John-Davis Chukwuemeka Oyedum, a communication network researcher, says characterizing machine-induced impulsive noise could accelerate the transition to fifth-generation (5G) enabled smart manufacturing in Nigeria and other emerging economies.
Oyedum, a Doctor of Philosophy (Ph.D.) student in Electrical and Computer Engineering at Rutgers University, New Jersey, said the approach could help manufacturers overcome severe wireless communication challenges caused by existing industrial machinery.
He said his research focuses on developing ultra-reliable, next-generation wireless networks for automated smart manufacturing at the Wireless Information Network Laboratory (WINLAB), a premier research facility at Rutgers University.
He noted that industrial environments pose unique challenges to wireless networks because heavy machinery generates sudden, high-power bursts of electromagnetic interference.
‘Standard 5G networks were designed under the assumption that factory noise is soft and continuous, like a gentle background hum,’ Oyedum explained.
‘However, real-world empirical measurements across sub-6 gigahertz (GHz) frequencies and the Frequency Range 3 (FR3) spectrum inside active factories showed the exact opposite. Heavy machinery such as electric welders and industrial motors generate sudden, high-power bursts of electrical static, known as machine-induced impulsive noise,’ he said.
He explained that these noise bursts severely disrupt wireless signals used to control automated machinery and industrial robots, causing data loss, signal drops, and critical communication delays.
This challenge is increasingly vital as industries adopt wireless sensors, robotics, and automated production systems. Consequently, his research seeks to model the behavior of impulsive noise and develop software-driven physical-layer solutions to mitigate its disruptive effects.
Oyedum highlighted that his research team has also engineered a physical-layer digital twin framework to simulate wireless communication within industrial environments.
‘A digital twin is essentially a virtual 3D replica of a real factory running on a computer,’ he said. ‘Most existing digital twins map physical machines and routers but overlook what happens to invisible radio waves in the air. Our framework allows engineers to simulate actual radio signals in real time and stress-test wireless networks before installing expensive physical infrastructure.’
This work was presented at the Institute of Electrical and Electronics Engineers (IEEE) Future Networks World Forum (FNWF) 2025.
He added that findings presented at the IEEE Global Communications Conference (GLOBECOM) 2025 revealed that as networks transition toward sixth-generation (6G) spectrums with higher subcarrier spacing, wireless links become even more vulnerable to impulsive noise.
‘Next-generation 5G and 6G networks use shorter signals to transmit information more rapidly,’ Oyedum noted. ‘This means a single, split-second burst of machine noise can completely wipe out an entire message. As networks become faster, they require stronger, targeted noise protection rather than relying on speed alone.’
To classify this dynamic interference in real time, the research team employed a statistical framework known as a Hidden Markov Model (HMM), parameterized using the Baum-Welch algorithm to identify underlying noise burst patterns.
Oyedum likened the HMM framework to a smart weather forecasting system for radio waves that continuously monitors changing conditions and predicts when sudden static ‘storms’ are about to strike, enabling wireless receivers to dynamically adjust and protect incoming data from corruption.
The researcher proposed a two-part defense system combining advanced Forward Error Correction (FEC) codes with spatially separated receiving antennas.
He explained that error-correcting codes-such as Reed-Solomon (RS), Polar, and Low-Density Parity-Check (LDPC) codes-attach redundant data to allow damaged information to automatically repair itself.
‘Multiple receiving antennas can also be positioned across the factory floor,’ he noted. ‘If a machine burst disrupts one antenna, another antenna located a few feet away can still capture a clean signal.’
By integrating this setup with Software-Defined Networking (SDN), a central controller can automatically detect noise bursts and instantly route data through a clean antenna while applying backup error-correction coding.
Oyedum noted that findings across research at WINLAB and the New Jersey Advanced Manufacturing Institute (NJAMI) identified four critical physical parameters that digital models must capture to accurately reflect industrial wireless environments: burst timing, energy spikes, state switching, and spatial spread.
‘Burst timing measures how long each noise burst lasts and how frequently it recurs,’ Oyedum explained. ‘Energy spikes capture the exact power surges across sub-6 GHz and FR3 frequencies. State switching tracks how rapidly an environment flips from quiet conditions to heavy interference, and spatial spread determines how noise travels across metal-heavy factory floors.’
Accurately modeling these four parameters allows engineers to construct virtual simulations that mirror real physical factories. Oyedum emphasized that these insights hold immense promise for accelerating industrial digitalization in Nigeria and other emerging economies.
‘Manufacturers in Nigeria and similar developing hubs are eager to adopt modern 5G sensors, but many still depend on older, heavy machinery that generates massive electrical static,’ he said. ‘Replacing all that equipment or constructing extensive physical shielding is prohibitively expensive. My research offers a practical, software-driven solution that enables factories to run reliable 5G and Beyond 5G (B5G) automation on top of existing machinery.’
He concluded that combining impulsive noise prediction models with strategic antenna placement, higher signal power, and advanced error-correction algorithms will lower the cost barrier to smart manufacturing, ensuring emerging economies can fully leverage Artificial Intelligence (AI) driven robotics, connected Internet of Things (IoT) sensors, and real-time industrial automation.