Abstract The design of a Printed Circuit Board (PCB) involves satisfying numerous design constraints in terms of the positioning of components and the routing of conducting tracks or traces between components. One common design constraint is limiting the maximum angular bend that a trace can make (e.g. to 45°) for several potential reasons including reliability. In this paper, we systematically investigate the failure characteristics of PCB external traces with different angle bends to understand the implications in a modern context with modern PCB manufacturing processes. This involves destructive testing of straight traces with no bends, traces with only 45° bends and traces with only 90° bends. There are three aspects to the testing: (1) maximal current testing, in which the maximum current that a trace can withstand before failure is measured, (2) failure location testing for traces with bends i.e. trace failed on straight segment or at the bend, and (3) time to failure testing. We did not find a large difference between the maximum currents that can be withstood by straight traces, traces with 45° bends and traces with 90° bends. However, some of the interesting results are that traces with 45° bends are significantly more likely to fail at the location of a bend than traces with 90° bends, and straight traces take much longer to fail on average than traces with bends for a given test current.

A variety of scheduling strategies can be employed in wireless systems to satisfy different system objectives and to cater for different traffic types. Static persistent resource allocations can be employed to transfer small M2M data packets efficiently compared to dynamic packet-by-packet scheduling, even when the M2M traffic model is non-deterministic. Recently, adaptive persistent allocations have been proposed in which the volume of allocated resources can change in sympathy with the instantaneous queue size at the M2M device and without expensive signaling on control channels. This increases the efficiency of resource usage at the expense of a (typically small) increased packet delay. In this paper, we derive a statistical model for the device queue size and packet delay in static and adaptive persistent allocations which can be used for any arrival process (i.e., Poisson or otherwise). The primary motivation is to assist with dimensioning of persistent allocations given a set of QoS requirements (such as a prescribed delay budget). We validate the statistical model via comparison with queue size and delay statistics obtained from a discrete event simulation of a persistent allocation system. The validation is performed for both exponential and gamma distributed packet inter-arrivals to demonstrate the model generality.

One of the major limitations of the LTE system in context of Machine-to-Machine (M2M) communications is the excessive control signalling required for dynamic scheduling of radio resources. Although the queue awareness and link adaptation features of the dynamic scheduler can be very beneficial for M2M traffic as well, these two properties come at a cost of significant control overhead. In this paper, we propose an adaptive semi-persistent LTE uplink scheduler which reduces the control overhead by a significant level and utilizes the uplink capacity efficiently to satisfy the Quality of Service (QoS) requirements of M2M traffic. The concept of Semi-Persistent Scheduling (SPS) has already been implemented for Voice over IP (VoIP) service in the LTE standard. However, we have further developed this into an adaptive SPS scheme which can adapt to the sudden bursts of stochastic M2M traffic flow by utilizing the device buffer information as well as respond to the channel fluctuations caused by fading while satisfying the M2M packet delay constraints. The performance of the proposed adaptive SPS scheme has been compared to an LTE dynamic scheduler using the OPNET simulation software for fading channels. The simulation results demonstrate the adaptive SPS scheme to outperform the dynamic scheduler in terms of both control channel utilization and satisfying QoS requirements of M2M traffic.

Some M2M applications such as event monitoring involve a group of devices in a vicinity that act in a co-ordinated manner. An LTE network can exploit the correlated traffic characteristics of such devices by proactively assigning resources to devices based upon the activity of neighboring devices in the same group. This can reduce latency compared to waiting for each device in the group to request resources reactively per the standard LTE protocol. In this paper, we specify a new low complexity predictive resource allocation algorithm, known as the one way algorithm, for use with delay sensitive event based M2M applications in the LTE uplink. This algorithm requires minimal incremental processing power and memory resources at the eNodeB, yet can reduce the mean uplink latency below the minimum possible value for a non-predictive resource allocation algorithm. We develop mathematical models for the probability of a prediction, the probability of a successful prediction, the probability of an unsuccessful prediction, resource usage/ wastage probabilities, and mean uplink latency. The validity of these models is demonstrated by comparison with the results from a simulation. The models can be used offline by network operators or online in real time by the eNodeB scheduler to optimize performance.

We propose a predictive resource allocation scheme for the LTE uplink based upon Maximum Likelihood Estimation of event propagation characteristics for M2M/Smart Grid applications. The LTE eNodeB estimates the inter-sensor propagation time of a disturbance using the pattern and timing of received Scheduling Requests (SRs) from sensors and then proceeds to predict the time at which the disturbance will reach downstream sensors, facilitating predictive uplink grants for these sensors in order to reduce the mean latency of their uplink data packets by up to 50% (according to a performance analysis) compared to the existing standard reactive LTE uplink resource allocation scheme. A further benefit is that when a predictive resource allocation is successful, the sensor does not need to send an SR, thereby freeing up uplink resources which can be critical with M2M communications. We consider various transition strategies from the estimation to prediction phases which reflect the compromise between estimation speed and accuracy, and also examine the concept of early and late prediction.

Wireless Network Control Systems (WNCSs) have been developed and deployed in many applications because of the flexibility afforded by the wireless communications medium in accommodating the control loop. In this paper, we examine a wide area WNCS using LTE as the wireless fabric. We introduce a novel concept known as “reflection” which is a subscription model that allows an LTE eNodeB to inform certain nodes in a group about low level radio signalling from certain other nodes in the group before any application data has been sent. The mean Round Trip Time (RTT) for control in an LTE based wide area WNCS which is enhanced for reflection is reduced by at least 6ms compared to a standard LTE based wide area WNCS. This is demonstrated using an OPNET simulation model that compares a reflection based architecture with a reference architecture. The technique has applications in Smart Grid control applications (such as Fault Detection and Isolation) and in other vertical industries in which WNCSs may be relevant.