the VOI in [19] which is used in a prioritization mechanism for multi-sensor scheduling.…”

On the resource allocation problem in wireless networked control systems

“…This cost is based on the statistical properties of the sensor measurements. We consider both the scenario where the communication is guaranteed to be successfully completed when a channel is allocated to a specific channel, such as in CAN-like network scenarios for communications [14], and the more general scenario where packet losses may occur during transmission [2]. The channel seen by each agent has different statistical properties, hence each agent has a different probability of success on each channel.…”

“…Inspired by the works of [12], [14], we analyze our setup for the LQG scenario, but unlike their approach, our work does not use current measurements. The communications resources are limited and as a result, which agents are allowed to transmit and which communication resource is allocated to each agent has to be carefully designed by a remote decision maker.…”

“…The communications resources are limited and as a result, which agents are allowed to transmit and which communication resource is allocated to each agent has to be carefully designed by a remote decision maker. To quantify the "value" of the measurements of each sensor, we propose and utilize the concept of "Cost of Information Loss" (CoIL), which is the opposite of the Value of Information (VoI) used in other papers (see, e.g., [14], [18]). This cost is based on the statistical properties of the sensor measurements.…”