We consider a time-slotted multihop wireless sensor network used for a remote estimation application. Sensor nodes sample processes of interest and convergecast the sampled data to a sink over the multihop wireless network. The processes are remotely estimated at the sink. For such applications, the Age of Information (AoI) is a relevant performance metric. Motivated by real-world protocols, such as 802.15.4e TSCH, we design novel periodic deterministic sampling and scheduling algorithms for minimizing the average AoI (AAoI) for convergecast in tree network topologies with K-hop interference constraints. We obtain an analytical characterization of the AAoI in terms of the period of the policy as well as hop distances of the source nodes to the sink. For periodic deterministic policies, the AAoI minimization problem is formulated as an integer programming problem which is relaxed to a linear programming problem to obtain numerical lower bounds. Novel analytical lower bounds on the AAoI for periodic deterministic policies in tree networks with K-hop interference constraints are obtained. We simulate and compare the AAoI performance of our periodic deterministic policies and show that they have an AAoI close to the lower bounds for that period. Through simulations, we also show that the AAoI of the proposed policies is smaller compared to other multihop AoI-minimizing policies in prior work, such as Age-Difference and Age-Debt. We also discuss extensions to the case of general functions of the AoI.

Sequential decision making for wireless networked control applications where the decision making agent is not colocated with the controlled system has been considered in prior work. The decision making agent obtains state observations and reward observations from the system through a communication network which may delay and lose such observations. Prior work has considered the relevance of information freshness for decision making agents which assume perfect knowledge of the system model. Motivated by the need to consider learning based decision making agents which have a fundamental exploration-exploitation tradeoff, we investigate the remote control of bandits over a bottlenecked communication network modelled as a Geo/Geo/1 queue. Using simulations, we investigate the regret performance of standard bandit algorithms such as$\epsilon$-greedy, upper confidence bound (UCB), and Thompson sampling in this context. We also propose reward-queue based bandit algorithms and evaluate their regret via simulations. We study the regret as a function of the queue arrival and service parameters as well as of the queueing discipline and investigate whether information freshness is relevant for remote learning based control of bandits.

We consider the optimal design of the frame-based irregular repetition slotted ALOHA (IRSA) protocol for minimizing the average age of information (AAoI) in grant-free massive machine-type communications (mMTC). To this end, we first characterize the AAoI as a function of the number of user elements (UEs), the frame duration, and the repetition distribution of IRSA. We present this characterization for IRSA schemes with packet recovery at the end of frame and packet generation either at the beginning of the frame or just in time before first transmission in a frame. We also propose and characterize the AAoI of a novel early packet recovery method which further reduces the average age of information. In all cases, the analysis reveals that, as a function of normalized channel traffic (defined as the ratio of number of UEs to frame duration), the AAoI first decreases linearly due to more frequent updates received from the UEs, and increases sharply beyond a critical point due to packet recovery failures caused by collisions. We then consider the problem of minimizing AAoI by optimizing over the normalized channel traffic and repetition distribution for all the proposed sampling and recovery schemes. The optimization problem is challenging since the objective function is semi-analytical and can only be completely characterized using simulations. In an asymptotic regime where the number of UEs as well as the frame size is large, we characterize the AAoI using upper and lower bounds. We also obtain a locally optimal normalized channel traffic and repetition distribution using differential evolution. Based on the insights obtained from the asymptotic analysis, we also propose a pragmatic approach to obtain a normalized channel traffic and repetition distribution for AAoI reduction in the non-asymptotic case. Finally, we empirically show that our AAoI minimizing schemes outperform conventional throughput optimal schemes.

We study stability properties of single-source single-destination delay tolerant networks with random packet arrivals and buffered relay nodes, using source spray and wait routing. We derive the stability threshold, the supremum of arrival rates for which the source queue is stable, as a function of the buffer space at the relays. In particular, we show that the stability threshold only doubles as the relays' buffer size increases from one to infinity for a network without packet delivery feedback. For the system without packet delivery feedback, we propose lower bounds for the average queueing delay and average delivery delay for packets and compare with simulations. We also obtain the stability threshold numerically for a network with instantaneous packet delivery feedback.

The optimal power-delay tradeoff is studied for a time-slotted independently and identically distributed fading point-to-point link, with perfect channel state information at both transmitter and receiver, and with random packet arrivals to the transmitter queue. It is assumed that the transmitter can control the number of packets served by controlling the transmit power in the slot. The optimal tradeoff between average power and average delay is analyzed for stationary and monotone transmitter policies. For such policies, an asymptotic lower bound on the minimum average delay of the packets is obtained, when average transmitter power approaches the minimum average power required for transmitter queue stability. The asymptotic lower bound on the minimum average delay is obtained from geometric upper bounds on the stationary distribution of the queue length. This approach, which uses geometric upper bounds, also leads to an intuitive explanation of the asymptotic behavior of average delay. The asymptotic lower bounds, along with previously known asymptotic upper bounds, are used to identify three new cases where the order of the asymptotic behavior differs from that obtained from a previously considered approximate model, in which the transmit power is a strictly convex function of real valued service batch size for every fade state.

We study the tradeoff between the average error probability and the average queueing delay of messages which randomly arrive to the transmitter of a point-to-point discrete memoryless channel that uses variable rate fixed codeword length random coding. Bounds to the exponential decay rate of the average error probability with average queueing delay in the regime of large average delay are obtained. Upper and lower bounds to the optimal average delay for a given average error probability constraint are presented. We then formulate a constrained Markov decision problem for characterizing the rate of transmission as a function of queue size given an average error probability constraint. Using a Lagrange multiplier the constrained Markov decision problem is then converted to a problem of minimizing the average cost for a Markov decision problem. A simple heuristic policy is proposed which approximately achieves the optimal average cost.

Average-delay optimal scheduling of messages arriving to the transmitter of a point-to-point channel is considered in this paper. We consider a discrete time batch-arrival batch-service queueing model for the communication scheme, with service time that may be a function of batch size. The question of delay optimality is addressed within the semi-Markov decision-theoretic framework. Approximations to the average-delay optimal policy are obtained.