Wireless Sensor Networks for Enabling Smart Production Lines in Industry 4.0

Abstract

With the deployment of data-driven assembly and production factories, challenges arise in sensor data acquisition and gathering. Different wireless technologies are currently used for transferring data, each with different advantages and constraints. In this paper, we present a hybrid network architecture for providing Quality of Service (QoS) in an industrial environment where guaranteed minimal data rates and maximal latency are of utmost importance for controlling devices and processes. The location of the access points (APs) is determined during the initial network-planning action, together with physical parameters such as frequency, transmit power, and modulation and coding schemes. Instead of performing network-planning just once before the network rollout, the network is monitored continuously by adding telemetry data to the frame header of all data streams, and the network is automatically reconfigured in real-time if the requirements are not met. By not using maximum transmit powers during the initial roll-out, more APs are needed, but coverage is guaranteed when new obstructions such as metallic racks or machinery are added. It is found that decreasing the transmit power by 6 dB gives the best trade-off between the number of required APs and network robustness. The proposed architecture is validated via simulations and via a proof-of-concept setup.

MmWave Physical Layer Network Modeling and Planning for Fixed Wireless Access Applications

“…In addition, non-line-of-sight links can be considered, e.g., by searching for paths with a single building reflection and using typical reflection loss values from the literature. The implementation of in-band telemetry monitoring allows for changing network configurations to maintain QoS [ 64 ] when the environment changes, e.g., when new buildings are built, when vegetation is added, or when devices in the network are broken. On higher layers, scheduling and medium access control mechanisms need to be investigated.…”

Geodesic Path Model for Indoor Propagation Loss Prediction of Narrowband Channels

“…Usually the path loss exponent n is either fitted to LOS measurements at a range of distances or set to the free space path loss (FSPL) exponent of 2. We prefer the latter, but because certain environments have a lower PL exponent than 2, such as highly metallic industrial environments or in this case a ship, both our own measurements as well as findings in literature indicate a PL exponent lower than 2 [1,[35][36][37]. As such, in our case, we will use a PL exponent of n = 2 for the office environment and n = 1.15 for the ship environment.…”

“…Radio channel models characterize radio propagation for a certain frequency and type of environment and are valuable for the design of wireless communication systems. Path loss (PL) models characterize signal attenuation between a transmitting (TX) and receiving (RX) antenna and allow signal strength prediction and coverage calculations during the network planning phase [1], but can also be used for sensing [2] and localization [3] applications.…”