A Pragmatic Relay Placement Approach in 3-D Space and Q-Learning-Based Transmission Scheme for Reliable Factory Automation Applications

Abstract

The biggest challenge in adopting industrial wireless sensor networks (IWSNs) for factory automation applications is to provide low latency and highly reliable communication in harsh factory environments. IEEE 802.15.4e low latency deterministic network (LLDN) mode attempts to address this requirement at the medium access control (MAC) layer. However, the measures offered by this mode are inadequate considering realistic factory environments, suffering from noise, interference, multipath fading, and resulting in frequent packet losses. Cooperative diversity using relay nodes and incorporation of forward error correction (FEC) techniques are the two conventional ways to enhance communication reliability. However, the challenge lies in the placement of relay nodes considering a realistic three-dimensional (3-D) factory space and satisfying various physical, performance, and energy-related constraints. Moreover, the versatile and dynamic behavior of factory environment demand that the solutions offered to enhance communication reliability must be generic and adaptive, thereby eliminating the need for unnecessary redesigns. This paper proposes a twofold solution to enhance the communication reliability offered by 802.15.4e LLDN. First, an efficient and pragmatic relay-placement strategy based on rainbow product ranking algorithm for a 3-D factory space. Second, an adaptive transmission scheme (ATS) inspired from reinforcement learning (RL) technique called Q-learning is proposed, which incorporates cooperative diversity and Reed Solomon (RS) block codes. The effectiveness of the proposed solution is established and demonstrated using a real-world case study.