Lawrence, Yao Receive NSF Award to Develop NextG Communications for Smart Manufacturing

Imagine an unmanned robotic warehouse forklift steadily approaching an aisle where boxes have fallen. Cameras may capture the hazard, but transmitting and processing the video back to the forklift is not instantaneous, especially over a busy manufacturing network. That delayed warning could leave the forklift on a dangerous path that risks the safety of the facility’s equipment and personnel.  

Delivering the information that smart manufacturing operations need, when they need it, is the goal of Victor Lawrence, senior research scientist, and Yu-Dong Yao, professor, from the Department of Electrical and Computer Engineering at Stevens. They are leading a collaboration with MIT, Nokia Bell Labs and Finnish partners to conduct a three-year National Science Foundation project, “Collaborative Research: VINES: Track 2: NSF-BF: Mission Critical Semantic Communications in NextG Networks for Smart Manufacturing.” 

“Today’s networks can move large amounts of data, but machines also need the right information at the right time,” said Yao, who has been instrumental in shaping and conducting the research. “We aim to help wireless networks deliver the information most relevant to a task, so robots and factory systems can make timely, reliable decisions.” 

Teaching wireless networks to communicate what matters most 

The key to the project is “semantic communication,” an approach that sends the information machines need to complete a task rather than transmitting every detail that sensors collect. 

By sending only the most relevant information, the system aims to reduce network traffic and help machines respond more quickly. For example, instead of receiving and having to review the entire video, that robotic forklift would get a clear warning: “the aisle is blocked,” along with the obstacle’s location. 

With Yao, Lawrence is exploring how future wireless networks can help machines make faster, more informed decisions in complex environments.

With its ability to reduce data transmission, improve efficiency and supporting intelligent networks, semantic communication becomes a powerful tool to drive modern communication systems.  

Also fundamental to the research is the concept of a “digital twin”: a computer model of a physical facility that is updated using sensor information, combining observations from different locations into a shared view.  

Artificial intelligence will help integrate sensing, computation and communication. 

“We will develop AI methods that consider the task, the surrounding environment and the network conditions,” Yao explained. “For example, to avoid a collision, a nearby worker’s location and movement matter more than background details. The system can also request more information when it needs a clearer picture.”  With that groundwork, the team will combine four approaches to help machines share information efficiently: 

• Task-oriented resource management: Distribute wireless resources according to what each part of the process needs. 

• Multimodal situation awareness: Combine camera feeds, radio measurements and other sensors to reliably understand the surroundings. 

• Semantic scene representations: Efficiently describe objects, their locations and how they relate to each other. 

• Adaptive coding: As network conditions change, automatically adjust how the data is compressed and protected, preserving the most important details. 

Together, these methods aim to reduce network traffic while ensuring machines receive the information they need to act. The resulting system will coordinate what to sense, where to process it and what to transmit.

“Robots move, wireless signals fluctuate and sensors provide incomplete views,” Yao noted. “We must evaluate the entire process, from sensing to communication to action, under these changing conditions. Our project will also model wireless coverage, helping the system anticipate both physical hazards and communication problems while consuming less data.” 

Robots could use nearby sensors to identify hazards and share essential information with other machines. And that forklift on a collision course with a fallen box? It could reroute before reaching the obstruction. 

For factories, that could mean safer operations, less downtime and more efficient use of network and computing resources. Information about each robot’s movements could also help the network anticipate and adjust for changes in wireless coverage.   

Bringing NextG technology to life 

This project builds on Stevens’ expertise in AI-based wireless communications, spectrum sensing, signal and system identification, robotics and advanced manufacturing, including research through the Center for Innovative Computing and Networked Systems. Co-principal investigator Lawrence holds more than 50 U.S. and international patents, and in 2025 the White House honored him with the National Medal of Technology and Innovation for his lifetime of achievements helping build the foundations for everything from digital modems, high-definition television, digital-video technologies and smartphones to subsea fiber-optic cable systems. 

Further bolstering the potential for technical innovation within this project are collaborations with MIT, Nokia Bell Labs and Finnish experts in robotics and industrial applications.  

The team will validate its approach in stages, beginning with laboratory testbeds at Nokia Bell Labs in New Jersey, then moving to robotics experiments in realistic smart manufacturing environments and industrial field trials in Finland. Researchers will measure communication delays, data volume, sensing accuracy and how reliably robots complete their tasks. They will also evaluate whether the system correctly answers questions and responds to commands. 

This NextG wireless project will also support next-generation engineers. A Stevens Ph.D. student will help develop and evaluate algorithms through simulations and experiments. Lawrence and Yao are also planning to bring the research into Stevens classrooms and share results through publications, software and datasets. 

“We are excited to bring our work in wireless communications and AI to address practical manufacturing problems and test our ideas with real robots and networks,” Yao said. “The collaboration also offers valuable hands-on research experience for our students. We hope this work will help industrial networks adapt to changing conditions, improve productivity and, most importantly, make manufacturing environments safer.”